/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-hip/checkouts/docs-5.7.1/include/hip/hip_runtime_api.h Source File

/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-hip/checkouts/docs-5.7.1/include/hip/hip_runtime_api.h Source File#

HIP Runtime API Reference: /home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-hip/checkouts/docs-5.7.1/include/hip/hip_runtime_api.h Source File
hip_runtime_api.h
Go to the documentation of this file.
1/*
2Copyright (c) 2015 - 2023 Advanced Micro Devices, Inc. All rights reserved.
3
4Permission is hereby granted, free of charge, to any person obtaining a copy
5of this software and associated documentation files (the "Software"), to deal
6in the Software without restriction, including without limitation the rights
7to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
8copies of the Software, and to permit persons to whom the Software is
9furnished to do so, subject to the following conditions:
10
11The above copyright notice and this permission notice shall be included in
12all copies or substantial portions of the Software.
13
14THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
15IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
16FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
17AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
18LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
19OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
20THE SOFTWARE.
21*/
22
30#ifndef HIP_INCLUDE_HIP_HIP_RUNTIME_API_H
31#define HIP_INCLUDE_HIP_HIP_RUNTIME_API_H
32
33
34#include <string.h> // for getDeviceProp
35#include <hip/hip_version.h>
36#include <hip/hip_common.h>
37
38enum {
43};
44// hack to get these to show up in Doxygen:
54typedef struct {
55 // 32-bit Atomics
56 unsigned hasGlobalInt32Atomics : 1;
58 unsigned hasSharedInt32Atomics : 1;
60 unsigned hasFloatAtomicAdd : 1;
61
62 // 64-bit Atomics
63 unsigned hasGlobalInt64Atomics : 1;
64 unsigned hasSharedInt64Atomics : 1;
65
66 // Doubles
67 unsigned hasDoubles : 1;
68
69 // Warp cross-lane operations
70 unsigned hasWarpVote : 1;
71 unsigned hasWarpBallot : 1;
72 unsigned hasWarpShuffle : 1;
73 unsigned hasFunnelShift : 1;
74
75 // Sync
76 unsigned hasThreadFenceSystem : 1;
77 unsigned hasSyncThreadsExt : 1;
78
79 // Misc
80 unsigned hasSurfaceFuncs : 1;
81 unsigned has3dGrid : 1;
82 unsigned hasDynamicParallelism : 1;
84
85typedef struct hipUUID_t {
86 char bytes[16];
87} hipUUID;
88
89//---
90// Common headers for both NVCC and HCC paths:
91
96typedef struct hipDeviceProp_t {
97 char name[256];
104 int maxGridSize[3];
109 int major;
112 int minor;
130 char gcnArchName[256];
138 unsigned int* hdpMemFlushCntl;
139 unsigned int* hdpRegFlushCntl;
140 size_t memPitch;
163
185
189typedef struct hipPointerAttribute_t {
190 union {
191 // Deprecated, use instead type
194 };
199 unsigned allocationFlags; /* flags specified when memory was allocated*/
200 /* peers? */
202
203// Ignoring error-code return values from hip APIs is discouraged. On C++17,
204// we can make that yield a warning
205#if __cplusplus >= 201703L
206#define __HIP_NODISCARD [[nodiscard]]
207#else
208#define __HIP_NODISCARD
209#endif
210
215// Developer note - when updating these, update the hipErrorName and hipErrorString functions in
216// NVCC and HCC paths Also update the hipCUDAErrorTohipError function in NVCC path.
217
223 // Deprecated
226 // Deprecated
248 // Deprecated
270 // Deprecated
282 704,
284 705,
289 712,
291 713,
293 719,
295 720,
326 // HSA Runtime Error Codes start here.
331 hipErrorTbd
332} hipError_t;
333
334#undef __HIP_NODISCARD
335
342
439
442
467
470 // Extended attributes for vendors
472
479
480#if (defined(__HIP_PLATFORM_HCC__) || defined(__HIP_PLATFORM_AMD__)) && !(defined(__HIP_PLATFORM_NVCC__) || defined(__HIP_PLATFORM_NVIDIA__))
481
482#include <stdint.h>
483#include <stddef.h>
484#ifndef GENERIC_GRID_LAUNCH
485#define GENERIC_GRID_LAUNCH 1
486#endif
487#include <hip/amd_detail/host_defines.h>
488#include <hip/driver_types.h>
489#include <hip/texture_types.h>
490#include <hip/surface_types.h>
491#if defined(_MSC_VER)
492#define DEPRECATED(msg) __declspec(deprecated(msg))
493#else // !defined(_MSC_VER)
494#define DEPRECATED(msg) __attribute__ ((deprecated(msg)))
495#endif // !defined(_MSC_VER)
496#define DEPRECATED_MSG "This API is marked as deprecated and may not be supported in future releases. For more details please refer https://github.com/ROCm-Developer-Tools/HIP/blob/master/docs/markdown/hip_deprecated_api_list.md"
497#define HIP_LAUNCH_PARAM_BUFFER_POINTER ((void*)0x01)
498#define HIP_LAUNCH_PARAM_BUFFER_SIZE ((void*)0x02)
499#define HIP_LAUNCH_PARAM_END ((void*)0x03)
500#ifdef __cplusplus
501 #define __dparm(x) \
502 = x
503#else
504 #define __dparm(x)
505#endif
506#ifdef __GNUC__
507#pragma GCC visibility push (default)
508#endif
509#ifdef __cplusplus
510namespace hip_impl {
511hipError_t hip_init();
512} // namespace hip_impl
513#endif
514// Structure definitions:
515#ifdef __cplusplus
516extern "C" {
517#endif
518//---
519// API-visible structures
520typedef struct ihipCtx_t* hipCtx_t;
521// Note many APIs also use integer deviceIds as an alternative to the device pointer:
522typedef int hipDevice_t;
529typedef struct ihipStream_t* hipStream_t;
530#define hipIpcMemLazyEnablePeerAccess 0
531#define HIP_IPC_HANDLE_SIZE 64
532typedef struct hipIpcMemHandle_st {
533 char reserved[HIP_IPC_HANDLE_SIZE];
535typedef struct hipIpcEventHandle_st {
536 char reserved[HIP_IPC_HANDLE_SIZE];
538typedef struct ihipModule_t* hipModule_t;
539typedef struct ihipModuleSymbol_t* hipFunction_t;
543typedef struct ihipMemPoolHandle_t* hipMemPool_t;
544
557typedef struct ihipEvent_t* hipEvent_t;
567//Flags that can be used with hipStreamCreateWithFlags.
569#define hipStreamDefault 0x00
570
572#define hipStreamNonBlocking 0x01
573
574//Flags that can be used with hipEventCreateWithFlags.
576#define hipEventDefault 0x0
577
579#define hipEventBlockingSync 0x1
580
582#define hipEventDisableTiming 0x2
583
585#define hipEventInterprocess 0x4
586
595#define hipEventDisableSystemFence 0x20000000
596
599#define hipEventReleaseToDevice 0x40000000
600
603#define hipEventReleaseToSystem 0x80000000
604
605//Flags that can be used with hipHostMalloc.
607#define hipHostMallocDefault 0x0
608
610#define hipHostMallocPortable 0x1
611
614#define hipHostMallocMapped 0x2
615
619#define hipHostMallocWriteCombined 0x4
620
625#define hipHostMallocNumaUser 0x20000000
626
628#define hipHostMallocCoherent 0x40000000
629
631#define hipHostMallocNonCoherent 0x80000000
632
634#define hipMemAttachGlobal 0x01
635
637#define hipMemAttachHost 0x02
638
640#define hipMemAttachSingle 0x04
641
642#define hipDeviceMallocDefault 0x0
643
645#define hipDeviceMallocFinegrained 0x1
646
648#define hipMallocSignalMemory 0x2
649
651#define hipDeviceMallocUncached 0x3
652
653//Flags that can be used with hipHostRegister.
655#define hipHostRegisterDefault 0x0
656
658#define hipHostRegisterPortable 0x1
659
662#define hipHostRegisterMapped 0x2
663
665#define hipHostRegisterIoMemory 0x4
666
668#define hipHostRegisterReadOnly 0x08
669
671#define hipExtHostRegisterCoarseGrained 0x8
672
674#define hipDeviceScheduleAuto 0x0
675
678#define hipDeviceScheduleSpin 0x1
679
682#define hipDeviceScheduleYield 0x2
683#define hipDeviceScheduleBlockingSync 0x4
684#define hipDeviceScheduleMask 0x7
685#define hipDeviceMapHost 0x8
686#define hipDeviceLmemResizeToMax 0x10
688#define hipArrayDefault 0x00
689#define hipArrayLayered 0x01
690#define hipArraySurfaceLoadStore 0x02
691#define hipArrayCubemap 0x04
692#define hipArrayTextureGather 0x08
693#define hipOccupancyDefault 0x00
694#define hipOccupancyDisableCachingOverride 0x01
695#define hipCooperativeLaunchMultiDeviceNoPreSync 0x01
696#define hipCooperativeLaunchMultiDeviceNoPostSync 0x02
697#define hipCpuDeviceId ((int)-1)
698#define hipInvalidDeviceId ((int)-2)
699//Flags that can be used with hipExtLaunch Set of APIs.
701#define hipExtAnyOrderLaunch 0x01
702// Flags to be used with hipStreamWaitValue32 and hipStreamWaitValue64.
703#define hipStreamWaitValueGte 0x0
704#define hipStreamWaitValueEq 0x1
705#define hipStreamWaitValueAnd 0x2
706#define hipStreamWaitValueNor 0x3
707// Stream per thread
709#define hipStreamPerThread ((hipStream_t)2)
710
711// Indicates that the external memory object is a dedicated resource
712#define hipExternalMemoryDedicated 0x1
762
892 unsigned char reserved[64];
894
948typedef struct dim3 {
949 uint32_t x;
950 uint32_t y;
951 uint32_t z;
952#ifdef __cplusplus
953 constexpr __host__ __device__ dim3(uint32_t _x = 1, uint32_t _y = 1, uint32_t _z = 1) : x(_x), y(_y), z(_z){};
954#endif
955} dim3;
959typedef struct hipLaunchParams_t {
960 void* func;
963 void **args;
964 size_t sharedMem;
970typedef struct hipFunctionLaunchParams_t {
972 unsigned int gridDimX;
973 unsigned int gridDimY;
974 unsigned int gridDimZ;
975 unsigned int blockDimX;
976 unsigned int blockDimY;
977 unsigned int blockDimZ;
978 unsigned int sharedMemBytes;
991typedef struct hipExternalMemoryHandleDesc_st {
993 union {
994 int fd;
995 struct {
996 void *handle;
997 const void *name;
998 } win32;
999 } handle;
1000 unsigned long long size;
1001 unsigned int flags;
1003typedef struct hipExternalMemoryBufferDesc_st {
1004 unsigned long long offset;
1005 unsigned long long size;
1006 unsigned int flags;
1015typedef struct hipExternalSemaphoreHandleDesc_st {
1017 union {
1018 int fd;
1019 struct {
1020 void* handle;
1021 const void* name;
1022 } win32;
1023 } handle;
1024 unsigned int flags;
1027typedef struct hipExternalSemaphoreSignalParams_st {
1028 struct {
1029 struct {
1030 unsigned long long value;
1031 } fence;
1032 struct {
1033 unsigned long long key;
1034 } keyedMutex;
1035 unsigned int reserved[12];
1036 } params;
1037 unsigned int flags;
1038 unsigned int reserved[16];
1043typedef struct hipExternalSemaphoreWaitParams_st {
1044 struct {
1045 struct {
1046 unsigned long long value;
1047 } fence;
1048 struct {
1049 unsigned long long key;
1050 unsigned int timeoutMs;
1051 } keyedMutex;
1052 unsigned int reserved[10];
1053 } params;
1054 unsigned int flags;
1055 unsigned int reserved[16];
1057
1058#if __HIP_HAS_GET_PCH
1063 void __hipGetPCH(const char** pch, unsigned int*size);
1064#endif
1065
1076
1089
1090typedef struct _hipGraphicsResource hipGraphicsResource;
1091
1093
1097typedef struct ihipGraph* hipGraph_t;
1101typedef struct hipGraphNode* hipGraphNode_t;
1105typedef struct hipGraphExec* hipGraphExec_t;
1106
1110typedef struct hipUserObject* hipUserObject_t;
1111
1112
1133
1134typedef void (*hipHostFn_t)(void* userData);
1147typedef struct hipMemsetParams {
1148 void* dst;
1149 unsigned int elementSize;
1150 size_t height;
1151 size_t pitch;
1152 unsigned int value;
1153 size_t width;
1155
1165
1189
1207
1219
1224
1234
1238
1249
1282
1294
1306
1310typedef struct ihipMemGenericAllocationHandle* hipMemGenericAllocationHandle_t;
1311
1319
1326
1334
1342
1346typedef struct hipArrayMapInfo {
1348 union {
1350 hipArray_t array;
1353 union {
1354 struct {
1355 unsigned int level;
1356 unsigned int layer;
1357 unsigned int offsetX;
1358 unsigned int offsetY;
1359 unsigned int offsetZ;
1360 unsigned int extentWidth;
1361 unsigned int extentHeight;
1362 unsigned int extentDepth;
1364 struct {
1365 unsigned int layer;
1366 unsigned long long offset;
1367 unsigned long long size;
1372 union {
1375 unsigned long long offset;
1376 unsigned int deviceBitMask;
1377 unsigned int flags;
1378 unsigned int reserved[2];
1380// Doxygen end group GlobalDefs
1402// TODO-ctx - more description on error codes.
1403hipError_t hipInit(unsigned int flags);
1443
1452hipError_t hipDeviceComputeCapability(int* major, int* minor, hipDevice_t device);
1461hipError_t hipDeviceGetName(char* name, int len, hipDevice_t device);
1484 int srcDevice, int dstDevice);
1493hipError_t hipDeviceGetPCIBusId(char* pciBusId, int len, int device);
1501hipError_t hipDeviceGetByPCIBusId(int* device, const char* pciBusId);
1510// doxygen end initialization
1718hipError_t hipDeviceGetLimit(size_t* pValue, enum hipLimit_t limit);
1737hipError_t hipDeviceSetLimit ( enum hipLimit_t limit, size_t value );
1757hipError_t hipGetDeviceFlags(unsigned int* flags);
1816hipError_t hipExtGetLinkTypeAndHopCount(int device1, int device2, uint32_t* linktype, uint32_t* hopcount);
1817// TODO: implement IPC apis
1884hipError_t hipIpcOpenMemHandle(void** devPtr, hipIpcMemHandle_t handle, unsigned int flags);
1906
1924
1943
1944// end doxygen Device
1968hipError_t hipFuncSetAttribute(const void* func, hipFuncAttribute attr, int value);
1994//doxygen end execution
2036const char* hipGetErrorName(hipError_t hip_error);
2045const char* hipGetErrorString(hipError_t hipError);
2055hipError_t hipDrvGetErrorName(hipError_t hipError, const char** errorString);
2065hipError_t hipDrvGetErrorString(hipError_t hipError, const char** errorString);
2066// end doxygen Error
2131hipError_t hipStreamCreateWithPriority(hipStream_t* stream, unsigned int flags, int priority);
2145hipError_t hipDeviceGetStreamPriorityRange(int* leastPriority, int* greatestPriority);
2219hipError_t hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int flags);
2233hipError_t hipStreamGetFlags(hipStream_t stream, unsigned int* flags);
2278hipError_t hipExtStreamCreateWithCUMask(hipStream_t* stream, uint32_t cuMaskSize, const uint32_t* cuMask);
2291hipError_t hipExtStreamGetCUMask(hipStream_t stream, uint32_t cuMaskSize, uint32_t* cuMask);
2295typedef void (*hipStreamCallback_t)(hipStream_t stream, hipError_t status, void* userData);
2312 unsigned int flags);
2313// end doxygen Stream
2356hipError_t hipStreamWaitValue32(hipStream_t stream, void* ptr, uint32_t value, unsigned int flags,
2357 uint32_t mask __dparm(0xFFFFFFFF));
2390hipError_t hipStreamWaitValue64(hipStream_t stream, void* ptr, uint64_t value, unsigned int flags,
2391 uint64_t mask __dparm(0xFFFFFFFFFFFFFFFF));
2411hipError_t hipStreamWriteValue32(hipStream_t stream, void* ptr, uint32_t value, unsigned int flags);
2431hipError_t hipStreamWriteValue64(hipStream_t stream, void* ptr, uint64_t value, unsigned int flags);
2432// end doxygen Stream Memory Operations
2508#ifdef __cplusplus
2509hipError_t hipEventRecord(hipEvent_t event, hipStream_t stream = NULL);
2510#else
2512#endif
2593// end doxygen Events
2625 hipDeviceptr_t ptr);
2626
2627
2662 hipDeviceptr_t ptr);
2679hipError_t hipDrvPointerGetAttributes(unsigned int numAttributes, hipPointer_attribute* attributes,
2680 void** data, hipDeviceptr_t ptr);
2702 const hipExternalSemaphoreHandleDesc* semHandleDesc);
2716 const hipExternalSemaphoreSignalParams* paramsArray,
2717 unsigned int numExtSems, hipStream_t stream);
2731 const hipExternalSemaphoreWaitParams* paramsArray,
2732 unsigned int numExtSems, hipStream_t stream);
2743
2777// end of external resource
2795hipError_t hipMalloc(void** ptr, size_t size);
2810hipError_t hipExtMallocWithFlags(void** ptr, size_t sizeBytes, unsigned int flags);
2823DEPRECATED("use hipHostMalloc instead")
2824hipError_t hipMallocHost(void** ptr, size_t size);
2837DEPRECATED("use hipHostMalloc instead")
2838hipError_t hipMemAllocHost(void** ptr, size_t size);
2880hipError_t hipHostMalloc(void** ptr, size_t size, unsigned int flags);
2914 size_t size,
2915 unsigned int flags __dparm(hipMemAttachGlobal));
2929 size_t count,
2930 int device,
2931 hipStream_t stream __dparm(0));
2951hipError_t hipMemAdvise(const void* dev_ptr,
2952 size_t count,
2953 hipMemoryAdvise advice,
2954 int device);
2970 size_t data_size,
2971 hipMemRangeAttribute attribute,
2972 const void* dev_ptr,
2973 size_t count);
2991 size_t* data_sizes,
2992 hipMemRangeAttribute* attributes,
2993 size_t num_attributes,
2994 const void* dev_ptr,
2995 size_t count);
3011 void* dev_ptr,
3012 size_t length __dparm(0),
3013 unsigned int flags __dparm(hipMemAttachSingle));
3014// end doxygen Managed Memory
3074hipError_t hipMallocAsync(void** dev_ptr, size_t size, hipStream_t stream);
3099hipError_t hipFreeAsync(void* dev_ptr, hipStream_t stream);
3127hipError_t hipMemPoolTrimTo(hipMemPool_t mem_pool, size_t min_bytes_to_hold);
3223hipError_t hipMemPoolSetAccess(hipMemPool_t mem_pool, const hipMemAccessDesc* desc_list, size_t count);
3328hipError_t hipMallocFromPoolAsync(void** dev_ptr, size_t size, hipMemPool_t mem_pool, hipStream_t stream);
3356 void* shared_handle,
3357 hipMemPool_t mem_pool,
3358 hipMemAllocationHandleType handle_type,
3359 unsigned int flags);
3384 hipMemPool_t* mem_pool,
3385 void* shared_handle,
3386 hipMemAllocationHandleType handle_type,
3387 unsigned int flags);
3437 void** dev_ptr,
3438 hipMemPool_t mem_pool,
3439 hipMemPoolPtrExportData* export_data);
3440// Doxygen end of ordered memory allocator
3458DEPRECATED("use hipHostMalloc instead")
3459hipError_t hipHostAlloc(void** ptr, size_t size, unsigned int flags);
3471hipError_t hipHostGetDevicePointer(void** devPtr, void* hstPtr, unsigned int flags);
3481hipError_t hipHostGetFlags(unsigned int* flagsPtr, void* hostPtr);
3518hipError_t hipHostRegister(void* hostPtr, size_t sizeBytes, unsigned int flags);
3546hipError_t hipMallocPitch(void** ptr, size_t* pitch, size_t width, size_t height);
3568hipError_t hipMemAllocPitch(hipDeviceptr_t* dptr, size_t* pitch, size_t widthInBytes, size_t height, unsigned int elementSizeBytes);
3592DEPRECATED("use hipHostFree instead")
3637hipError_t hipMemcpy(void* dst, const void* src, size_t sizeBytes, hipMemcpyKind kind);
3652hipError_t hipMemcpyWithStream(void* dst, const void* src, size_t sizeBytes,
3653 hipMemcpyKind kind, hipStream_t stream);
3671hipError_t hipMemcpyHtoD(hipDeviceptr_t dst, void* src, size_t sizeBytes);
3689hipError_t hipMemcpyDtoH(void* dst, hipDeviceptr_t src, size_t sizeBytes);
3725hipError_t hipMemcpyHtoDAsync(hipDeviceptr_t dst, void* src, size_t sizeBytes, hipStream_t stream);
3743hipError_t hipMemcpyDtoHAsync(void* dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream);
3762 hipStream_t stream);
3763
3779 hipModule_t hmod, const char* name);
3780
3790hipError_t hipGetSymbolAddress(void** devPtr, const void* symbol);
3791
3801hipError_t hipGetSymbolSize(size_t* size, const void* symbol);
3802
3821hipError_t hipMemcpyToSymbol(const void* symbol, const void* src,
3822 size_t sizeBytes, size_t offset __dparm(0),
3824
3838hipError_t hipMemcpyToSymbolAsync(const void* symbol, const void* src,
3839 size_t sizeBytes, size_t offset,
3840 hipMemcpyKind kind, hipStream_t stream __dparm(0));
3841
3854hipError_t hipMemcpyFromSymbol(void* dst, const void* symbol,
3855 size_t sizeBytes, size_t offset __dparm(0),
3857
3871hipError_t hipMemcpyFromSymbolAsync(void* dst, const void* symbol,
3872 size_t sizeBytes, size_t offset,
3873 hipMemcpyKind kind,
3874 hipStream_t stream __dparm(0));
3903hipError_t hipMemcpyAsync(void* dst, const void* src, size_t sizeBytes, hipMemcpyKind kind,
3904 hipStream_t stream __dparm(0));
3914hipError_t hipMemset(void* dst, int value, size_t sizeBytes);
3924hipError_t hipMemsetD8(hipDeviceptr_t dest, unsigned char value, size_t count);
3940hipError_t hipMemsetD8Async(hipDeviceptr_t dest, unsigned char value, size_t count, hipStream_t stream __dparm(0));
3950hipError_t hipMemsetD16(hipDeviceptr_t dest, unsigned short value, size_t count);
3966hipError_t hipMemsetD16Async(hipDeviceptr_t dest, unsigned short value, size_t count, hipStream_t stream __dparm(0));
3976hipError_t hipMemsetD32(hipDeviceptr_t dest, int value, size_t count);
3992hipError_t hipMemsetAsync(void* dst, int value, size_t sizeBytes, hipStream_t stream __dparm(0));
4008hipError_t hipMemsetD32Async(hipDeviceptr_t dst, int value, size_t count,
4009 hipStream_t stream __dparm(0));
4020hipError_t hipMemset2D(void* dst, size_t pitch, int value, size_t width, size_t height);
4032hipError_t hipMemset2DAsync(void* dst, size_t pitch, int value, size_t width, size_t height,hipStream_t stream __dparm(0));
4041hipError_t hipMemset3D(hipPitchedPtr pitchedDevPtr, int value, hipExtent extent );
4051hipError_t hipMemset3DAsync(hipPitchedPtr pitchedDevPtr, int value, hipExtent extent ,hipStream_t stream __dparm(0));
4068hipError_t hipMemGetInfo(size_t* free, size_t* total);
4069
4081hipError_t hipMemPtrGetInfo(void* ptr, size_t* size);
4094hipError_t hipMallocArray(hipArray** array, const hipChannelFormatDesc* desc, size_t width,
4095 size_t height __dparm(0), unsigned int flags __dparm(hipArrayDefault));
4106hipError_t hipArrayCreate(hipArray** pHandle, const HIP_ARRAY_DESCRIPTOR* pAllocateArray);
4160 struct hipExtent extent, unsigned int flags);
4173hipError_t hipArrayGetInfo(hipChannelFormatDesc* desc, hipExtent* extent, unsigned int* flags,
4174 hipArray* array);
4227hipError_t hipMemcpy2D(void* dst, size_t dpitch, const void* src, size_t spitch, size_t width,
4228 size_t height, hipMemcpyKind kind);
4267hipError_t hipMemcpy2DAsync(void* dst, size_t dpitch, const void* src, size_t spitch, size_t width,
4268 size_t height, hipMemcpyKind kind, hipStream_t stream __dparm(0));
4286hipError_t hipMemcpy2DToArray(hipArray* dst, size_t wOffset, size_t hOffset, const void* src,
4287 size_t spitch, size_t width, size_t height, hipMemcpyKind kind);
4306hipError_t hipMemcpy2DToArrayAsync(hipArray* dst, size_t wOffset, size_t hOffset, const void* src,
4307 size_t spitch, size_t width, size_t height, hipMemcpyKind kind,
4308 hipStream_t stream __dparm(0));
4327hipError_t hipMemcpyToArray(hipArray* dst, size_t wOffset, size_t hOffset, const void* src,
4328 size_t count, hipMemcpyKind kind);
4347hipError_t hipMemcpyFromArray(void* dst, hipArray_const_t srcArray, size_t wOffset, size_t hOffset,
4348 size_t count, hipMemcpyKind kind);
4366hipError_t hipMemcpy2DFromArray( void* dst, size_t dpitch, hipArray_const_t src, size_t wOffset, size_t hOffset, size_t width, size_t height, hipMemcpyKind kind);
4385hipError_t hipMemcpy2DFromArrayAsync( void* dst, size_t dpitch, hipArray_const_t src, size_t wOffset, size_t hOffset, size_t width, size_t height, hipMemcpyKind kind, hipStream_t stream __dparm(0));
4399hipError_t hipMemcpyAtoH(void* dst, hipArray* srcArray, size_t srcOffset, size_t count);
4413hipError_t hipMemcpyHtoA(hipArray* dstArray, size_t dstOffset, const void* srcHost, size_t count);
4460// doxygen end Memory
4488hipError_t hipDeviceCanAccessPeer(int* canAccessPeer, int deviceId, int peerDeviceId);
4505hipError_t hipDeviceEnablePeerAccess(int peerDeviceId, unsigned int flags);
4531#ifndef USE_PEER_NON_UNIFIED
4532#define USE_PEER_NON_UNIFIED 1
4533#endif
4534#if USE_PEER_NON_UNIFIED == 1
4546hipError_t hipMemcpyPeer(void* dst, int dstDeviceId, const void* src, int srcDeviceId,
4547 size_t sizeBytes);
4560hipError_t hipMemcpyPeerAsync(void* dst, int dstDeviceId, const void* src, int srcDevice,
4561 size_t sizeBytes, hipStream_t stream __dparm(0));
4562#endif
4563// doxygen end PeerToPeer
4596hipError_t hipCtxCreate(hipCtx_t* ctx, unsigned int flags, hipDevice_t device);
4798hipError_t hipCtxGetFlags(unsigned int* flags);
4821hipError_t hipCtxEnablePeerAccess(hipCtx_t peerCtx, unsigned int flags);
4842// doxygen end Context deprecated
4858hipError_t hipDevicePrimaryCtxGetState(hipDevice_t dev, unsigned int* flags, int* active);
4908// doxygen end Context Management
4935hipError_t hipModuleLoad(hipModule_t* module, const char* fname);
4956hipError_t hipModuleGetFunction(hipFunction_t* function, hipModule_t module, const char* kname);
4965hipError_t hipFuncGetAttributes(struct hipFuncAttributes* attr, const void* func);
4995hipError_t hipModuleLoadData(hipModule_t* module, const void* image);
5008hipError_t hipModuleLoadDataEx(hipModule_t* module, const void* image, unsigned int numOptions,
5009 hipJitOption* options, void** optionValues);
5038hipError_t hipModuleLaunchKernel(hipFunction_t f, unsigned int gridDimX, unsigned int gridDimY,
5039 unsigned int gridDimZ, unsigned int blockDimX,
5040 unsigned int blockDimY, unsigned int blockDimZ,
5041 unsigned int sharedMemBytes, hipStream_t stream,
5042 void** kernelParams, void** extra);
5069 unsigned int gridDimY, unsigned int gridDimZ,
5070 unsigned int blockDimX, unsigned int blockDimY,
5071 unsigned int blockDimZ, unsigned int sharedMemBytes,
5072 hipStream_t stream, void** kernelParams);
5088 unsigned int numDevices,
5089 unsigned int flags);
5108hipError_t hipLaunchCooperativeKernel(const void* f, dim3 gridDim, dim3 blockDimX,
5109 void** kernelParams, unsigned int sharedMemBytes,
5110 hipStream_t stream);
5122 int numDevices, unsigned int flags);
5135 int numDevices, unsigned int flags);
5136// doxygen end Module
5163//TODO - Match CUoccupancyB2DSize
5165 hipFunction_t f, size_t dynSharedMemPerBlk,
5166 int blockSizeLimit);
5182//TODO - Match CUoccupancyB2DSize
5184 hipFunction_t f, size_t dynSharedMemPerBlk,
5185 int blockSizeLimit, unsigned int flags);
5195 int* numBlocks, hipFunction_t f, int blockSize, size_t dynSharedMemPerBlk);
5206 int* numBlocks, hipFunction_t f, int blockSize, size_t dynSharedMemPerBlk, unsigned int flags);
5216 int* numBlocks, const void* f, int blockSize, size_t dynSharedMemPerBlk);
5227 int* numBlocks, const void* f, int blockSize, size_t dynSharedMemPerBlk, unsigned int flags __dparm(hipOccupancyDefault));
5242hipError_t hipOccupancyMaxPotentialBlockSize(int* gridSize, int* blockSize,
5243 const void* f, size_t dynSharedMemPerBlk,
5244 int blockSizeLimit);
5245// doxygen end Occupancy
5259// TODO - expand descriptions:
5265DEPRECATED("use roctracer/rocTX instead")
5272DEPRECATED("use roctracer/rocTX instead")
5274// doxygen end profiler
5301hipError_t hipConfigureCall(dim3 gridDim, dim3 blockDim, size_t sharedMem __dparm(0), hipStream_t stream __dparm(0));
5312hipError_t hipSetupArgument(const void* arg, size_t size, size_t offset);
5321hipError_t hipLaunchByPtr(const void* func);
5339 dim3 blockDim,
5340 size_t sharedMem __dparm(0),
5341 hipStream_t stream __dparm(0));
5362 dim3 *blockDim,
5363 size_t *sharedMem,
5364 hipStream_t *stream);
5380hipError_t hipLaunchKernel(const void* function_address,
5381 dim3 numBlocks,
5382 dim3 dimBlocks,
5383 void** args,
5384 size_t sharedMemBytes __dparm(0),
5385 hipStream_t stream __dparm(0));
5386
5399
5408//TODO: Move this to hip_ext.h
5429hipError_t hipExtLaunchKernel(const void* function_address, dim3 numBlocks, dim3 dimBlocks,
5430 void** args, size_t sharedMemBytes, hipStream_t stream,
5431 hipEvent_t startEvent, hipEvent_t stopEvent, int flags);
5432// doxygen end Clang launch
5459 hipTextureObject_t* pTexObject,
5460 const hipResourceDesc* pResDesc,
5461 const hipTextureDesc* pTexDesc,
5462 const struct hipResourceViewDesc* pResViewDesc);
5463
5473
5485 hipArray_const_t array);
5486
5497 hipResourceDesc* pResDesc,
5498 hipTextureObject_t textureObject);
5499
5510 struct hipResourceViewDesc* pResViewDesc,
5511 hipTextureObject_t textureObject);
5512
5523 hipTextureDesc* pTexDesc,
5524 hipTextureObject_t textureObject);
5525
5538 hipTextureObject_t* pTexObject,
5539 const HIP_RESOURCE_DESC* pResDesc,
5540 const HIP_TEXTURE_DESC* pTexDesc,
5541 const HIP_RESOURCE_VIEW_DESC* pResViewDesc);
5542
5552 hipTextureObject_t texObject);
5553
5564 HIP_RESOURCE_DESC* pResDesc,
5565 hipTextureObject_t texObject);
5566
5577 HIP_RESOURCE_VIEW_DESC* pResViewDesc,
5578 hipTextureObject_t texObject);
5579
5590 HIP_TEXTURE_DESC* pTexDesc,
5591 hipTextureObject_t texObject);
5592
5608 hipMipmappedArray_t *mipmappedArray,
5609 const struct hipChannelFormatDesc* desc,
5610 struct hipExtent extent,
5611 unsigned int numLevels,
5612 unsigned int flags __dparm(0));
5613
5624hipError_t hipFreeMipmappedArray(hipMipmappedArray_t mipmappedArray);
5625
5639 hipArray_t *levelArray,
5640 hipMipmappedArray_const_t mipmappedArray,
5641 unsigned int level);
5642
5655 hipMipmappedArray_t* pHandle,
5656 HIP_ARRAY3D_DESCRIPTOR* pMipmappedArrayDesc,
5657 unsigned int numMipmapLevels);
5658
5669hipError_t hipMipmappedArrayDestroy(hipMipmappedArray_t hMipmappedArray);
5670
5684 hipArray_t* pLevelArray,
5685 hipMipmappedArray_t hMipMappedArray,
5686 unsigned int level);
5687
5708 const textureReference* tex,
5709 hipMipmappedArray_const_t mipmappedArray,
5710 const hipChannelFormatDesc* desc);
5711
5724 const textureReference** texref,
5725 const void* symbol);
5739 textureReference* texRef,
5740 int dim,
5741 enum hipTextureAddressMode am);
5756 textureReference* tex,
5757 hipArray_const_t array,
5758 unsigned int flags);
5772 textureReference* texRef,
5773 enum hipTextureFilterMode fm);
5787 textureReference* texRef,
5788 unsigned int Flags);
5803 textureReference* texRef,
5804 hipArray_Format fmt,
5805 int NumPackedComponents);
5820 size_t* offset,
5821 const textureReference* tex,
5822 const void* devPtr,
5823 const hipChannelFormatDesc* desc,
5824 size_t size __dparm(UINT_MAX));
5841 size_t* offset,
5842 const textureReference* tex,
5843 const void* devPtr,
5844 const hipChannelFormatDesc* desc,
5845 size_t width,
5846 size_t height,
5847 size_t pitch);
5860 const textureReference* tex,
5861 hipArray_const_t array,
5862 const hipChannelFormatDesc* desc);
5874 size_t* offset,
5875 const textureReference* texref);
5897 hipDeviceptr_t* dev_ptr,
5898 const textureReference* texRef);
5911 enum hipTextureAddressMode* pam,
5912 const textureReference* texRef,
5913 int dim);
5925 enum hipTextureFilterMode* pfm,
5926 const textureReference* texRef);
5938 unsigned int* pFlags,
5939 const textureReference* texRef);
5952 hipArray_Format* pFormat,
5953 int* pNumChannels,
5954 const textureReference* texRef);
5966 int* pmaxAnsio,
5967 const textureReference* texRef);
5979 enum hipTextureFilterMode* pfm,
5980 const textureReference* texRef);
5992 float* pbias,
5993 const textureReference* texRef);
6006 float* pminMipmapLevelClamp,
6007 float* pmaxMipmapLevelClamp,
6008 const textureReference* texRef);
6020 hipMipmappedArray_t* pArray,
6021 const textureReference* texRef);
6035 size_t* ByteOffset,
6036 textureReference* texRef,
6037 hipDeviceptr_t dptr,
6038 size_t bytes);
6052 textureReference* texRef,
6053 const HIP_ARRAY_DESCRIPTOR* desc,
6054 hipDeviceptr_t dptr,
6055 size_t Pitch);
6067 textureReference* texRef,
6068 unsigned int maxAniso);
6080 textureReference* texRef,
6081 float* pBorderColor);
6093 textureReference* texRef,
6094 enum hipTextureFilterMode fm);
6106 textureReference* texRef,
6107 float bias);
6120 textureReference* texRef,
6121 float minMipMapLevelClamp,
6122 float maxMipMapLevelClamp);
6135 textureReference* texRef,
6136 struct hipMipmappedArray* mipmappedArray,
6137 unsigned int Flags);
6138
6139// doxygen end deprecated texture management
6144// doxygen end Texture management
6156// This group is for HIPrtc
6157
6158// doxygen end Runtime
6177const char* hipApiName(uint32_t id);
6186const char* hipKernelNameRef(const hipFunction_t f);
6196const char* hipKernelNameRefByPtr(const void* hostFunction, hipStream_t stream);
6206
6207// doxygen end Callback
6233
6247
6262 unsigned long long* pId);
6263
6281 unsigned long long* id_out __dparm(0),
6282 hipGraph_t* graph_out __dparm(0),
6283 const hipGraphNode_t** dependencies_out __dparm(0),
6284 size_t* numDependencies_out __dparm(0));
6285
6299
6313 size_t numDependencies,
6314 unsigned int flags __dparm(0));
6315
6327
6340hipError_t hipGraphCreate(hipGraph_t* pGraph, unsigned int flags);
6341
6354
6369 const hipGraphNode_t* to, size_t numDependencies);
6370
6385 const hipGraphNode_t* to, size_t numDependencies);
6386
6405 size_t* numEdges);
6406
6423hipError_t hipGraphGetNodes(hipGraph_t graph, hipGraphNode_t* nodes, size_t* numNodes);
6424
6442 size_t* pNumRootNodes);
6443
6461 size_t* pNumDependencies);
6462
6481 size_t* pNumDependentNodes);
6482
6495
6507
6519hipError_t hipGraphClone(hipGraph_t* pGraphClone, hipGraph_t originalGraph);
6520
6534 hipGraph_t clonedGraph);
6535
6553 hipGraphNode_t* pErrorNode, char* pLogBuffer, size_t bufferSize);
6554
6568 unsigned long long flags);
6569
6581
6593
6605
6606// Check whether an executable graph can be updated with a graph and perform the update if possible.
6621 hipGraphNode_t* hErrorNode_out,
6622 hipGraphExecUpdateResult* updateResult_out);
6623
6637 const hipGraphNode_t* pDependencies, size_t numDependencies,
6638 const hipKernelNodeParams* pNodeParams);
6639
6650
6661
6673 const hipKernelNodeParams* pNodeParams);
6674
6688 const hipGraphNode_t* pDependencies, size_t numDependencies,
6689 const hipMemcpy3DParms* pCopyParams);
6700
6711
6723 const hipKernelNodeAttrValue* value);
6735 hipKernelNodeAttrValue* value);
6747 hipMemcpy3DParms* pNodeParams);
6748
6765 const hipGraphNode_t* pDependencies, size_t numDependencies,
6766 void* dst, const void* src, size_t count, hipMemcpyKind kind);
6767
6781 size_t count, hipMemcpyKind kind);
6782
6798 void* dst, const void* src, size_t count,
6799 hipMemcpyKind kind);
6800
6818 const hipGraphNode_t* pDependencies,
6819 size_t numDependencies, void* dst, const void* symbol,
6820 size_t count, size_t offset, hipMemcpyKind kind);
6821
6836 size_t count, size_t offset, hipMemcpyKind kind);
6837
6854 void* dst, const void* symbol, size_t count,
6855 size_t offset, hipMemcpyKind kind);
6856
6874 const hipGraphNode_t* pDependencies,
6875 size_t numDependencies, const void* symbol,
6876 const void* src, size_t count, size_t offset,
6877 hipMemcpyKind kind);
6878
6893 const void* src, size_t count, size_t offset,
6894 hipMemcpyKind kind);
6895
6896
6912 const void* symbol, const void* src,
6913 size_t count, size_t offset, hipMemcpyKind kind);
6914
6928 const hipGraphNode_t* pDependencies, size_t numDependencies,
6929 const hipMemsetParams* pMemsetParams);
6930
6941
6952
6964 const hipMemsetParams* pNodeParams);
6965
6979 const hipGraphNode_t* pDependencies, size_t numDependencies,
6980 const hipHostNodeParams* pNodeParams);
6981
6992
7003
7015 const hipHostNodeParams* pNodeParams);
7016
7030 const hipGraphNode_t* pDependencies, size_t numDependencies,
7031 hipGraph_t childGraph);
7032
7043
7055 hipGraph_t childGraph);
7056
7069 const hipGraphNode_t* pDependencies, size_t numDependencies);
7070
7071
7085 const hipGraphNode_t* pDependencies, size_t numDependencies,
7086 hipEvent_t event);
7087
7098
7109
7121 hipEvent_t event);
7122
7136 const hipGraphNode_t* pDependencies, size_t numDependencies,
7137 hipEvent_t event);
7138
7139
7150
7161
7173 hipEvent_t event);
7174
7188 const hipGraphNode_t* pDependencies, size_t numDependencies, hipMemAllocNodeParams* pNodeParams);
7189
7200
7214 const hipGraphNode_t* pDependencies, size_t numDependencies, void* dev_ptr);
7215
7226
7238
7250
7259
7273 unsigned int initialRefcount, unsigned int flags);
7274
7285
7296
7309 unsigned int count __dparm(1), unsigned int flags __dparm(0));
7310
7322 unsigned int count __dparm(1));
7323
7334hipError_t hipGraphDebugDotPrint(hipGraph_t graph, const char* path, unsigned int flags);
7335
7351
7375 unsigned int isEnabled);
7397 unsigned int* isEnabled);
7398
7412 const hipGraphNode_t* pDependencies, size_t numDependencies,
7413 const hipExternalSemaphoreWaitNodeParams* nodeParams);
7414
7428 const hipGraphNode_t* pDependencies, size_t numDependencies,
7429 const hipExternalSemaphoreSignalNodeParams* nodeParams);
7440 const hipExternalSemaphoreSignalNodeParams* nodeParams);
7451 const hipExternalSemaphoreWaitNodeParams* nodeParams);
7462 const hipExternalSemaphoreSignalNodeParams* params_out);
7473 const hipExternalSemaphoreWaitNodeParams* params_out);
7485 const hipExternalSemaphoreSignalNodeParams* nodeParams);
7497 const hipExternalSemaphoreWaitNodeParams* nodeParams);
7498// doxygen end graph API
7526hipError_t hipMemAddressFree(void* devPtr, size_t size);
7527
7542hipError_t hipMemAddressReserve(void** ptr, size_t size, size_t alignment, void* addr, unsigned long long flags);
7543
7557hipError_t hipMemCreate(hipMemGenericAllocationHandle_t* handle, size_t size, const hipMemAllocationProp* prop, unsigned long long flags);
7558
7572hipError_t hipMemExportToShareableHandle(void* shareableHandle, hipMemGenericAllocationHandle_t handle, hipMemAllocationHandleType handleType, unsigned long long flags);
7573
7586hipError_t hipMemGetAccess(unsigned long long* flags, const hipMemLocation* location, void* ptr);
7587
7602
7615
7629
7644hipError_t hipMemMap(void* ptr, size_t size, size_t offset, hipMemGenericAllocationHandle_t handle, unsigned long long flags);
7645
7658hipError_t hipMemMapArrayAsync(hipArrayMapInfo* mapInfoList, unsigned int count, hipStream_t stream);
7659
7671
7684
7698hipError_t hipMemSetAccess(void* ptr, size_t size, const hipMemAccessDesc* desc, size_t count);
7699
7711hipError_t hipMemUnmap(void* ptr, size_t size);
7712
7713// doxygen end virtual memory management API
7726typedef unsigned int GLuint;
7728typedef unsigned int GLenum;
7729
7743hipError_t hipGLGetDevices(unsigned int* pHipDeviceCount, int* pHipDevices,
7744 unsigned int hipDeviceCount, hipGLDeviceList deviceList);
7756 unsigned int flags);
7769 GLenum target, unsigned int flags);
7781 hipStream_t stream __dparm(0) );
7796 unsigned int arrayIndex, unsigned int mipLevel);
7808 hipGraphicsResource_t resource);
7820 hipStream_t stream __dparm(0));
7830// doxygen end GL Interop
7865// end of surface
7869#ifdef __cplusplus
7870} /* extern "c" */
7871#endif
7872#ifdef __cplusplus
7873#if defined(__clang__) && defined(__HIP__)
7874template <typename T>
7875static hipError_t __host__ inline hipOccupancyMaxPotentialBlockSize(int* gridSize, int* blockSize,
7876 T f, size_t dynSharedMemPerBlk = 0, int blockSizeLimit = 0) {
7877 return hipOccupancyMaxPotentialBlockSize(gridSize, blockSize, reinterpret_cast<const void*>(f),dynSharedMemPerBlk,blockSizeLimit);
7878}
7879template <typename T>
7880static hipError_t __host__ inline hipOccupancyMaxPotentialBlockSizeWithFlags(int* gridSize, int* blockSize,
7881 T f, size_t dynSharedMemPerBlk = 0, int blockSizeLimit = 0, unsigned int flags = 0 ) {
7882 return hipOccupancyMaxPotentialBlockSize(gridSize, blockSize, reinterpret_cast<const void*>(f),dynSharedMemPerBlk,blockSizeLimit);
7883}
7884#endif // defined(__clang__) && defined(__HIP__)
7885
7895template <typename T>
7896hipError_t hipGetSymbolAddress(void** devPtr, const T &symbol) {
7897 return ::hipGetSymbolAddress(devPtr, (const void *)&symbol);
7898}
7909template <typename T>
7910hipError_t hipGetSymbolSize(size_t* size, const T &symbol) {
7911 return ::hipGetSymbolSize(size, (const void *)&symbol);
7912}
7913
7922template <typename T>
7923hipError_t hipMemcpyToSymbol(const T& symbol, const void* src, size_t sizeBytes,
7924 size_t offset __dparm(0),
7926 return ::hipMemcpyToSymbol((const void*)&symbol, src, sizeBytes, offset, kind);
7927}
7936template <typename T>
7937hipError_t hipMemcpyToSymbolAsync(const T& symbol, const void* src, size_t sizeBytes, size_t offset,
7938 hipMemcpyKind kind, hipStream_t stream __dparm(0)) {
7939 return ::hipMemcpyToSymbolAsync((const void*)&symbol, src, sizeBytes, offset, kind, stream);
7940}
7948template <typename T>
7949hipError_t hipMemcpyFromSymbol(void* dst, const T &symbol,
7950 size_t sizeBytes, size_t offset __dparm(0),
7952 return ::hipMemcpyFromSymbol(dst, (const void*)&symbol, sizeBytes, offset, kind);
7953}
7961template <typename T>
7962hipError_t hipMemcpyFromSymbolAsync(void* dst, const T& symbol, size_t sizeBytes, size_t offset,
7963 hipMemcpyKind kind, hipStream_t stream __dparm(0)) {
7964 return ::hipMemcpyFromSymbolAsync(dst, (const void*)&symbol, sizeBytes, offset, kind, stream);
7965}
7966
7978template <class T>
7980 int* numBlocks, T f, int blockSize, size_t dynSharedMemPerBlk) {
7982 numBlocks, reinterpret_cast<const void*>(f), blockSize, dynSharedMemPerBlk);
7983}
7997template <class T>
7999 int* numBlocks, T f, int blockSize, size_t dynSharedMemPerBlk, unsigned int flags) {
8001 numBlocks, reinterpret_cast<const void*>(f), blockSize, dynSharedMemPerBlk, flags);
8002}
8023template<typename UnaryFunction, class T>
8024static hipError_t __host__ inline hipOccupancyMaxPotentialBlockSizeVariableSMemWithFlags(
8025 int* min_grid_size,
8026 int* block_size,
8027 T func,
8028 UnaryFunction block_size_to_dynamic_smem_size,
8029 int block_size_limit = 0,
8030 unsigned int flags = 0) {
8031 if (min_grid_size == nullptr || block_size == nullptr ||
8032 reinterpret_cast<const void*>(func) == nullptr) {
8033 return hipErrorInvalidValue;
8034 }
8035
8036 int dev;
8037 hipError_t status;
8038 if ((status = hipGetDevice(&dev)) != hipSuccess) {
8039 return status;
8040 }
8041
8042 int max_threads_per_cu;
8043 if ((status = hipDeviceGetAttribute(&max_threads_per_cu,
8045 return status;
8046 }
8047
8048 int warp_size;
8049 if ((status = hipDeviceGetAttribute(&warp_size,
8051 return status;
8052 }
8053
8054 int max_cu_count;
8055 if ((status = hipDeviceGetAttribute(&max_cu_count,
8057 return status;
8058 }
8059
8060 struct hipFuncAttributes attr;
8061 if ((status = hipFuncGetAttributes(&attr, reinterpret_cast<const void*>(func))) != hipSuccess) {
8062 return status;
8063 }
8064
8065 // Initial limits for the execution
8066 const int func_max_threads_per_block = attr.maxThreadsPerBlock;
8067 if (block_size_limit == 0) {
8068 block_size_limit = func_max_threads_per_block;
8069 }
8070
8071 if (func_max_threads_per_block < block_size_limit) {
8072 block_size_limit = func_max_threads_per_block;
8073 }
8074
8075 const int block_size_limit_aligned =
8076 ((block_size_limit + (warp_size - 1)) / warp_size) * warp_size;
8077
8078 // For maximum search
8079 int max_threads = 0;
8080 int max_block_size{};
8081 int max_num_blocks{};
8082 for (int block_size_check_aligned = block_size_limit_aligned;
8083 block_size_check_aligned > 0;
8084 block_size_check_aligned -= warp_size) {
8085 // Make sure the logic uses the requested limit and not aligned
8086 int block_size_check = (block_size_limit < block_size_check_aligned) ?
8087 block_size_limit : block_size_check_aligned;
8088
8089 size_t dyn_smem_size = block_size_to_dynamic_smem_size(block_size_check);
8090 int optimal_blocks;
8092 &optimal_blocks, func, block_size_check, dyn_smem_size, flags)) != hipSuccess) {
8093 return status;
8094 }
8095
8096 int total_threads = block_size_check * optimal_blocks;
8097 if (total_threads > max_threads) {
8098 max_block_size = block_size_check;
8099 max_num_blocks = optimal_blocks;
8100 max_threads = total_threads;
8101 }
8102
8103 // Break if the logic reached possible maximum
8104 if (max_threads_per_cu == max_threads) {
8105 break;
8106 }
8107 }
8108
8109 // Grid size is the number of blocks per CU * CU count
8110 *min_grid_size = max_num_blocks * max_cu_count;
8111 *block_size = max_block_size;
8112
8113 return status;
8114}
8115
8135template<typename UnaryFunction, class T>
8136static hipError_t __host__ inline hipOccupancyMaxPotentialBlockSizeVariableSMem(
8137 int* min_grid_size,
8138 int* block_size,
8139 T func,
8140 UnaryFunction block_size_to_dynamic_smem_size,
8141 int block_size_limit = 0)
8142{
8143 return hipOccupancyMaxPotentialBlockSizeVariableSMemWithFlags(min_grid_size, block_size, func,
8144 block_size_to_dynamic_smem_size, block_size_limit);
8145}
8160template <typename F>
8161inline hipError_t hipOccupancyMaxPotentialBlockSize(int* gridSize, int* blockSize,
8162 F kernel, size_t dynSharedMemPerBlk, uint32_t blockSizeLimit) {
8163return hipOccupancyMaxPotentialBlockSize(gridSize, blockSize,(hipFunction_t)kernel, dynSharedMemPerBlk, blockSizeLimit);
8164}
8181template <class T>
8182inline hipError_t hipLaunchCooperativeKernel(T f, dim3 gridDim, dim3 blockDim,
8183 void** kernelParams, unsigned int sharedMemBytes, hipStream_t stream) {
8184 return hipLaunchCooperativeKernel(reinterpret_cast<const void*>(f), gridDim,
8185 blockDim, kernelParams, sharedMemBytes, stream);
8186}
8201template <class T>
8203 unsigned int numDevices, unsigned int flags = 0) {
8204 return hipLaunchCooperativeKernelMultiDevice(launchParamsList, numDevices, flags);
8205}
8220template <class T>
8222 unsigned int numDevices, unsigned int flags = 0) {
8223 return hipExtLaunchMultiKernelMultiDevice(launchParamsList, numDevices, flags);
8224}
8236template <class T, int dim, enum hipTextureReadMode readMode>
8238static inline hipError_t hipBindTexture(size_t* offset, const struct texture<T, dim, readMode>& tex,
8239 const void* devPtr, size_t size = UINT_MAX) {
8240 return hipBindTexture(offset, &tex, devPtr, &tex.channelDesc, size);
8241}
8254template <class T, int dim, enum hipTextureReadMode readMode>
8256static inline hipError_t
8257 hipBindTexture(size_t* offset, const struct texture<T, dim, readMode>& tex, const void* devPtr,
8258 const struct hipChannelFormatDesc& desc, size_t size = UINT_MAX) {
8259 return hipBindTexture(offset, &tex, devPtr, &desc, size);
8260}
8274template<class T, int dim, enum hipTextureReadMode readMode>
8276static inline hipError_t hipBindTexture2D(
8277 size_t *offset,
8278 const struct texture<T, dim, readMode> &tex,
8279 const void *devPtr,
8280 size_t width,
8281 size_t height,
8282 size_t pitch)
8283{
8284 return hipBindTexture2D(offset, &tex, devPtr, &tex.channelDesc, width, height, pitch);
8285}
8300template<class T, int dim, enum hipTextureReadMode readMode>
8302static inline hipError_t hipBindTexture2D(
8303 size_t *offset,
8304 const struct texture<T, dim, readMode> &tex,
8305 const void *devPtr,
8306 const struct hipChannelFormatDesc &desc,
8307 size_t width,
8308 size_t height,
8309 size_t pitch)
8310{
8311 return hipBindTexture2D(offset, &tex, devPtr, &desc, width, height, pitch);
8312}
8313
8323template<class T, int dim, enum hipTextureReadMode readMode>
8325static inline hipError_t hipBindTextureToArray(
8326 const struct texture<T, dim, readMode> &tex,
8327 hipArray_const_t array)
8328{
8329 struct hipChannelFormatDesc desc;
8330 hipError_t err = hipGetChannelDesc(&desc, array);
8331 return (err == hipSuccess) ? hipBindTextureToArray(&tex, array, &desc) : err;
8332}
8343template<class T, int dim, enum hipTextureReadMode readMode>
8345static inline hipError_t hipBindTextureToArray(
8346 const struct texture<T, dim, readMode> &tex,
8347 hipArray_const_t array,
8348 const struct hipChannelFormatDesc &desc)
8349{
8350 return hipBindTextureToArray(&tex, array, &desc);
8351}
8361template<class T, int dim, enum hipTextureReadMode readMode>
8364 const struct texture<T, dim, readMode> &tex,
8365 hipMipmappedArray_const_t mipmappedArray)
8366{
8367 struct hipChannelFormatDesc desc;
8368 hipArray_t levelArray;
8369 hipError_t err = hipGetMipmappedArrayLevel(&levelArray, mipmappedArray, 0);
8370 if (err != hipSuccess) {
8371 return err;
8372 }
8373 err = hipGetChannelDesc(&desc, levelArray);
8374 return (err == hipSuccess) ? hipBindTextureToMipmappedArray(&tex, mipmappedArray, &desc) : err;
8375}
8386template<class T, int dim, enum hipTextureReadMode readMode>
8389 const struct texture<T, dim, readMode> &tex,
8390 hipMipmappedArray_const_t mipmappedArray,
8391 const struct hipChannelFormatDesc &desc)
8392{
8393 return hipBindTextureToMipmappedArray(&tex, mipmappedArray, &desc);
8394}
8403template<class T, int dim, enum hipTextureReadMode readMode>
8405static inline hipError_t hipUnbindTexture(
8406 const struct texture<T, dim, readMode> &tex)
8407{
8408 return hipUnbindTexture(&tex);
8409}
8433static inline hipError_t hipMallocAsync(
8434 void** dev_ptr,
8435 size_t size,
8436 hipMemPool_t mem_pool,
8437 hipStream_t stream) {
8438 return hipMallocFromPoolAsync(dev_ptr, size, mem_pool, stream);
8439}
8450template<class T>
8451static inline hipError_t hipMallocAsync(
8452 T** dev_ptr,
8453 size_t size,
8454 hipMemPool_t mem_pool,
8455 hipStream_t stream) {
8456 return hipMallocFromPoolAsync(reinterpret_cast<void**>(dev_ptr), size, mem_pool, stream);
8457}
8468template<class T>
8469static inline hipError_t hipMallocAsync(
8470 T** dev_ptr,
8471 size_t size,
8472 hipStream_t stream) {
8473 return hipMallocAsync(reinterpret_cast<void**>(dev_ptr), size, stream);
8474}
8485template<class T>
8487 T** dev_ptr,
8488 size_t size,
8489 hipMemPool_t mem_pool,
8490 hipStream_t stream) {
8491 return hipMallocFromPoolAsync(reinterpret_cast<void**>(dev_ptr), size, mem_pool, stream);
8492}
8496#endif // __cplusplus
8497
8498#ifdef __GNUC__
8499#pragma GCC visibility pop
8500#endif
8501
8502
8503#elif !(defined(__HIP_PLATFORM_HCC__) || defined(__HIP_PLATFORM_AMD__)) && (defined(__HIP_PLATFORM_NVCC__) || defined(__HIP_PLATFORM_NVIDIA__))
8504#include "hip/nvidia_detail/nvidia_hip_runtime_api.h"
8505#else
8506#error("Must define exactly one of __HIP_PLATFORM_AMD__ or __HIP_PLATFORM_NVIDIA__");
8507#endif
8508
8509
8521#if defined(__cplusplus) && !defined(__HIP_DISABLE_CPP_FUNCTIONS__)
8522template <class T>
8523static inline hipError_t hipMalloc(T** devPtr, size_t size) {
8524 return hipMalloc((void**)devPtr, size);
8525}
8538template <class T>
8539static inline hipError_t hipHostMalloc(T** ptr, size_t size,
8540 unsigned int flags = hipHostMallocDefault) {
8541 return hipHostMalloc((void**)ptr, size, flags);
8542}
8557template <class T>
8558static inline hipError_t hipMallocManaged(T** devPtr, size_t size,
8559 unsigned int flags = hipMemAttachGlobal) {
8560 return hipMallocManaged((void**)devPtr, size, flags);
8561}
8562
8563#endif
8564#endif
8565// doxygen end HIP API
8569#include <hip/amd_detail/amd_hip_runtime_pt_api.h>
8570
8571#if USE_PROF_API
8572#include <hip/amd_detail/hip_prof_str.h>
8573#endif
hipMemcpyKind
Definition driver_types.h:349
@ hipMemcpyDeviceToHost
Device-to-Host Copy.
Definition driver_types.h:352
@ hipMemcpyHostToDevice
Host-to-Device Copy.
Definition driver_types.h:351
hipPointer_attribute
Definition driver_types.h:446
void * hipDeviceptr_t
Definition driver_types.h:42
hipFunction_attribute
Definition driver_types.h:432
hipArray_Format
Definition driver_types.h:60
hipResourceType
Definition driver_types.h:139
const char * hipApiName(uint32_t id)
Returns HIP API name by ID.
const char * hipKernelNameRef(const hipFunction_t f)
Returns kernel name reference by function name.
const char * hipKernelNameRefByPtr(const void *hostFunction, hipStream_t stream)
Retrives kernel for a given host pointer, unless stated otherwise.
int hipGetStreamDeviceId(hipStream_t stream)
Returns device ID on the stream.
hipError_t hipSetupArgument(const void *arg, size_t size, size_t offset)
Set a kernel argument.
hipError_t __hipPopCallConfiguration(dim3 *gridDim, dim3 *blockDim, size_t *sharedMem, hipStream_t *stream)
Pop configuration of a kernel launch.
hipError_t __hipPushCallConfiguration(dim3 gridDim, dim3 blockDim, size_t sharedMem __dparm(0), hipStream_t stream __dparm(0))
Push configuration of a kernel launch.
hipError_t hipLaunchHostFunc(hipStream_t stream, hipHostFn_t fn, void *userData)
Enqueues a host function call in a stream.
hipError_t hipExtLaunchKernel(const void *function_address, dim3 numBlocks, dim3 dimBlocks, void **args, size_t sharedMemBytes, hipStream_t stream, hipEvent_t startEvent, hipEvent_t stopEvent, int flags)
Launches kernel from the pointer address, with arguments and shared memory on stream.
hipError_t hipConfigureCall(dim3 gridDim, dim3 blockDim, size_t sharedMem __dparm(0), hipStream_t stream __dparm(0))
Configure a kernel launch.
hipError_t hipLaunchByPtr(const void *func)
Launch a kernel.
hipError_t hipLaunchKernel(const void *function_address, dim3 numBlocks, dim3 dimBlocks, void **args, size_t sharedMemBytes __dparm(0), hipStream_t stream __dparm(0))
C compliant kernel launch API.
hipError_t hipDrvMemcpy2DUnaligned(const hip_Memcpy2D *pCopy)
hipError_t hipCtxGetFlags(unsigned int *flags)
Return flags used for creating default context.
hipError_t hipCtxPopCurrent(hipCtx_t *ctx)
Pop the current/default context and return the popped context.
hipError_t hipCtxGetSharedMemConfig(hipSharedMemConfig *pConfig)
Get Shared memory bank configuration.
hipError_t hipCtxGetCurrent(hipCtx_t *ctx)
Get the handle of the current/ default context.
hipError_t hipCtxSetCacheConfig(hipFuncCache_t cacheConfig)
Set L1/Shared cache partition.
hipError_t hipCtxSetCurrent(hipCtx_t ctx)
Set the passed context as current/default.
hipError_t hipCtxEnablePeerAccess(hipCtx_t peerCtx, unsigned int flags)
Enables direct access to memory allocations in a peer context.
hipError_t hipCtxGetDevice(hipDevice_t *device)
Get the handle of the device associated with current/default context.
hipError_t hipCtxDestroy(hipCtx_t ctx)
Destroy a HIP context.
hipError_t hipCtxPushCurrent(hipCtx_t ctx)
Push the context to be set as current/ default context.
hipError_t hipCtxGetCacheConfig(hipFuncCache_t *cacheConfig)
Get Cache configuration for a specific function.
hipError_t hipCtxCreate(hipCtx_t *ctx, unsigned int flags, hipDevice_t device)
Create a context and set it as current/default context.
hipError_t hipCtxGetApiVersion(hipCtx_t ctx, int *apiVersion)
Returns the approximate HIP api version.
hipError_t hipCtxSynchronize(void)
Blocks until the default context has completed all preceding requested tasks.
hipError_t hipCtxSetSharedMemConfig(hipSharedMemConfig config)
Set Shared memory bank configuration.
hipError_t hipCtxDisablePeerAccess(hipCtx_t peerCtx)
Disable direct access from current context's virtual address space to memory allocations physically l...
hipError_t hipDevicePrimaryCtxSetFlags(hipDevice_t dev, unsigned int flags)
Set flags for the primary context.
hipError_t hipDevicePrimaryCtxGetState(hipDevice_t dev, unsigned int *flags, int *active)
Get the state of the primary context.
hipError_t hipDevicePrimaryCtxRelease(hipDevice_t dev)
Release the primary context on the GPU.
hipError_t hipDevicePrimaryCtxRetain(hipCtx_t *pctx, hipDevice_t dev)
Retain the primary context on the GPU.
hipError_t hipDevicePrimaryCtxReset(hipDevice_t dev)
Resets the primary context on the GPU.
hipError_t hipGetDeviceFlags(unsigned int *flags)
Gets the flags set for current device.
hipError_t hipIpcGetEventHandle(hipIpcEventHandle_t *handle, hipEvent_t event)
Gets an opaque interprocess handle for an event.
hipError_t hipDeviceGetDefaultMemPool(hipMemPool_t *mem_pool, int device)
Returns the default memory pool of the specified device.
hipError_t hipDeviceGetSharedMemConfig(hipSharedMemConfig *pConfig)
Returns bank width of shared memory for current device.
hipError_t hipDeviceSetMemPool(int device, hipMemPool_t mem_pool)
Sets the current memory pool of a device.
hipError_t hipIpcOpenMemHandle(void **devPtr, hipIpcMemHandle_t handle, unsigned int flags)
Opens an interprocess memory handle exported from another process and returns a device pointer usable...
hipError_t hipGetDeviceProperties(hipDeviceProp_t *prop, int deviceId)
Returns device properties.
hipError_t hipDeviceGetCacheConfig(hipFuncCache_t *cacheConfig)
Get Cache configuration for a specific Device.
hipError_t hipSetDevice(int deviceId)
Set default device to be used for subsequent hip API calls from this thread.
hipError_t hipExtGetLinkTypeAndHopCount(int device1, int device2, uint32_t *linktype, uint32_t *hopcount)
Returns the link type and hop count between two devices.
hipError_t hipSetDeviceFlags(unsigned flags)
The current device behavior is changed according the flags passed.
hipError_t hipDeviceGetAttribute(int *pi, hipDeviceAttribute_t attr, int deviceId)
Query for a specific device attribute.
hipError_t hipGetDevice(int *deviceId)
Return the default device id for the calling host thread.
hipError_t hipGetDeviceCount(int *count)
Return number of compute-capable devices.
hipError_t hipDeviceGetMemPool(hipMemPool_t *mem_pool, int device)
Gets the current memory pool for the specified device.
hipError_t hipDeviceReset(void)
The state of current device is discarded and updated to a fresh state.
hipError_t hipDeviceGetLimit(size_t *pValue, enum hipLimit_t limit)
Gets resource limits of current device.
hipError_t hipDeviceSetSharedMemConfig(hipSharedMemConfig config)
The bank width of shared memory on current device is set.
hipError_t hipDeviceSetLimit(enum hipLimit_t limit, size_t value)
Sets resource limits of current device.
hipError_t hipDeviceSetCacheConfig(hipFuncCache_t cacheConfig)
Set L1/Shared cache partition.
hipError_t hipIpcCloseMemHandle(void *devPtr)
Close memory mapped with hipIpcOpenMemHandle.
hipError_t hipIpcOpenEventHandle(hipEvent_t *event, hipIpcEventHandle_t handle)
Opens an interprocess event handles.
hipError_t hipDeviceSynchronize(void)
Waits on all active streams on current device.
hipError_t hipChooseDevice(int *device, const hipDeviceProp_t *prop)
Device which matches hipDeviceProp_t is returned.
hipError_t hipIpcGetMemHandle(hipIpcMemHandle_t *handle, void *devPtr)
Gets an interprocess memory handle for an existing device memory allocation.
hipError_t hipInit(unsigned int flags)
Explicitly initializes the HIP runtime.
hipError_t hipDeviceComputeCapability(int *major, int *minor, hipDevice_t device)
Returns the compute capability of the device.
hipError_t hipDeviceGetName(char *name, int len, hipDevice_t device)
Returns an identifer string for the device.
hipError_t hipDeviceGetByPCIBusId(int *device, const char *pciBusId)
Returns a handle to a compute device.
hipError_t hipDeviceGetUuid(hipUUID *uuid, hipDevice_t device)
Returns an UUID for the device.[BETA].
hipError_t hipDeviceTotalMem(size_t *bytes, hipDevice_t device)
Returns the total amount of memory on the device.
hipError_t hipDeviceGetPCIBusId(char *pciBusId, int len, int device)
Returns a PCI Bus Id string for the device, overloaded to take int device ID.
hipError_t hipDeviceGet(hipDevice_t *device, int ordinal)
Returns a handle to a compute device.
hipError_t hipRuntimeGetVersion(int *runtimeVersion)
Returns the approximate HIP Runtime version.
hipError_t hipDeviceGetP2PAttribute(int *value, hipDeviceP2PAttr attr, int srcDevice, int dstDevice)
Returns a value for attribute of link between two devices.
hipError_t hipDriverGetVersion(int *driverVersion)
Returns the approximate HIP driver version.
const char * hipGetErrorString(hipError_t hipError)
Return handy text string message to explain the error which occurred.
hipError_t hipPeekAtLastError(void)
Return last error returned by any HIP runtime API call.
hipError_t hipDrvGetErrorString(hipError_t hipError, const char **errorString)
Return handy text string message to explain the error which occurred.
hipError_t hipGetLastError(void)
Return last error returned by any HIP runtime API call and resets the stored error code to hipSuccess...
hipError_t hipDrvGetErrorName(hipError_t hipError, const char **errorString)
Return hip error as text string form.
const char * hipGetErrorName(hipError_t hip_error)
Return hip error as text string form.
hipError_t hipEventSynchronize(hipEvent_t event)
Wait for an event to complete.
hipError_t hipEventRecord(hipEvent_t event, hipStream_t stream)
Record an event in the specified stream.
hipError_t hipEventQuery(hipEvent_t event)
Query event status.
hipError_t hipEventCreate(hipEvent_t *event)
hipError_t hipEventDestroy(hipEvent_t event)
Destroy the specified event.
hipError_t hipEventElapsedTime(float *ms, hipEvent_t start, hipEvent_t stop)
Return the elapsed time between two events.
hipError_t hipEventCreateWithFlags(hipEvent_t *event, unsigned flags)
Create an event with the specified flags.
hipError_t hipFuncSetSharedMemConfig(const void *func, hipSharedMemConfig config)
Set shared memory configuation for a specific function.
hipError_t hipFuncSetAttribute(const void *func, hipFuncAttribute attr, int value)
Set attribute for a specific function.
hipError_t hipFuncSetCacheConfig(const void *func, hipFuncCache_t config)
Set Cache configuration for a specific function.
hipError_t hipDestroyExternalMemory(hipExternalMemory_t extMem)
Destroys an external memory object.
hipError_t hipImportExternalSemaphore(hipExternalSemaphore_t *extSem_out, const hipExternalSemaphoreHandleDesc *semHandleDesc)
Imports an external semaphore.
hipError_t hipImportExternalMemory(hipExternalMemory_t *extMem_out, const hipExternalMemoryHandleDesc *memHandleDesc)
Imports an external memory object.
hipError_t hipDestroyExternalSemaphore(hipExternalSemaphore_t extSem)
Destroys an external semaphore object and releases any references to the underlying resource....
hipError_t hipWaitExternalSemaphoresAsync(const hipExternalSemaphore_t *extSemArray, const hipExternalSemaphoreWaitParams *paramsArray, unsigned int numExtSems, hipStream_t stream)
Waits on a set of external semaphore objects.
hipError_t hipSignalExternalSemaphoresAsync(const hipExternalSemaphore_t *extSemArray, const hipExternalSemaphoreSignalParams *paramsArray, unsigned int numExtSems, hipStream_t stream)
Signals a set of external semaphore objects.
hipError_t hipExternalMemoryGetMappedBuffer(void **devPtr, hipExternalMemory_t extMem, const hipExternalMemoryBufferDesc *bufferDesc)
Maps a buffer onto an imported memory object.
hipError_t hipGLGetDevices(unsigned int *pHipDeviceCount, int *pHipDevices, unsigned int hipDeviceCount, hipGLDeviceList deviceList)
Queries devices associated with the current OpenGL context.
hipError_t hipGraphicsGLRegisterImage(hipGraphicsResource **resource, GLuint image, GLenum target, unsigned int flags)
Register a GL Image for interop and returns the corresponding graphic resource.
hipError_t hipGraphicsUnregisterResource(hipGraphicsResource_t resource)
Unregisters a graphics resource.
hipError_t hipGraphicsGLRegisterBuffer(hipGraphicsResource **resource, GLuint buffer, unsigned int flags)
Registers a GL Buffer for interop and returns corresponding graphics resource.
hipError_t hipGraphicsMapResources(int count, hipGraphicsResource_t *resources, hipStream_t stream __dparm(0))
Maps a graphics resource for access.
unsigned int GLenum
Definition hip_runtime_api.h:7728
hipError_t hipGraphicsSubResourceGetMappedArray(hipArray_t *array, hipGraphicsResource_t resource, unsigned int arrayIndex, unsigned int mipLevel)
Get an array through which to access a subresource of a mapped graphics resource.
unsigned int GLuint
Definition hip_runtime_api.h:7726
hipError_t hipGraphicsUnmapResources(int count, hipGraphicsResource_t *resources, hipStream_t stream __dparm(0))
Unmaps graphics resources.
hipError_t hipGraphicsResourceGetMappedPointer(void **devPtr, size_t *size, hipGraphicsResource_t resource)
Gets device accessible address of a graphics resource.
hipLimit_t
Definition hip_runtime_api.h:558
hipMemAllocationHandleType
Definition hip_runtime_api.h:869
hipGraphicsResource * hipGraphicsResource_t
Definition hip_runtime_api.h:1092
struct ihipCtx_t * hipCtx_t
Definition hip_runtime_api.h:520
#define __dparm(x)
Definition hip_runtime_api.h:504
struct ihipStream_t * hipStream_t
Definition hip_runtime_api.h:529
hipGLDeviceList
Definition hip_runtime_api.h:1069
int hipDevice_t
Definition hip_runtime_api.h:522
hipAccessProperty
Definition hip_runtime_api.h:1173
hipMemRangeAttribute
Definition hip_runtime_api.h:751
#define hipMemAttachSingle
Definition hip_runtime_api.h:640
hipMemoryAdvise
Definition hip_runtime_api.h:718
#define hipArrayDefault
Definition hip_runtime_api.h:688
hipSharedMemConfig
Definition hip_runtime_api.h:938
hipComputeMode
Definition hip_runtime_api.h:473
struct ihipEvent_t * hipEvent_t
Definition hip_runtime_api.h:557
hipStreamCaptureMode
Definition hip_runtime_api.h:1208
hipGraphNodeType
Definition hip_runtime_api.h:1116
hipExternalMemoryHandleType
Definition hip_runtime_api.h:982
#define hipMemAttachGlobal
Definition hip_runtime_api.h:634
hipFuncAttribute
Definition hip_runtime_api.h:921
hipExternalSemaphoreHandleType
Definition hip_runtime_api.h:1009
hipJitOption
Definition hip_runtime_api.h:898
hipDeviceP2PAttr
Definition hip_runtime_api.h:523
hipGraphDebugDotFlags
Definition hip_runtime_api.h:1250
hipUserObjectRetainFlags
Definition hip_runtime_api.h:1235
enum __HIP_NODISCARD hipError_t hipError_t
Definition hip_runtime_api.h:218
void(* hipHostFn_t)(void *userData)
Definition hip_runtime_api.h:1134
#define hipOccupancyDefault
Definition hip_runtime_api.h:693
hipStreamUpdateCaptureDependenciesFlags
Definition hip_runtime_api.h:1220
struct ihipGraph * hipGraph_t
Definition hip_runtime_api.h:1097
hipMemHandleType
Definition hip_runtime_api.h:1323
hipFuncCache_t
Definition hip_runtime_api.h:929
hipUserObjectFlags
Definition hip_runtime_api.h:1231
#define __HIP_NODISCARD
Definition hip_runtime_api.h:208
#define DEPRECATED_MSG
Definition hip_runtime_api.h:496
hipGraphMemAttributeType
Definition hip_runtime_api.h:1225
hipMemPoolAttr
Definition hip_runtime_api.h:767
struct ihipModule_t * hipModule_t
Definition hip_runtime_api.h:538
struct hipUserObject * hipUserObject_t
Definition hip_runtime_api.h:1110
hipMemOperationType
Definition hip_runtime_api.h:1330
void * hipExternalSemaphore_t
Definition hip_runtime_api.h:1026
hipGraphicsRegisterFlags
Definition hip_runtime_api.h:1080
hipMemRangeCoherencyMode
Definition hip_runtime_api.h:739
hipMemAccessFlags
Definition hip_runtime_api.h:842
hipMemAllocationGranularity_flags
Definition hip_runtime_api.h:1315
struct hipGraphExec * hipGraphExec_t
Definition hip_runtime_api.h:1105
hipGraphExecUpdateResult
Definition hip_runtime_api.h:1193
struct ihipModuleSymbol_t * hipFunction_t
Definition hip_runtime_api.h:539
void * hipExternalMemory_t
Definition hip_runtime_api.h:1008
hipStreamCaptureStatus
Definition hip_runtime_api.h:1213
hipDeviceAttribute_t
Definition hip_runtime_api.h:340
#define DEPRECATED(msg)
Definition hip_runtime_api.h:494
#define hipHostMallocDefault
Definition hip_runtime_api.h:607
struct _hipGraphicsResource hipGraphicsResource
Definition hip_runtime_api.h:1090
hipMemAllocationType
Definition hip_runtime_api.h:857
struct hipGraphNode * hipGraphNode_t
Definition hip_runtime_api.h:1101
struct ihipMemGenericAllocationHandle * hipMemGenericAllocationHandle_t
Definition hip_runtime_api.h:1310
hipMemoryType
Definition hip_runtime_api.h:175
hipGraphInstantiateFlags
Definition hip_runtime_api.h:1239
hipKernelNodeAttrID
Definition hip_runtime_api.h:1169
hipMemLocationType
Definition hip_runtime_api.h:825
#define HIP_IPC_HANDLE_SIZE
Definition hip_runtime_api.h:531
struct ihipMemPoolHandle_t * hipMemPool_t
Definition hip_runtime_api.h:543
hipArraySparseSubresourceType
Definition hip_runtime_api.h:1338
@ hipLimitMallocHeapSize
limit of heap size in bytes on the current device
Definition hip_runtime_api.h:561
@ hipLimitStackSize
limit of stack size in bytes on the current device
Definition hip_runtime_api.h:559
@ hipLimitRange
supported limit range
Definition hip_runtime_api.h:562
@ hipLimitPrintfFifoSize
size limit in bytes of fifo used by printf call on the device
Definition hip_runtime_api.h:560
@ hipMemHandleTypeWin32
Allows a Win32 NT handle for exporting. (HANDLE)
Definition hip_runtime_api.h:872
@ hipMemHandleTypeWin32Kmt
Allows a Win32 KMT handle for exporting. (D3DKMT_HANDLE)
Definition hip_runtime_api.h:873
@ hipMemHandleTypeNone
Does not allow any export mechanism.
Definition hip_runtime_api.h:870
@ hipMemHandleTypePosixFileDescriptor
Allows a file descriptor for exporting. Permitted only on POSIX systems.
Definition hip_runtime_api.h:871
@ hipGLDeviceListNextFrame
frame.
Definition hip_runtime_api.h:1073
@ hipGLDeviceListAll
All hip devices used by current OpenGL context.
Definition hip_runtime_api.h:1070
@ hipGLDeviceListCurrentFrame
frame
Definition hip_runtime_api.h:1071
@ hipAccessPropertyNormal
Definition hip_runtime_api.h:1174
@ hipAccessPropertyPersisting
Definition hip_runtime_api.h:1176
@ hipAccessPropertyStreaming
Definition hip_runtime_api.h:1175
@ hipMemRangeAttributeLastPrefetchLocation
prefetched
Definition hip_runtime_api.h:757
@ hipMemRangeAttributePreferredLocation
The preferred location of the range.
Definition hip_runtime_api.h:754
@ hipMemRangeAttributeAccessedBy
Definition hip_runtime_api.h:755
@ hipMemRangeAttributeReadMostly
Definition hip_runtime_api.h:752
@ hipMemRangeAttributeCoherencyMode
Definition hip_runtime_api.h:759
@ hipMemAdviseUnsetAccessedBy
Definition hip_runtime_api.h:727
@ hipMemAdviseUnsetCoarseGrain
Restores cache coherency policy back to fine-grain.
Definition hip_runtime_api.h:734
@ hipMemAdviseSetCoarseGrain
Definition hip_runtime_api.h:729
@ hipMemAdviseUnsetPreferredLocation
Clear the preferred location for the data.
Definition hip_runtime_api.h:724
@ hipMemAdviseSetAccessedBy
Definition hip_runtime_api.h:725
@ hipMemAdviseSetPreferredLocation
Definition hip_runtime_api.h:722
@ hipMemAdviseSetReadMostly
Definition hip_runtime_api.h:719
@ hipMemAdviseUnsetReadMostly
Undo the effect of hipMemAdviseSetReadMostly.
Definition hip_runtime_api.h:721
@ hipSharedMemBankSizeFourByte
Definition hip_runtime_api.h:940
@ hipSharedMemBankSizeEightByte
Definition hip_runtime_api.h:942
@ hipSharedMemBankSizeDefault
The compiler selects a device-specific value for the banking.
Definition hip_runtime_api.h:939
@ hipComputeModeProhibited
Definition hip_runtime_api.h:476
@ hipComputeModeExclusive
Definition hip_runtime_api.h:475
@ hipComputeModeDefault
Definition hip_runtime_api.h:474
@ hipComputeModeExclusiveProcess
Definition hip_runtime_api.h:477
@ hipStreamCaptureModeRelaxed
Definition hip_runtime_api.h:1211
@ hipStreamCaptureModeThreadLocal
Definition hip_runtime_api.h:1210
@ hipStreamCaptureModeGlobal
Definition hip_runtime_api.h:1209
@ hipGraphNodeTypeGraph
Node which executes an embedded graph.
Definition hip_runtime_api.h:1121
@ hipGraphNodeTypeMemset
Memset node.
Definition hip_runtime_api.h:1119
@ hipGraphNodeTypeEventRecord
External event record node.
Definition hip_runtime_api.h:1124
@ hipGraphNodeTypeExtSemaphoreSignal
External Semaphore signal node.
Definition hip_runtime_api.h:1125
@ hipGraphNodeTypeMemcpy
Memcpy node.
Definition hip_runtime_api.h:1118
@ hipGraphNodeTypeWaitEvent
External event wait node.
Definition hip_runtime_api.h:1123
@ hipGraphNodeTypeCount
Definition hip_runtime_api.h:1131
@ hipGraphNodeTypeKernel
GPU kernel node.
Definition hip_runtime_api.h:1117
@ hipGraphNodeTypeHost
Host (executable) node.
Definition hip_runtime_api.h:1120
@ hipGraphNodeTypeMemAlloc
Memory alloc node.
Definition hip_runtime_api.h:1127
@ hipGraphNodeTypeMemFree
Memory free node.
Definition hip_runtime_api.h:1128
@ hipGraphNodeTypeMemcpyFromSymbol
MemcpyFromSymbol node.
Definition hip_runtime_api.h:1129
@ hipGraphNodeTypeMemcpyToSymbol
MemcpyToSymbol node.
Definition hip_runtime_api.h:1130
@ hipGraphNodeTypeEmpty
Empty (no-op) node.
Definition hip_runtime_api.h:1122
@ hipGraphNodeTypeExtSemaphoreWait
External Semaphore wait node.
Definition hip_runtime_api.h:1126
@ hipExternalMemoryHandleTypeD3D11Resource
Definition hip_runtime_api.h:988
@ hipExternalMemoryHandleTypeD3D12Resource
Definition hip_runtime_api.h:987
@ hipExternalMemoryHandleTypeOpaqueFd
Definition hip_runtime_api.h:983
@ hipExternalMemoryHandleTypeD3D12Heap
Definition hip_runtime_api.h:986
@ hipExternalMemoryHandleTypeOpaqueWin32Kmt
Definition hip_runtime_api.h:985
@ hipExternalMemoryHandleTypeOpaqueWin32
Definition hip_runtime_api.h:984
@ hipExternalMemoryHandleTypeD3D11ResourceKmt
Definition hip_runtime_api.h:989
@ hipFuncAttributePreferredSharedMemoryCarveout
Definition hip_runtime_api.h:923
@ hipFuncAttributeMaxDynamicSharedMemorySize
Definition hip_runtime_api.h:922
@ hipFuncAttributeMax
Definition hip_runtime_api.h:924
@ hipExternalSemaphoreHandleTypeOpaqueWin32Kmt
Definition hip_runtime_api.h:1012
@ hipExternalSemaphoreHandleTypeOpaqueFd
Definition hip_runtime_api.h:1010
@ hipExternalSemaphoreHandleTypeOpaqueWin32
Definition hip_runtime_api.h:1011
@ hipExternalSemaphoreHandleTypeD3D12Fence
Definition hip_runtime_api.h:1013
@ hipJitOptionCacheMode
Definition hip_runtime_api.h:913
@ hipJitOptionFastCompile
Definition hip_runtime_api.h:915
@ hipJitOptionInfoLogBuffer
Definition hip_runtime_api.h:902
@ hipJitOptionErrorLogBuffer
Definition hip_runtime_api.h:904
@ hipJitOptionTarget
Definition hip_runtime_api.h:908
@ hipJitOptionThreadsPerBlock
Definition hip_runtime_api.h:900
@ hipJitOptionNumOptions
Definition hip_runtime_api.h:916
@ hipJitOptionGenerateLineInfo
Definition hip_runtime_api.h:912
@ hipJitOptionOptimizationLevel
Definition hip_runtime_api.h:906
@ hipJitOptionErrorLogBufferSizeBytes
Definition hip_runtime_api.h:905
@ hipJitOptionWallTime
Definition hip_runtime_api.h:901
@ hipJitOptionInfoLogBufferSizeBytes
Definition hip_runtime_api.h:903
@ hipJitOptionTargetFromContext
Definition hip_runtime_api.h:907
@ hipJitOptionGenerateDebugInfo
Definition hip_runtime_api.h:910
@ hipJitOptionLogVerbose
Definition hip_runtime_api.h:911
@ hipJitOptionMaxRegisters
Definition hip_runtime_api.h:899
@ hipJitOptionSm3xOpt
Definition hip_runtime_api.h:914
@ hipJitOptionFallbackStrategy
Definition hip_runtime_api.h:909
@ hipDevP2PAttrAccessSupported
Definition hip_runtime_api.h:525
@ hipDevP2PAttrHipArrayAccessSupported
Definition hip_runtime_api.h:527
@ hipDevP2PAttrNativeAtomicSupported
Definition hip_runtime_api.h:526
@ hipDevP2PAttrPerformanceRank
Definition hip_runtime_api.h:524
@ hipGraphDebugDotFlagsKernelNodeAttributes
Definition hip_runtime_api.h:1263
@ hipGraphDebugDotFlagsMemsetNodeParams
Definition hip_runtime_api.h:1255
@ hipGraphDebugDotFlagsKernelNodeParams
Definition hip_runtime_api.h:1253
@ hipGraphDebugDotFlagsExtSemasWaitNodeParams
Definition hip_runtime_api.h:1261
@ hipGraphDebugDotFlagsHandles
Definition hip_runtime_api.h:1265
@ hipGraphDebugDotFlagsExtSemasSignalNodeParams
Definition hip_runtime_api.h:1259
@ hipGraphDebugDotFlagsHostNodeParams
Definition hip_runtime_api.h:1256
@ hipGraphDebugDotFlagsEventNodeParams
Definition hip_runtime_api.h:1257
@ hipGraphDebugDotFlagsVerbose
Definition hip_runtime_api.h:1251
@ hipGraphDebugDotFlagsMemcpyNodeParams
Definition hip_runtime_api.h:1254
@ hipGraphUserObjectMove
Add new reference or retain.
Definition hip_runtime_api.h:1236
@ hipStreamAddCaptureDependencies
Add new nodes to the dependency set.
Definition hip_runtime_api.h:1221
@ hipStreamSetCaptureDependencies
Replace the dependency set with the new nodes.
Definition hip_runtime_api.h:1222
@ hipMemHandleTypeGeneric
Generic handle type.
Definition hip_runtime_api.h:1324
@ hipFuncCachePreferNone
no preference for shared memory or L1 (default)
Definition hip_runtime_api.h:930
@ hipFuncCachePreferEqual
prefer equal size L1 cache and shared memory
Definition hip_runtime_api.h:933
@ hipFuncCachePreferL1
prefer larger L1 cache and smaller shared memory
Definition hip_runtime_api.h:932
@ hipFuncCachePreferShared
prefer larger shared memory and smaller L1 cache
Definition hip_runtime_api.h:931
@ hipUserObjectNoDestructorSync
Destructor execution is not synchronized.
Definition hip_runtime_api.h:1232
@ hipGraphMemAttrReservedMemHigh
High watermark of memory, in bytes, currently allocated for graphs.
Definition hip_runtime_api.h:1229
@ hipGraphMemAttrUsedMemCurrent
Amount of memory, in bytes, currently associated with graphs.
Definition hip_runtime_api.h:1226
@ hipGraphMemAttrUsedMemHigh
High watermark of memory, in bytes, associated with graphs since the last time.
Definition hip_runtime_api.h:1227
@ hipGraphMemAttrReservedMemCurrent
Amount of memory, in bytes, currently allocated for graphs.
Definition hip_runtime_api.h:1228
@ hipMemPoolAttrUsedMemCurrent
Definition hip_runtime_api.h:814
@ hipMemPoolAttrReservedMemHigh
Definition hip_runtime_api.h:809
@ hipMemPoolAttrReservedMemCurrent
Definition hip_runtime_api.h:803
@ hipMemPoolAttrReleaseThreshold
Definition hip_runtime_api.h:798
@ hipMemPoolAttrUsedMemHigh
Definition hip_runtime_api.h:820
@ hipMemPoolReuseAllowInternalDependencies
Definition hip_runtime_api.h:789
@ hipMemPoolReuseFollowEventDependencies
Definition hip_runtime_api.h:776
@ hipMemPoolReuseAllowOpportunistic
Definition hip_runtime_api.h:782
@ hipMemOperationTypeMap
Map operation.
Definition hip_runtime_api.h:1331
@ hipMemOperationTypeUnmap
Unmap operation.
Definition hip_runtime_api.h:1332
@ hipGraphicsRegisterFlagsReadOnly
HIP will not write to this registered resource.
Definition hip_runtime_api.h:1082
@ hipGraphicsRegisterFlagsTextureGather
HIP will perform texture gather operations on this registered resource.
Definition hip_runtime_api.h:1086
@ hipGraphicsRegisterFlagsWriteDiscard
HIP will only write and will not read from this registered resource.
Definition hip_runtime_api.h:1083
@ hipGraphicsRegisterFlagsNone
Definition hip_runtime_api.h:1081
@ hipGraphicsRegisterFlagsSurfaceLoadStore
HIP will bind this resource to a surface.
Definition hip_runtime_api.h:1085
@ hipMemRangeCoherencyModeFineGrain
Definition hip_runtime_api.h:740
@ hipMemRangeCoherencyModeIndeterminate
Definition hip_runtime_api.h:744
@ hipMemRangeCoherencyModeCoarseGrain
Definition hip_runtime_api.h:742
@ hipMemAccessFlagsProtRead
Set the address range read accessible.
Definition hip_runtime_api.h:844
@ hipMemAccessFlagsProtNone
Default, make the address range not accessible.
Definition hip_runtime_api.h:843
@ hipMemAccessFlagsProtReadWrite
Set the address range read-write accessible.
Definition hip_runtime_api.h:845
@ hipMemAllocationGranularityMinimum
Minimum granularity.
Definition hip_runtime_api.h:1316
@ hipMemAllocationGranularityRecommended
Recommended granularity for performance.
Definition hip_runtime_api.h:1317
@ hipGraphExecUpdateErrorNotSupported
The update failed because something about the node is not supported.
Definition hip_runtime_api.h:1203
@ hipGraphExecUpdateErrorNodeTypeChanged
The update failed because a node type changed.
Definition hip_runtime_api.h:1198
@ hipGraphExecUpdateErrorTopologyChanged
The update failed because the topology changed.
Definition hip_runtime_api.h:1197
@ hipGraphExecUpdateErrorUnsupportedFunctionChange
Definition hip_runtime_api.h:1205
@ hipGraphExecUpdateError
Definition hip_runtime_api.h:1195
@ hipGraphExecUpdateErrorFunctionChanged
The update failed because the function of a kernel node changed.
Definition hip_runtime_api.h:1199
@ hipGraphExecUpdateSuccess
The update succeeded.
Definition hip_runtime_api.h:1194
@ hipGraphExecUpdateErrorParametersChanged
The update failed because the parameters changed in a way that is not supported.
Definition hip_runtime_api.h:1201
@ hipStreamCaptureStatusInvalidated
Definition hip_runtime_api.h:1216
@ hipStreamCaptureStatusNone
Stream is not capturing.
Definition hip_runtime_api.h:1214
@ hipStreamCaptureStatusActive
Stream is actively capturing.
Definition hip_runtime_api.h:1215
@ hipDeviceAttributeDirectManagedMemAccessFromHost
Definition hip_runtime_api.h:358
@ hipDeviceAttributeSurfaceAlignment
Cuda only. Alignment requirement for surfaces.
Definition hip_runtime_api.h:428
@ hipDeviceAttributeMaxGridDimX
Max grid size in width.
Definition hip_runtime_api.h:375
@ hipDeviceAttributeMaxSurfaceCubemapLayered
Cuda only. Maximum dimension of Cubemap layered surface.
Definition hip_runtime_api.h:384
@ hipDeviceAttributeMaxSurface3D
Maximum dimension (width, height, depth) of 3D surface.
Definition hip_runtime_api.h:382
@ hipDeviceAttributeMaxPitch
Maximum pitch in bytes allowed by memory copies.
Definition hip_runtime_api.h:405
@ hipDeviceAttributeTccDriver
Cuda only. Whether device is a Tesla device using TCC driver.
Definition hip_runtime_api.h:429
@ hipDeviceAttributeHostNativeAtomicSupported
Cuda only. Link between the device and the host supports native atomic operations.
Definition hip_runtime_api.h:361
@ hipDeviceAttributePageableMemoryAccessUsesHostPageTables
Device accesses pageable memory via the host's page tables.
Definition hip_runtime_api.h:414
@ hipDeviceAttributeMaxBlockDimY
Max block size in height.
Definition hip_runtime_api.h:373
@ hipDeviceAttributeMaxSurfaceCubemap
Cuda only. Maximum dimensions of Cubemap surface.
Definition hip_runtime_api.h:383
@ hipDeviceAttributeName
Device name.
Definition hip_runtime_api.h:411
@ hipDeviceAttributePageableMemoryAccess
Definition hip_runtime_api.h:412
@ hipDeviceAttributeClockRate
Peak clock frequency in kilohertz.
Definition hip_runtime_api.h:349
@ hipDeviceAttributeCudaCompatibleEnd
Definition hip_runtime_api.h:440
@ hipDeviceAttributeMaxTexture1DMipmap
Cuda only. Maximum size of 1D mipmapped texture.
Definition hip_runtime_api.h:389
@ hipDeviceAttributeComputeCapabilityMajor
Major compute capability version number.
Definition hip_runtime_api.h:369
@ hipDeviceAttributeGlobalL1CacheSupported
Cuda only. Device supports caching globals in L1.
Definition hip_runtime_api.h:360
@ hipDeviceAttributeCooperativeMultiDeviceUnmatchedSharedMem
Definition hip_runtime_api.h:456
@ hipDeviceAttributeAccessPolicyMaxWindowSize
Cuda only. The maximum size of the window policy in bytes.
Definition hip_runtime_api.h:344
@ hipDeviceAttributeMaxTexture3DHeight
Maximum dimension height of 3D texture.
Definition hip_runtime_api.h:397
@ hipDeviceAttributeComputeCapabilityMinor
Minor compute capability version number.
Definition hip_runtime_api.h:408
@ hipDeviceAttributeMaxTextureCubemap
Cuda only. Maximum dimensions of Cubemap texture.
Definition hip_runtime_api.h:400
@ hipDeviceAttributeDeviceOverlap
Definition hip_runtime_api.h:356
@ hipDeviceAttributeMaxTexture2DLinear
Cuda only. Maximum dimensions (width, height, pitch) of 2D textures bound to pitched memory.
Definition hip_runtime_api.h:394
@ hipDeviceAttributeMaxSurface2D
Maximum dimension (width, height) of 2D surface.
Definition hip_runtime_api.h:380
@ hipDeviceAttributeMaxSurface1DLayered
Cuda only. Maximum dimensions of 1D layered surface.
Definition hip_runtime_api.h:379
@ hipDeviceAttributeHdpRegFlushCntl
Address of the HDP_REG_COHERENCY_FLUSH_CNTL register.
Definition hip_runtime_api.h:449
@ hipDeviceAttributeComputeMode
Compute mode that device is currently in.
Definition hip_runtime_api.h:350
@ hipDeviceAttributeCooperativeMultiDeviceUnmatchedFunc
Definition hip_runtime_api.h:450
@ hipDeviceAttributeAsicRevision
Revision of the GPU in this device.
Definition hip_runtime_api.h:459
@ hipDeviceAttributeCooperativeMultiDeviceLaunch
Support cooperative launch on multiple devices.
Definition hip_runtime_api.h:355
@ hipDeviceAttributeMaxTexture2DHeight
Maximum dimension hight of 2D texture.
Definition hip_runtime_api.h:391
@ hipDeviceAttributePersistingL2CacheMaxSize
Cuda11 only. Maximum l2 persisting lines capacity in bytes.
Definition hip_runtime_api.h:418
@ hipDeviceAttributeEccEnabled
Whether ECC support is enabled.
Definition hip_runtime_api.h:343
@ hipDeviceAttributeCanUseHostPointerForRegisteredMem
Definition hip_runtime_api.h:347
@ hipDeviceAttributePciBusId
PCI Bus ID.
Definition hip_runtime_api.h:415
@ hipDeviceAttributeL2CacheSize
Size of L2 cache in bytes. 0 if the device doesn't have L2 cache.
Definition hip_runtime_api.h:365
@ hipDeviceAttributeKernelExecTimeout
Run time limit for kernels executed on the device.
Definition hip_runtime_api.h:364
@ hipDeviceAttributeMaxThreadsDim
Maximum dimension of a block.
Definition hip_runtime_api.h:402
@ hipDeviceAttributeSingleToDoublePrecisionPerfRatio
Cuda only. Performance ratio of single precision to double precision.
Definition hip_runtime_api.h:426
@ hipDeviceAttributeMaxGridDimY
Max grid size in height.
Definition hip_runtime_api.h:376
@ hipDeviceAttributeMultiprocessorCount
Number of multiprocessors on the device.
Definition hip_runtime_api.h:410
@ hipDeviceAttributeUuid
Cuda only. Unique ID in 16 byte.
Definition hip_runtime_api.h:435
@ hipDeviceAttributeAmdSpecificBegin
Definition hip_runtime_api.h:441
@ hipDeviceAttributeMaxRegistersPerMultiprocessor
32-bit registers available per block.
Definition hip_runtime_api.h:421
@ hipDeviceAttributeGcnArchName
Device gcnArch name in 256 bytes.
Definition hip_runtime_api.h:447
@ hipDeviceAttributeIsMultiGpuBoard
Multiple GPU devices.
Definition hip_runtime_api.h:363
@ hipDeviceAttributeSharedMemPerBlockOptin
Cuda only. Maximum shared memory per block usable by special opt in.
Definition hip_runtime_api.h:424
@ hipDeviceAttributeMaxSurface2DLayered
Cuda only. Maximum dimensions of 2D layered surface.
Definition hip_runtime_api.h:381
@ hipDeviceAttributeAmdSpecificEnd
Definition hip_runtime_api.h:468
@ hipDeviceAttributeMemoryClockRate
Peak memory clock frequency in kilohertz.
Definition hip_runtime_api.h:407
@ hipDeviceAttributeGcnArch
Device gcn architecture.
Definition hip_runtime_api.h:446
@ hipDeviceAttributeMaxGridDimZ
Max grid size in depth.
Definition hip_runtime_api.h:377
@ hipDeviceAttributeHdpMemFlushCntl
Address of the HDP_MEM_COHERENCY_FLUSH_CNTL register.
Definition hip_runtime_api.h:448
@ hipDeviceAttributeFineGrainSupport
'1' if Device supports fine grain, '0' otherwise
Definition hip_runtime_api.h:465
@ hipDeviceAttributeCooperativeLaunch
Support cooperative launch.
Definition hip_runtime_api.h:354
@ hipDeviceAttributeUnifiedAddressing
Cuda only. An unified address space shared with the host.
Definition hip_runtime_api.h:434
@ hipDeviceAttributeAsyncEngineCount
Cuda only. Asynchronous engines number.
Definition hip_runtime_api.h:345
@ hipDeviceAttributeMultiGpuBoardGroupID
Cuda only. Unique ID of device group on the same multi-GPU board.
Definition hip_runtime_api.h:409
@ hipDeviceAttributeStreamPrioritiesSupported
Cuda only. Whether to support stream priorities.
Definition hip_runtime_api.h:427
@ hipDeviceAttributeMaxSharedMemoryPerBlock
Maximum shared memory available per block in bytes.
Definition hip_runtime_api.h:423
@ hipDeviceAttributeLocalL1CacheSupported
caching locals in L1 is supported
Definition hip_runtime_api.h:366
@ hipDeviceAttributeCanUseStreamWaitValue
Definition hip_runtime_api.h:460
@ hipDeviceAttributeMaxRegistersPerBlock
Definition hip_runtime_api.h:419
@ hipDeviceAttributeMaxThreadsPerBlock
Maximum number of threads per block.
Definition hip_runtime_api.h:403
@ hipDeviceAttributeMemoryBusWidth
Global memory bus width in bits.
Definition hip_runtime_api.h:406
@ hipDeviceAttributeComputePreemptionSupported
Cuda only. Device supports Compute Preemption.
Definition hip_runtime_api.h:351
@ hipDeviceAttributeMaxSurface1D
Maximum size of 1D surface.
Definition hip_runtime_api.h:378
@ hipDeviceAttributeMaxTexture3DAlt
Cuda only. Maximum dimensions of alternate 3D texture.
Definition hip_runtime_api.h:399
@ hipDeviceAttributeMaxTexture1DLinear
Definition hip_runtime_api.h:387
@ hipDeviceAttributeMaxTexture2DGather
Cuda only. Maximum dimensions of 2D texture if gather operations performed.
Definition hip_runtime_api.h:392
@ hipDeviceAttributeWallClockRate
Constant frequency of wall clock in kilohertz.
Definition hip_runtime_api.h:466
@ hipDeviceAttributePciDeviceId
PCI Device ID.
Definition hip_runtime_api.h:416
@ hipDeviceAttributeCanMapHostMemory
Whether host memory can be mapped into device address space.
Definition hip_runtime_api.h:346
@ hipDeviceAttributeMaxTexture1DWidth
Maximum size of 1D texture.
Definition hip_runtime_api.h:385
@ hipDeviceAttributeConcurrentManagedAccess
Device can coherently access managed memory concurrently with the CPU.
Definition hip_runtime_api.h:353
@ hipDeviceAttributeVendorSpecificBegin
Definition hip_runtime_api.h:469
@ hipDeviceAttributeArch
Device architecture.
Definition hip_runtime_api.h:444
@ hipDeviceAttributeManagedMemory
Device supports allocating managed memory on this system.
Definition hip_runtime_api.h:370
@ hipDeviceAttributeIntegrated
Device is integrated GPU.
Definition hip_runtime_api.h:362
@ hipDeviceAttributeMaxTexture2DLayered
Cuda only. Maximum dimensions of 2D layered texture.
Definition hip_runtime_api.h:393
@ hipDeviceAttributeCooperativeMultiDeviceUnmatchedGridDim
Definition hip_runtime_api.h:452
@ hipDeviceAttributeMaxTexture3DDepth
Maximum dimension depth of 3D texture.
Definition hip_runtime_api.h:398
@ hipDeviceAttributeImageSupport
'1' if Device supports image, '0' otherwise.
Definition hip_runtime_api.h:462
@ hipDeviceAttributeTotalGlobalMem
Global memory available on devicice.
Definition hip_runtime_api.h:433
@ hipDeviceAttributeMaxBlocksPerMultiProcessor
Cuda only. Max block size per multiprocessor.
Definition hip_runtime_api.h:371
@ hipDeviceAttributeIsLargeBar
Whether it is LargeBar.
Definition hip_runtime_api.h:458
@ hipDeviceAttributeMaxTexture2DWidth
Maximum dimension width of 2D texture.
Definition hip_runtime_api.h:390
@ hipDeviceAttributeMaxBlockDimX
Max block size in width.
Definition hip_runtime_api.h:372
@ hipDeviceAttributeMaxTexture3DWidth
Maximum dimension width of 3D texture.
Definition hip_runtime_api.h:396
@ hipDeviceAttributeTotalConstantMemory
Constant memory size in bytes.
Definition hip_runtime_api.h:432
@ hipDeviceAttributeTextureAlignment
Alignment requirement for textures.
Definition hip_runtime_api.h:430
@ hipDeviceAttributeMaxTexture1DLayered
Cuda only. Maximum dimensions of 1D layered texture.
Definition hip_runtime_api.h:386
@ hipDeviceAttributeMaxSharedMemoryPerMultiprocessor
Maximum Shared Memory PerMultiprocessor.
Definition hip_runtime_api.h:445
@ hipDeviceAttributeReservedSharedMemPerBlock
Cuda11 only. Shared memory reserved by CUDA driver per block.
Definition hip_runtime_api.h:422
@ hipDeviceAttributeCudaCompatibleBegin
Definition hip_runtime_api.h:341
@ hipDeviceAttributeConcurrentKernels
Device can possibly execute multiple kernels concurrently.
Definition hip_runtime_api.h:352
@ hipDeviceAttributeLuid
Cuda only. 8-byte locally unique identifier in 8 bytes. Undefined on TCC and non-Windows platforms.
Definition hip_runtime_api.h:367
@ hipDeviceAttributePciDomainID
PCI Domain ID.
Definition hip_runtime_api.h:417
@ hipDeviceAttributeMaxThreadsPerMultiProcessor
Maximum resident threads per multiprocessor.
Definition hip_runtime_api.h:404
@ hipDeviceAttributeTexturePitchAlignment
Pitch alignment requirement for 2D texture references bound to pitched memory;.
Definition hip_runtime_api.h:431
@ hipDeviceAttributePhysicalMultiProcessorCount
Definition hip_runtime_api.h:463
@ hipDeviceAttributeSharedMemPerMultiprocessor
Cuda only. Shared memory available per multiprocessor.
Definition hip_runtime_api.h:425
@ hipDeviceAttributeLuidDeviceNodeMask
Cuda only. Luid device node mask. Undefined on TCC and non-Windows platforms.
Definition hip_runtime_api.h:368
@ hipDeviceAttributeMaxTextureCubemapLayered
Cuda only. Maximum dimensions of Cubemap layered texture.
Definition hip_runtime_api.h:401
@ hipDeviceAttributeCooperativeMultiDeviceUnmatchedBlockDim
Definition hip_runtime_api.h:454
@ hipDeviceAttributeVirtualMemoryManagementSupported
Device supports HIP virtual memory management.
Definition hip_runtime_api.h:438
@ hipDeviceAttributeMemoryPoolsSupported
Device supports HIP Stream Ordered Memory Allocator.
Definition hip_runtime_api.h:437
@ hipDeviceAttributeClockInstructionRate
Frequency in khz of the timer used by the device-side "clock*".
Definition hip_runtime_api.h:443
@ hipDeviceAttributeMaxTexture2DMipmap
Cuda only. Maximum dimensions of 2D mipmapped texture.
Definition hip_runtime_api.h:395
@ hipDeviceAttributeMaxBlockDimZ
Max block size in depth.
Definition hip_runtime_api.h:374
@ hipDeviceAttributeWarpSize
Warp size in threads.
Definition hip_runtime_api.h:436
@ hipMemAllocationTypeMax
Definition hip_runtime_api.h:863
@ hipMemAllocationTypePinned
Definition hip_runtime_api.h:862
@ hipMemAllocationTypeInvalid
Definition hip_runtime_api.h:858
@ hipMemoryTypeDevice
Definition hip_runtime_api.h:177
@ hipMemoryTypeHost
Memory is physically located on host.
Definition hip_runtime_api.h:176
@ hipMemoryTypeArray
Definition hip_runtime_api.h:179
@ hipMemoryTypeManaged
Definition hip_runtime_api.h:182
@ hipMemoryTypeUnified
Not used currently.
Definition hip_runtime_api.h:181
@ hipGraphInstantiateFlagAutoFreeOnLaunch
Automatically free memory allocated in a graph before relaunching.
Definition hip_runtime_api.h:1240
@ hipGraphInstantiateFlagUseNodePriority
Run the graph using the per-node priority attributes rather than the priority of the stream it is lau...
Definition hip_runtime_api.h:1246
@ hipGraphInstantiateFlagUpload
Automatically upload the graph after instantiaton.
Definition hip_runtime_api.h:1242
@ hipGraphInstantiateFlagDeviceLaunch
Instantiate the graph to be launchable from the device.
Definition hip_runtime_api.h:1244
@ hipKernelNodeAttributeCooperative
Definition hip_runtime_api.h:1171
@ hipKernelNodeAttributeAccessPolicyWindow
Definition hip_runtime_api.h:1170
@ hipMemLocationTypeInvalid
Definition hip_runtime_api.h:826
@ hipMemLocationTypeDevice
Device location, thus it's HIP device ID.
Definition hip_runtime_api.h:827
@ hipArraySparseSubresourceTypeSparseLevel
Sparse level.
Definition hip_runtime_api.h:1339
@ hipArraySparseSubresourceTypeMiptail
Miptail.
Definition hip_runtime_api.h:1340
hipError_t hipGraphKernelNodeSetParams(hipGraphNode_t node, const hipKernelNodeParams *pNodeParams)
Sets a kernel node's parameters.
hipError_t hipGraphMemFreeNodeGetParams(hipGraphNode_t node, void *dev_ptr)
Returns parameters for memory free node.
hipError_t hipGraphNodeGetDependencies(hipGraphNode_t node, hipGraphNode_t *pDependencies, size_t *pNumDependencies)
Returns a node's dependencies.
hipError_t hipGraphCreate(hipGraph_t *pGraph, unsigned int flags)
Creates a graph.
hipError_t hipGraphAddMemcpyNode1D(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, void *dst, const void *src, size_t count, hipMemcpyKind kind)
Creates a 1D memcpy node and adds it to a graph.
hipError_t hipDeviceSetGraphMemAttribute(int device, hipGraphMemAttributeType attr, void *value)
Set the mem attribute for graphs.
hipError_t hipGraphMemcpyNodeSetParams(hipGraphNode_t node, const hipMemcpy3DParms *pNodeParams)
Sets a memcpy node's parameters.
hipError_t hipUserObjectCreate(hipUserObject_t *object_out, void *ptr, hipHostFn_t destroy, unsigned int initialRefcount, unsigned int flags)
Create an instance of userObject to manage lifetime of a resource.
hipError_t hipGraphExecUpdate(hipGraphExec_t hGraphExec, hipGraph_t hGraph, hipGraphNode_t *hErrorNode_out, hipGraphExecUpdateResult *updateResult_out)
Check whether an executable graph can be updated with a graph and perform the update if * possible.
hipError_t hipGraphNodeFindInClone(hipGraphNode_t *pNode, hipGraphNode_t originalNode, hipGraph_t clonedGraph)
Finds a cloned version of a node.
hipError_t hipGraphAddEventWaitNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, hipEvent_t event)
Creates an event wait node and adds it to a graph.
hipError_t hipGraphMemcpyNodeSetParamsToSymbol(hipGraphNode_t node, const void *symbol, const void *src, size_t count, size_t offset, hipMemcpyKind kind)
Sets a memcpy node's parameters to copy to a symbol on the device.
hipError_t hipGraphRetainUserObject(hipGraph_t graph, hipUserObject_t object, unsigned int count __dparm(1), unsigned int flags __dparm(0))
Retain user object for graphs.
hipError_t hipStreamIsCapturing(hipStream_t stream, hipStreamCaptureStatus *pCaptureStatus)
Get stream's capture state.
hipError_t hipGraphNodeGetEnabled(hipGraphExec_t hGraphExec, hipGraphNode_t hNode, unsigned int *isEnabled)
Query whether a node in the given graphExec is enabled.
hipError_t hipGraphAddChildGraphNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, hipGraph_t childGraph)
Creates a child graph node and adds it to a graph.
hipError_t hipGraphExternalSemaphoresSignalNodeSetParams(hipGraphNode_t hNode, const hipExternalSemaphoreSignalNodeParams *nodeParams)
Updates node parameters in the external semaphore signal node.
hipError_t hipGraphDestroyNode(hipGraphNode_t node)
Remove a node from the graph.
hipError_t hipGraphMemcpyNodeSetParamsFromSymbol(hipGraphNode_t node, void *dst, const void *symbol, size_t count, size_t offset, hipMemcpyKind kind)
Sets a memcpy node's parameters to copy from a symbol on the device.
hipError_t hipGraphRemoveDependencies(hipGraph_t graph, const hipGraphNode_t *from, const hipGraphNode_t *to, size_t numDependencies)
Removes dependency edges from a graph.
hipError_t hipGraphMemcpyNodeSetParams1D(hipGraphNode_t node, void *dst, const void *src, size_t count, hipMemcpyKind kind)
Sets a memcpy node's parameters to perform a 1-dimensional copy.
hipError_t hipGraphExecMemcpyNodeSetParams1D(hipGraphExec_t hGraphExec, hipGraphNode_t node, void *dst, const void *src, size_t count, hipMemcpyKind kind)
Sets the parameters for a memcpy node in the given graphExec to perform a 1-dimensional copy.
hipError_t hipGraphExecDestroy(hipGraphExec_t graphExec)
Destroys an executable graph.
hipError_t hipGraphAddExternalSemaphoresSignalNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, const hipExternalSemaphoreSignalNodeParams *nodeParams)
Creates a external semaphor signal node and adds it to a graph.
hipError_t hipGraphExecExternalSemaphoresWaitNodeSetParams(hipGraphExec_t hGraphExec, hipGraphNode_t hNode, const hipExternalSemaphoreWaitNodeParams *nodeParams)
Updates node parameters in the external semaphore wait node in the given graphExec.
hipError_t hipGraphExecHostNodeSetParams(hipGraphExec_t hGraphExec, hipGraphNode_t node, const hipHostNodeParams *pNodeParams)
Sets the parameters for a host node in the given graphExec.
hipError_t hipGraphExecChildGraphNodeSetParams(hipGraphExec_t hGraphExec, hipGraphNode_t node, hipGraph_t childGraph)
Updates node parameters in the child graph node in the given graphExec.
hipError_t hipStreamGetCaptureInfo(hipStream_t stream, hipStreamCaptureStatus *pCaptureStatus, unsigned long long *pId)
Get capture status of a stream.
hipError_t hipGraphAddDependencies(hipGraph_t graph, const hipGraphNode_t *from, const hipGraphNode_t *to, size_t numDependencies)
Adds dependency edges to a graph.
hipError_t hipGraphExecKernelNodeSetParams(hipGraphExec_t hGraphExec, hipGraphNode_t node, const hipKernelNodeParams *pNodeParams)
Sets the parameters for a kernel node in the given graphExec.
hipError_t hipGraphNodeGetDependentNodes(hipGraphNode_t node, hipGraphNode_t *pDependentNodes, size_t *pNumDependentNodes)
Returns a node's dependent nodes.
hipError_t hipDeviceGetGraphMemAttribute(int device, hipGraphMemAttributeType attr, void *value)
Get the mem attribute for graphs.
hipError_t hipGraphHostNodeGetParams(hipGraphNode_t node, hipHostNodeParams *pNodeParams)
Returns a host node's parameters.
hipError_t hipGraphInstantiateWithFlags(hipGraphExec_t *pGraphExec, hipGraph_t graph, unsigned long long flags)
Creates an executable graph from a graph.
hipError_t hipGraphExternalSemaphoresWaitNodeGetParams(hipGraphNode_t hNode, const hipExternalSemaphoreWaitNodeParams *params_out)
Returns external semaphore wait node params.
hipError_t hipGraphUpload(hipGraphExec_t graphExec, hipStream_t stream)
uploads an executable graph in a stream
hipError_t hipGraphExecExternalSemaphoresSignalNodeSetParams(hipGraphExec_t hGraphExec, hipGraphNode_t hNode, const hipExternalSemaphoreSignalNodeParams *nodeParams)
Updates node parameters in the external semaphore signal node in the given graphExec.
hipError_t hipGraphExecMemcpyNodeSetParamsToSymbol(hipGraphExec_t hGraphExec, hipGraphNode_t node, const void *symbol, const void *src, size_t count, size_t offset, hipMemcpyKind kind)
Sets the parameters for a memcpy node in the given graphExec to copy to a symbol on the device.
hipError_t hipGraphKernelNodeGetParams(hipGraphNode_t node, hipKernelNodeParams *pNodeParams)
Gets kernel node's parameters.
hipError_t hipGraphExternalSemaphoresSignalNodeGetParams(hipGraphNode_t hNode, const hipExternalSemaphoreSignalNodeParams *params_out)
Returns external semaphore signal node params.
hipError_t hipGraphAddMemcpyNodeFromSymbol(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, void *dst, const void *symbol, size_t count, size_t offset, hipMemcpyKind kind)
Creates a memcpy node to copy from a symbol on the device and adds it to a graph.
hipError_t hipGraphKernelNodeCopyAttributes(hipGraphNode_t hSrc, hipGraphNode_t hDst)
Copies attributes from source node to destination node.
hipError_t hipGraphAddMemFreeNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, void *dev_ptr)
Creates a memory free node and adds it to a graph.
hipError_t hipUserObjectRelease(hipUserObject_t object, unsigned int count __dparm(1))
Release number of references to resource.
hipError_t hipGraphMemcpyNodeGetParams(hipGraphNode_t node, hipMemcpy3DParms *pNodeParams)
Gets a memcpy node's parameters.
hipError_t hipGraphAddEventRecordNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, hipEvent_t event)
Creates an event record node and adds it to a graph.
hipError_t hipStreamUpdateCaptureDependencies(hipStream_t stream, hipGraphNode_t *dependencies, size_t numDependencies, unsigned int flags __dparm(0))
Update the set of dependencies in a capturing stream.
hipError_t hipStreamBeginCapture(hipStream_t stream, hipStreamCaptureMode mode)
Begins graph capture on a stream.
hipError_t hipStreamEndCapture(hipStream_t stream, hipGraph_t *pGraph)
Ends capture on a stream, returning the captured graph.
hipError_t hipGraphNodeGetType(hipGraphNode_t node, hipGraphNodeType *pType)
Returns a node's type.
hipError_t hipGraphNodeSetEnabled(hipGraphExec_t hGraphExec, hipGraphNode_t hNode, unsigned int isEnabled)
Enables or disables the specified node in the given graphExec.
hipError_t hipGraphExecMemcpyNodeSetParams(hipGraphExec_t hGraphExec, hipGraphNode_t node, hipMemcpy3DParms *pNodeParams)
Sets the parameters for a memcpy node in the given graphExec.
hipError_t hipGraphHostNodeSetParams(hipGraphNode_t node, const hipHostNodeParams *pNodeParams)
Sets a host node's parameters.
hipError_t hipGraphMemAllocNodeGetParams(hipGraphNode_t node, hipMemAllocNodeParams *pNodeParams)
Returns parameters for memory allocation node.
hipError_t hipUserObjectRetain(hipUserObject_t object, unsigned int count __dparm(1))
Retain number of references to resource.
hipError_t hipGraphExecEventRecordNodeSetEvent(hipGraphExec_t hGraphExec, hipGraphNode_t hNode, hipEvent_t event)
Sets the event for an event record node in the given graphExec.
hipError_t hipGraphExternalSemaphoresWaitNodeSetParams(hipGraphNode_t hNode, const hipExternalSemaphoreWaitNodeParams *nodeParams)
Updates node parameters in the external semaphore wait node.
hipError_t hipGraphMemsetNodeGetParams(hipGraphNode_t node, hipMemsetParams *pNodeParams)
Gets a memset node's parameters.
hipError_t hipGraphChildGraphNodeGetGraph(hipGraphNode_t node, hipGraph_t *pGraph)
Gets a handle to the embedded graph of a child graph node.
hipError_t hipGraphKernelNodeGetAttribute(hipGraphNode_t hNode, hipKernelNodeAttrID attr, hipKernelNodeAttrValue *value)
Gets a node attribute.
hipError_t hipThreadExchangeStreamCaptureMode(hipStreamCaptureMode *mode)
Swaps the stream capture mode of a thread.
hipError_t hipGraphLaunch(hipGraphExec_t graphExec, hipStream_t stream)
launches an executable graph in a stream
hipError_t hipGraphDebugDotPrint(hipGraph_t graph, const char *path, unsigned int flags)
Write a DOT file describing graph structure.
hipError_t hipGraphExecMemsetNodeSetParams(hipGraphExec_t hGraphExec, hipGraphNode_t node, const hipMemsetParams *pNodeParams)
Sets the parameters for a memset node in the given graphExec.
hipError_t hipGraphGetEdges(hipGraph_t graph, hipGraphNode_t *from, hipGraphNode_t *to, size_t *numEdges)
Returns a graph's dependency edges.
hipError_t hipGraphAddKernelNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, const hipKernelNodeParams *pNodeParams)
Creates a kernel execution node and adds it to a graph.
hipError_t hipGraphExecEventWaitNodeSetEvent(hipGraphExec_t hGraphExec, hipGraphNode_t hNode, hipEvent_t event)
Sets the event for an event record node in the given graphExec.
hipError_t hipGraphAddExternalSemaphoresWaitNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, const hipExternalSemaphoreWaitNodeParams *nodeParams)
Creates a external semaphor wait node and adds it to a graph.
hipError_t hipGraphEventWaitNodeSetEvent(hipGraphNode_t node, hipEvent_t event)
Sets an event wait node's event.
hipError_t hipGraphEventRecordNodeSetEvent(hipGraphNode_t node, hipEvent_t event)
Sets an event record node's event.
hipError_t hipGraphEventWaitNodeGetEvent(hipGraphNode_t node, hipEvent_t *event_out)
Returns the event associated with an event wait node.
hipError_t hipGraphKernelNodeSetAttribute(hipGraphNode_t hNode, hipKernelNodeAttrID attr, const hipKernelNodeAttrValue *value)
Sets a node attribute.
hipError_t hipGraphMemsetNodeSetParams(hipGraphNode_t node, const hipMemsetParams *pNodeParams)
Sets a memset node's parameters.
hipError_t hipGraphEventRecordNodeGetEvent(hipGraphNode_t node, hipEvent_t *event_out)
Returns the event associated with an event record node.
hipError_t hipGraphAddEmptyNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies)
Creates an empty node and adds it to a graph.
hipError_t hipGraphAddMemcpyNodeToSymbol(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, const void *symbol, const void *src, size_t count, size_t offset, hipMemcpyKind kind)
Creates a memcpy node to copy to a symbol on the device and adds it to a graph.
hipError_t hipGraphDestroy(hipGraph_t graph)
Destroys a graph.
hipError_t hipGraphExecMemcpyNodeSetParamsFromSymbol(hipGraphExec_t hGraphExec, hipGraphNode_t node, void *dst, const void *symbol, size_t count, size_t offset, hipMemcpyKind kind)
Sets the parameters for a memcpy node in the given graphExec to copy from a symbol on the.
hipError_t hipGraphAddMemAllocNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, hipMemAllocNodeParams *pNodeParams)
Creates a memory allocation node and adds it to a graph.
hipError_t hipGraphAddHostNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, const hipHostNodeParams *pNodeParams)
Creates a host execution node and adds it to a graph.
hipError_t hipGraphGetNodes(hipGraph_t graph, hipGraphNode_t *nodes, size_t *numNodes)
Returns graph nodes.
hipError_t hipGraphAddMemsetNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, const hipMemsetParams *pMemsetParams)
Creates a memset node and adds it to a graph.
hipError_t hipGraphInstantiate(hipGraphExec_t *pGraphExec, hipGraph_t graph, hipGraphNode_t *pErrorNode, char *pLogBuffer, size_t bufferSize)
Creates an executable graph from a graph.
hipError_t hipDeviceGraphMemTrim(int device)
Free unused memory on specific device used for graph back to OS.
hipError_t hipGraphClone(hipGraph_t *pGraphClone, hipGraph_t originalGraph)
Clones a graph.
hipError_t hipGraphGetRootNodes(hipGraph_t graph, hipGraphNode_t *pRootNodes, size_t *pNumRootNodes)
Returns graph's root nodes.
hipError_t hipGraphAddMemcpyNode(hipGraphNode_t *pGraphNode, hipGraph_t graph, const hipGraphNode_t *pDependencies, size_t numDependencies, const hipMemcpy3DParms *pCopyParams)
Creates a memcpy node and adds it to a graph.
hipError_t hipStreamGetCaptureInfo_v2(hipStream_t stream, hipStreamCaptureStatus *captureStatus_out, unsigned long long *id_out __dparm(0), hipGraph_t *graph_out __dparm(0), const hipGraphNode_t **dependencies_out __dparm(0), size_t *numDependencies_out __dparm(0))
Get stream's capture state.
hipError_t hipGraphReleaseUserObject(hipGraph_t graph, hipUserObject_t object, unsigned int count __dparm(1))
Release user object from graphs.
hipError_t hipMemcpyToSymbol(const void *symbol, const void *src, size_t sizeBytes, size_t offset __dparm(0), hipMemcpyKind kind __dparm(hipMemcpyHostToDevice))
Copies data to the given symbol on the device. Symbol HIP APIs allow a kernel to define a device-side...
hipError_t hipMemcpyAtoH(void *dst, hipArray *srcArray, size_t srcOffset, size_t count)
Copies data between host and device.
hipError_t hipHostAlloc(void **ptr, size_t size, unsigned int flags)
Allocate device accessible page locked host memory [Deprecated].
hipError_t hipMemcpy2DToArrayAsync(hipArray *dst, size_t wOffset, size_t hOffset, const void *src, size_t spitch, size_t width, size_t height, hipMemcpyKind kind, hipStream_t stream __dparm(0))
Copies data between host and device.
hipError_t hipMemsetD16Async(hipDeviceptr_t dest, unsigned short value, size_t count, hipStream_t stream __dparm(0))
Fills the first sizeBytes bytes of the memory area pointed to by dest with the constant short value v...
hipError_t hipFreeHost(void *ptr)
Free memory allocated by the hcc hip host memory allocation API [Deprecated].
hipError_t hipMemcpyFromSymbol(void *dst, const void *symbol, size_t sizeBytes, size_t offset __dparm(0), hipMemcpyKind kind __dparm(hipMemcpyDeviceToHost))
Copies data from the given symbol on the device.
hipError_t hipHostFree(void *ptr)
Free memory allocated by the hcc hip host memory allocation API This API performs an implicit hipDevi...
hipError_t hipMemcpyToArray(hipArray *dst, size_t wOffset, size_t hOffset, const void *src, size_t count, hipMemcpyKind kind)
Copies data between host and device.
hipError_t hipMemcpy2DFromArrayAsync(void *dst, size_t dpitch, hipArray_const_t src, size_t wOffset, size_t hOffset, size_t width, size_t height, hipMemcpyKind kind, hipStream_t stream __dparm(0))
Copies data between host and device asynchronously.
hipError_t hipMemGetInfo(size_t *free, size_t *total)
Query memory info.
hipError_t hipExtMallocWithFlags(void **ptr, size_t sizeBytes, unsigned int flags)
Allocate memory on the default accelerator.
hipError_t hipMemcpy3D(const struct hipMemcpy3DParms *p)
Copies data between host and device.
hipError_t hipMemcpy3DAsync(const struct hipMemcpy3DParms *p, hipStream_t stream __dparm(0))
Copies data between host and device asynchronously.
hipError_t hipMemset3D(hipPitchedPtr pitchedDevPtr, int value, hipExtent extent)
Fills synchronously the memory area pointed to by pitchedDevPtr with the constant value.
hipError_t hipModuleGetGlobal(hipDeviceptr_t *dptr, size_t *bytes, hipModule_t hmod, const char *name)
Returns a global pointer from a module. Returns in *dptr and *bytes the pointer and size of the globa...
hipError_t hipMemsetD16(hipDeviceptr_t dest, unsigned short value, size_t count)
Fills the first sizeBytes bytes of the memory area pointed to by dest with the constant short value v...
hipError_t hipMemcpyFromSymbolAsync(void *dst, const void *symbol, size_t sizeBytes, size_t offset, hipMemcpyKind kind, hipStream_t stream __dparm(0))
Copies data from the given symbol on the device asynchronously.
hipError_t hipMalloc(void **ptr, size_t size)
Allocate memory on the default accelerator.
hipError_t hipHostUnregister(void *hostPtr)
Un-register host pointer.
hipError_t hipHostGetFlags(unsigned int *flagsPtr, void *hostPtr)
Return flags associated with host pointer.
hipError_t hipMemsetD32(hipDeviceptr_t dest, int value, size_t count)
Fills the memory area pointed to by dest with the constant integer value for specified number of time...
hipError_t hipDrvMemcpy3DAsync(const HIP_MEMCPY3D *pCopy, hipStream_t stream)
Copies data between host and device asynchronously.
hipError_t hipMallocHost(void **ptr, size_t size)
Allocate pinned host memory [Deprecated].
hipError_t hipArrayGetDescriptor(HIP_ARRAY_DESCRIPTOR *pArrayDescriptor, hipArray *array)
Gets a 1D or 2D array descriptor.
hipError_t hipMalloc3DArray(hipArray **array, const struct hipChannelFormatDesc *desc, struct hipExtent extent, unsigned int flags)
Allocate an array on the device.
hipError_t hipFree(void *ptr)
Free memory allocated by the hcc hip memory allocation API. This API performs an implicit hipDeviceSy...
hipError_t hipMemcpyToSymbolAsync(const void *symbol, const void *src, size_t sizeBytes, size_t offset, hipMemcpyKind kind, hipStream_t stream __dparm(0))
Copies data to the given symbol on the device asynchronously.
hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, const void *ptr)
Return attributes for the specified pointer.
hipError_t hipMallocPitch(void **ptr, size_t *pitch, size_t width, size_t height)
hipError_t hipMemcpyDtoD(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes)
Copy data from Device to Device.
hipError_t hipMemcpy2D(void *dst, size_t dpitch, const void *src, size_t spitch, size_t width, size_t height, hipMemcpyKind kind)
Copies data between host and device.
hipError_t hipMemcpyFromArray(void *dst, hipArray_const_t srcArray, size_t wOffset, size_t hOffset, size_t count, hipMemcpyKind kind)
Copies data between host and device.
hipError_t hipHostGetDevicePointer(void **devPtr, void *hstPtr, unsigned int flags)
Get Device pointer from Host Pointer allocated through hipHostMalloc.
hipError_t hipMemcpyParam2DAsync(const hip_Memcpy2D *pCopy, hipStream_t stream __dparm(0))
Copies memory for 2D arrays.
hipError_t hipMemsetD8Async(hipDeviceptr_t dest, unsigned char value, size_t count, hipStream_t stream __dparm(0))
Fills the first sizeBytes bytes of the memory area pointed to by dest with the constant byte value va...
hipError_t hipMemcpy2DFromArray(void *dst, size_t dpitch, hipArray_const_t src, size_t wOffset, size_t hOffset, size_t width, size_t height, hipMemcpyKind kind)
Copies data between host and device.
hipError_t hipDrvMemcpy3D(const HIP_MEMCPY3D *pCopy)
Copies data between host and device.
hipError_t hipMemcpyHtoD(hipDeviceptr_t dst, void *src, size_t sizeBytes)
Copy data from Host to Device.
hipError_t hipMemcpyParam2D(const hip_Memcpy2D *pCopy)
Copies memory for 2D arrays.
hipError_t hipMemsetAsync(void *dst, int value, size_t sizeBytes, hipStream_t stream __dparm(0))
Fills the first sizeBytes bytes of the memory area pointed to by dev with the constant byte value val...
hipError_t hipMemset3DAsync(hipPitchedPtr pitchedDevPtr, int value, hipExtent extent, hipStream_t stream __dparm(0))
Fills asynchronously the memory area pointed to by pitchedDevPtr with the constant value.
hipError_t hipHostMalloc(void **ptr, size_t size, unsigned int flags)
Allocates device accessible page locked (pinned) host memory.
hipError_t hipMemcpy2DToArray(hipArray *dst, size_t wOffset, size_t hOffset, const void *src, size_t spitch, size_t width, size_t height, hipMemcpyKind kind)
Copies data between host and device.
hipError_t hipMemcpyAsync(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream __dparm(0))
Copy data from src to dst asynchronously.
hipError_t hipArray3DCreate(hipArray **array, const HIP_ARRAY3D_DESCRIPTOR *pAllocateArray)
Create a 3D array memory pointer on the device.
hipError_t hipHostRegister(void *hostPtr, size_t sizeBytes, unsigned int flags)
Register host memory so it can be accessed from the current device.
hipError_t hipMemcpy(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind)
Copy data from src to dst.
hipError_t hipMemset(void *dst, int value, size_t sizeBytes)
Fills the first sizeBytes bytes of the memory area pointed to by dest with the constant byte value va...
hipError_t hipMemcpy2DAsync(void *dst, size_t dpitch, const void *src, size_t spitch, size_t width, size_t height, hipMemcpyKind kind, hipStream_t stream __dparm(0))
Copies data between host and device.
hipError_t hipDrvPointerGetAttributes(unsigned int numAttributes, hipPointer_attribute *attributes, void **data, hipDeviceptr_t ptr)
Returns information about the specified pointer.[BETA].
hipError_t hipMalloc3D(hipPitchedPtr *pitchedDevPtr, hipExtent extent)
Create a 3D memory pointer on the device.
hipError_t hipMallocArray(hipArray **array, const hipChannelFormatDesc *desc, size_t width, size_t height __dparm(0), unsigned int flags __dparm(hipArrayDefault))
Allocate an array on the device.
hipError_t hipPointerSetAttribute(const void *value, hipPointer_attribute attribute, hipDeviceptr_t ptr)
Sets information on the specified pointer.[BETA].
hipError_t hipMemcpyDtoDAsync(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream)
Copy data from Device to Device asynchronously.
hipError_t hipMemAllocPitch(hipDeviceptr_t *dptr, size_t *pitch, size_t widthInBytes, size_t height, unsigned int elementSizeBytes)
hipError_t hipMemsetD8(hipDeviceptr_t dest, unsigned char value, size_t count)
Fills the first sizeBytes bytes of the memory area pointed to by dest with the constant byte value va...
hipError_t hipMemcpyDtoHAsync(void *dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream)
Copy data from Device to Host asynchronously.
hipError_t hipMemset2DAsync(void *dst, size_t pitch, int value, size_t width, size_t height, hipStream_t stream __dparm(0))
Fills asynchronously the memory area pointed to by dst with the constant value.
hipError_t hipArrayGetInfo(hipChannelFormatDesc *desc, hipExtent *extent, unsigned int *flags, hipArray *array)
Gets info about the specified array.
hipError_t hipArray3DGetDescriptor(HIP_ARRAY3D_DESCRIPTOR *pArrayDescriptor, hipArray *array)
Gets a 3D array descriptor.
hipError_t hipMemset2D(void *dst, size_t pitch, int value, size_t width, size_t height)
Fills the memory area pointed to by dst with the constant value.
hipError_t hipFreeArray(hipArray *array)
Frees an array on the device.
hipError_t hipMemcpyHtoDAsync(hipDeviceptr_t dst, void *src, size_t sizeBytes, hipStream_t stream)
Copy data from Host to Device asynchronously.
hipError_t hipGetSymbolSize(size_t *size, const void *symbol)
Gets the size of the given symbol on the device.
hipError_t hipMemcpyDtoH(void *dst, hipDeviceptr_t src, size_t sizeBytes)
Copy data from Device to Host.
hipError_t hipMemcpyWithStream(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind, hipStream_t stream)
Memory copy on the stream. It allows single or multiple devices to do memory copy on single or multip...
hipError_t hipGetSymbolAddress(void **devPtr, const void *symbol)
Gets device pointer associated with symbol on the device.
hipError_t hipArrayCreate(hipArray **pHandle, const HIP_ARRAY_DESCRIPTOR *pAllocateArray)
Create an array memory pointer on the device.
hipError_t hipMemAllocHost(void **ptr, size_t size)
Allocate pinned host memory [Deprecated].
hipError_t hipPointerGetAttribute(void *data, hipPointer_attribute attribute, hipDeviceptr_t ptr)
Returns information about the specified pointer.[BETA].
hipError_t hipArrayDestroy(hipArray *array)
Destroy an array memory pointer on the device.
hipError_t hipMemcpyHtoA(hipArray *dstArray, size_t dstOffset, const void *srcHost, size_t count)
Copies data between host and device.
hipError_t hipMemPtrGetInfo(void *ptr, size_t *size)
Get allocated memory size via memory pointer.
hipError_t hipMemsetD32Async(hipDeviceptr_t dst, int value, size_t count, hipStream_t stream __dparm(0))
Fills the memory area pointed to by dev with the constant integer value for specified number of times...
hipError_t hipStreamAttachMemAsync(hipStream_t stream, void *dev_ptr, size_t length __dparm(0), unsigned int flags __dparm(hipMemAttachSingle))
Attach memory to a stream asynchronously in HIP.
hipError_t hipMemRangeGetAttributes(void **data, size_t *data_sizes, hipMemRangeAttribute *attributes, size_t num_attributes, const void *dev_ptr, size_t count)
Query attributes of a given memory range in HIP.
hipError_t hipMemAdvise(const void *dev_ptr, size_t count, hipMemoryAdvise advice, int device)
Advise about the usage of a given memory range to HIP.
hipError_t hipMemPrefetchAsync(const void *dev_ptr, size_t count, int device, hipStream_t stream __dparm(0))
Prefetches memory to the specified destination device using HIP.
hipError_t hipMemRangeGetAttribute(void *data, size_t data_size, hipMemRangeAttribute attribute, const void *dev_ptr, size_t count)
Query an attribute of a given memory range in HIP.
hipError_t hipMallocManaged(void **dev_ptr, size_t size, unsigned int flags __dparm(hipMemAttachGlobal))
Allocates memory that will be automatically managed by HIP.
hipError_t hipModuleLaunchCooperativeKernelMultiDevice(hipFunctionLaunchParams *launchParamsList, unsigned int numDevices, unsigned int flags)
Launches kernels on multiple devices where thread blocks can cooperate and synchronize as they execut...
hipError_t hipFuncGetAttributes(struct hipFuncAttributes *attr, const void *func)
Find out attributes for a given function.
hipError_t hipModuleGetTexRef(textureReference **texRef, hipModule_t hmod, const char *name)
returns the handle of the texture reference with the name from the module.
hipError_t hipExtLaunchMultiKernelMultiDevice(hipLaunchParams *launchParamsList, int numDevices, unsigned int flags)
Launches kernels on multiple devices and guarantees all specified kernels are dispatched on respectiv...
hipError_t hipModuleLaunchKernel(hipFunction_t f, unsigned int gridDimX, unsigned int gridDimY, unsigned int gridDimZ, unsigned int blockDimX, unsigned int blockDimY, unsigned int blockDimZ, unsigned int sharedMemBytes, hipStream_t stream, void **kernelParams, void **extra)
launches kernel f with launch parameters and shared memory on stream with arguments passed to kernelp...
hipError_t hipModuleLoad(hipModule_t *module, const char *fname)
Loads code object from file into a module the currrent context.
hipError_t hipModuleLoadDataEx(hipModule_t *module, const void *image, unsigned int numOptions, hipJitOption *options, void **optionValues)
builds module from code object which resides in host memory. Image is pointer to that location....
hipError_t hipFuncGetAttribute(int *value, hipFunction_attribute attrib, hipFunction_t hfunc)
Find out a specific attribute for a given function.
hipError_t hipLaunchCooperativeKernelMultiDevice(hipLaunchParams *launchParamsList, int numDevices, unsigned int flags)
Launches kernels on multiple devices where thread blocks can cooperate and synchronize as they execut...
hipError_t hipModuleLaunchCooperativeKernel(hipFunction_t f, unsigned int gridDimX, unsigned int gridDimY, unsigned int gridDimZ, unsigned int blockDimX, unsigned int blockDimY, unsigned int blockDimZ, unsigned int sharedMemBytes, hipStream_t stream, void **kernelParams)
launches kernel f with launch parameters and shared memory on stream with arguments passed to kernelP...
hipError_t hipModuleGetFunction(hipFunction_t *function, hipModule_t module, const char *kname)
Function with kname will be extracted if present in module.
hipError_t hipLaunchCooperativeKernel(const void *f, dim3 gridDim, dim3 blockDimX, void **kernelParams, unsigned int sharedMemBytes, hipStream_t stream)
launches kernel f with launch parameters and shared memory on stream with arguments passed to kernelp...
hipError_t hipModuleLoadData(hipModule_t *module, const void *image)
builds module from code object which resides in host memory. Image is pointer to that location.
hipError_t hipModuleUnload(hipModule_t module)
Frees the module.
hipError_t hipOccupancyMaxActiveBlocksPerMultiprocessorWithFlags(int *numBlocks, const void *f, int blockSize, size_t dynSharedMemPerBlk, unsigned int flags __dparm(hipOccupancyDefault))
Returns occupancy for a device function.
hipError_t hipModuleOccupancyMaxPotentialBlockSize(int *gridSize, int *blockSize, hipFunction_t f, size_t dynSharedMemPerBlk, int blockSizeLimit)
determine the grid and block sizes to achieves maximum occupancy for a kernel
hipError_t hipOccupancyMaxActiveBlocksPerMultiprocessor(int *numBlocks, const void *f, int blockSize, size_t dynSharedMemPerBlk)
Returns occupancy for a device function.
hipError_t hipOccupancyMaxPotentialBlockSize(int *gridSize, int *blockSize, const void *f, size_t dynSharedMemPerBlk, int blockSizeLimit)
determine the grid and block sizes to achieves maximum occupancy for a kernel
hipError_t hipModuleOccupancyMaxActiveBlocksPerMultiprocessor(int *numBlocks, hipFunction_t f, int blockSize, size_t dynSharedMemPerBlk)
Returns occupancy for a device function.
hipError_t hipModuleOccupancyMaxActiveBlocksPerMultiprocessorWithFlags(int *numBlocks, hipFunction_t f, int blockSize, size_t dynSharedMemPerBlk, unsigned int flags)
Returns occupancy for a device function.
hipError_t hipModuleOccupancyMaxPotentialBlockSizeWithFlags(int *gridSize, int *blockSize, hipFunction_t f, size_t dynSharedMemPerBlk, int blockSizeLimit, unsigned int flags)
determine the grid and block sizes to achieves maximum occupancy for a kernel
hipError_t hipDeviceCanAccessPeer(int *canAccessPeer, int deviceId, int peerDeviceId)
Determine if a device can access a peer's memory.
hipError_t hipDeviceEnablePeerAccess(int peerDeviceId, unsigned int flags)
Enable direct access from current device's virtual address space to memory allocations physically loc...
hipError_t hipDeviceDisablePeerAccess(int peerDeviceId)
Disable direct access from current device's virtual address space to memory allocations physically lo...
hipError_t hipMemGetAddressRange(hipDeviceptr_t *pbase, size_t *psize, hipDeviceptr_t dptr)
Get information on memory allocations.
hipError_t hipProfilerStart()
Start recording of profiling information. When using this API, start the profiler with profiling disa...
hipError_t hipProfilerStop()
Stop recording of profiling information. When using this API, start the profiler with profiling disab...
hipError_t hipDeviceGetStreamPriorityRange(int *leastPriority, int *greatestPriority)
Returns numerical values that correspond to the least and greatest stream priority.
hipError_t hipStreamDestroy(hipStream_t stream)
Destroys the specified stream.
hipError_t hipStreamGetFlags(hipStream_t stream, unsigned int *flags)
Return flags associated with this stream.
hipError_t hipStreamAddCallback(hipStream_t stream, hipStreamCallback_t callback, void *userData, unsigned int flags)
Adds a callback to be called on the host after all currently enqueued items in the stream have comple...
void(* hipStreamCallback_t)(hipStream_t stream, hipError_t status, void *userData)
Definition hip_runtime_api.h:2295
hipError_t hipStreamGetDevice(hipStream_t stream, hipDevice_t *device)
Get the device assocaited with the stream.
hipError_t hipStreamQuery(hipStream_t stream)
Return hipSuccess if all of the operations in the specified stream have completed,...
hipError_t hipStreamSynchronize(hipStream_t stream)
Wait for all commands in stream to complete.
hipError_t hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int flags)
Make the specified compute stream wait for an event.
hipError_t hipStreamCreateWithPriority(hipStream_t *stream, unsigned int flags, int priority)
Create an asynchronous stream with the specified priority.
hipError_t hipExtStreamCreateWithCUMask(hipStream_t *stream, uint32_t cuMaskSize, const uint32_t *cuMask)
Create an asynchronous stream with the specified CU mask.
hipError_t hipStreamGetPriority(hipStream_t stream, int *priority)
Query the priority of a stream.
hipError_t hipExtStreamGetCUMask(hipStream_t stream, uint32_t cuMaskSize, uint32_t *cuMask)
Get CU mask associated with an asynchronous stream.
hipError_t hipStreamCreateWithFlags(hipStream_t *stream, unsigned int flags)
Create an asynchronous stream.
hipError_t hipStreamCreate(hipStream_t *stream)
Create an asynchronous stream.
hipError_t hipStreamWriteValue64(hipStream_t stream, void *ptr, uint64_t value, unsigned int flags)
Enqueues a write command to the stream.[BETA].
hipError_t hipStreamWriteValue32(hipStream_t stream, void *ptr, uint32_t value, unsigned int flags)
Enqueues a write command to the stream.[BETA].
hipError_t hipStreamWaitValue32(hipStream_t stream, void *ptr, uint32_t value, unsigned int flags, uint32_t mask __dparm(0xFFFFFFFF))
Enqueues a wait command to the stream.[BETA].
hipError_t hipStreamWaitValue64(hipStream_t stream, void *ptr, uint64_t value, unsigned int flags, uint64_t mask __dparm(0xFFFFFFFFFFFFFFFF))
Enqueues a wait command to the stream.[BETA].
hipError_t hipMemPoolImportPointer(void **dev_ptr, hipMemPool_t mem_pool, hipMemPoolPtrExportData *export_data)
Import a memory pool allocation from another process.
hipError_t hipMemPoolImportFromShareableHandle(hipMemPool_t *mem_pool, void *shared_handle, hipMemAllocationHandleType handle_type, unsigned int flags)
Imports a memory pool from a shared handle.
hipError_t hipMallocFromPoolAsync(void **dev_ptr, size_t size, hipMemPool_t mem_pool, hipStream_t stream)
Allocates memory from a specified pool with stream ordered semantics.
hipError_t hipFreeAsync(void *dev_ptr, hipStream_t stream)
Frees memory with stream ordered semantics.
hipError_t hipMemPoolTrimTo(hipMemPool_t mem_pool, size_t min_bytes_to_hold)
Releases freed memory back to the OS.
hipError_t hipMemPoolCreate(hipMemPool_t *mem_pool, const hipMemPoolProps *pool_props)
Creates a memory pool.
hipError_t hipMemPoolSetAccess(hipMemPool_t mem_pool, const hipMemAccessDesc *desc_list, size_t count)
Controls visibility of the specified pool between devices.
hipError_t hipMemPoolDestroy(hipMemPool_t mem_pool)
Destroys the specified memory pool.
hipError_t hipMemPoolSetAttribute(hipMemPool_t mem_pool, hipMemPoolAttr attr, void *value)
Sets attributes of a memory pool.
hipError_t hipMemPoolGetAttribute(hipMemPool_t mem_pool, hipMemPoolAttr attr, void *value)
Gets attributes of a memory pool.
hipError_t hipMemPoolExportToShareableHandle(void *shared_handle, hipMemPool_t mem_pool, hipMemAllocationHandleType handle_type, unsigned int flags)
Exports a memory pool to the requested handle type.
hipError_t hipMemPoolExportPointer(hipMemPoolPtrExportData *export_data, void *dev_ptr)
Export data to share a memory pool allocation between processes.
hipError_t hipMallocAsync(void **dev_ptr, size_t size, hipStream_t stream)
Allocates memory with stream ordered semantics.
hipError_t hipMemPoolGetAccess(hipMemAccessFlags *flags, hipMemPool_t mem_pool, hipMemLocation *location)
Returns the accessibility of a pool from a device.
hipError_t hipDestroySurfaceObject(hipSurfaceObject_t surfaceObject)
Destroy a surface object.
hipError_t hipCreateSurfaceObject(hipSurfaceObject_t *pSurfObject, const hipResourceDesc *pResDesc)
Create a surface object.
hipError_t hipTexRefGetFilterMode(enum hipTextureFilterMode *pfm, const textureReference *texRef)
Gets filter mode for a texture reference.
hipError_t hipBindTextureToMipmappedArray(const textureReference *tex, hipMipmappedArray_const_t mipmappedArray, const hipChannelFormatDesc *desc)
Binds a mipmapped array to a texture.
hipError_t hipTexRefGetMipmapLevelClamp(float *pminMipmapLevelClamp, float *pmaxMipmapLevelClamp, const textureReference *texRef)
Gets the minimum and maximum mipmap level clamps for a texture reference.
hipError_t hipTexRefGetAddressMode(enum hipTextureAddressMode *pam, const textureReference *texRef, int dim)
Gets the address mode for a texture reference.
hipError_t hipTexRefGetFlags(unsigned int *pFlags, const textureReference *texRef)
Gets flags for a texture reference.
hipError_t hipTexRefSetAddress(size_t *ByteOffset, textureReference *texRef, hipDeviceptr_t dptr, size_t bytes)
Sets an bound address for a texture reference.
hipError_t hipTexRefSetFilterMode(textureReference *texRef, enum hipTextureFilterMode fm)
Set filter mode for a texture reference.
hipError_t hipTexRefSetMipmappedArray(textureReference *texRef, struct hipMipmappedArray *mipmappedArray, unsigned int Flags)
Binds mipmapped array to a texture reference.
hipError_t hipTexRefGetMipmapFilterMode(enum hipTextureFilterMode *pfm, const textureReference *texRef)
Gets the mipmap filter mode for a texture reference.
hipError_t hipGetTextureAlignmentOffset(size_t *offset, const textureReference *texref)
Get the offset of the alignment in a texture.
hipError_t hipBindTexture2D(size_t *offset, const textureReference *tex, const void *devPtr, const hipChannelFormatDesc *desc, size_t width, size_t height, size_t pitch)
Binds a 2D memory area to a texture.
hipError_t hipTexRefGetMaxAnisotropy(int *pmaxAnsio, const textureReference *texRef)
Gets the maximum anisotropy for a texture reference.
hipError_t hipUnbindTexture(const textureReference *tex)
Unbinds a texture.
hipError_t hipTexRefSetFormat(textureReference *texRef, hipArray_Format fmt, int NumPackedComponents)
Set format for a texture reference.
hipError_t hipTexRefGetAddress(hipDeviceptr_t *dev_ptr, const textureReference *texRef)
Gets the the address for a texture reference.
hipError_t hipTexRefSetFlags(textureReference *texRef, unsigned int Flags)
Set flags for a texture reference.
hipError_t hipTexRefSetAddressMode(textureReference *texRef, int dim, enum hipTextureAddressMode am)
Sets address mode for a texture reference.
hipError_t hipTexRefGetMipMappedArray(hipMipmappedArray_t *pArray, const textureReference *texRef)
Gets the mipmapped array bound to a texture reference.
hipError_t hipTexRefSetMaxAnisotropy(textureReference *texRef, unsigned int maxAniso)
Sets the maximum anisotropy for a texture reference.
hipError_t hipTexRefSetMipmapLevelClamp(textureReference *texRef, float minMipMapLevelClamp, float maxMipMapLevelClamp)
Sets mipmap level clamp for a texture reference.
hipError_t hipTexRefSetBorderColor(textureReference *texRef, float *pBorderColor)
Sets border color for a texture reference.
hipError_t hipTexRefSetArray(textureReference *tex, hipArray_const_t array, unsigned int flags)
Binds an array as a texture reference.
hipError_t hipTexRefSetMipmapLevelBias(textureReference *texRef, float bias)
Sets mipmap level bias for a texture reference.
hipError_t hipTexRefGetMipmapLevelBias(float *pbias, const textureReference *texRef)
Gets the mipmap level bias for a texture reference.
hipError_t hipBindTextureToArray(const textureReference *tex, hipArray_const_t array, const hipChannelFormatDesc *desc)
Binds a memory area to a texture.
hipError_t hipBindTexture(size_t *offset, const textureReference *tex, const void *devPtr, const hipChannelFormatDesc *desc, size_t size __dparm(UINT_MAX))
Binds a memory area to a texture.
hipError_t hipGetTextureReference(const textureReference **texref, const void *symbol)
Gets the texture reference related with the symbol.
hipError_t hipTexRefSetAddress2D(textureReference *texRef, const HIP_ARRAY_DESCRIPTOR *desc, hipDeviceptr_t dptr, size_t Pitch)
Set a bind an address as a 2D texture reference.
hipError_t hipTexRefSetMipmapFilterMode(textureReference *texRef, enum hipTextureFilterMode fm)
Sets mipmap filter mode for a texture reference.
hipError_t hipTexRefGetFormat(hipArray_Format *pFormat, int *pNumChannels, const textureReference *texRef)
Gets texture format for a texture reference.
hipError_t hipFreeMipmappedArray(hipMipmappedArray_t mipmappedArray)
Frees a mipmapped array on the device.
hipError_t hipGetMipmappedArrayLevel(hipArray_t *levelArray, hipMipmappedArray_const_t mipmappedArray, unsigned int level)
Gets a mipmap level of a HIP mipmapped array.
hipError_t hipMipmappedArrayGetLevel(hipArray_t *pLevelArray, hipMipmappedArray_t hMipMappedArray, unsigned int level)
Get a mipmapped array on a mipmapped level.
hipError_t hipMipmappedArrayDestroy(hipMipmappedArray_t hMipmappedArray)
Destroy a mipmapped array.
hipError_t hipTexObjectCreate(hipTextureObject_t *pTexObject, const HIP_RESOURCE_DESC *pResDesc, const HIP_TEXTURE_DESC *pTexDesc, const HIP_RESOURCE_VIEW_DESC *pResViewDesc)
Creates a texture object.
hipError_t hipGetTextureObjectResourceDesc(hipResourceDesc *pResDesc, hipTextureObject_t textureObject)
Gets resource descriptor for the texture object.
hipError_t hipGetTextureObjectResourceViewDesc(struct hipResourceViewDesc *pResViewDesc, hipTextureObject_t textureObject)
Gets resource view descriptor for the texture object.
hipError_t hipTexObjectGetTextureDesc(HIP_TEXTURE_DESC *pTexDesc, hipTextureObject_t texObject)
Gets texture descriptor of a texture object.
hipError_t hipCreateTextureObject(hipTextureObject_t *pTexObject, const hipResourceDesc *pResDesc, const hipTextureDesc *pTexDesc, const struct hipResourceViewDesc *pResViewDesc)
Creates a texture object.
hipError_t hipTexObjectDestroy(hipTextureObject_t texObject)
Destroys a texture object.
hipError_t hipMipmappedArrayCreate(hipMipmappedArray_t *pHandle, HIP_ARRAY3D_DESCRIPTOR *pMipmappedArrayDesc, unsigned int numMipmapLevels)
Create a mipmapped array.
hipError_t hipGetChannelDesc(hipChannelFormatDesc *desc, hipArray_const_t array)
Gets the channel descriptor in an array.
hipError_t hipTexObjectGetResourceDesc(HIP_RESOURCE_DESC *pResDesc, hipTextureObject_t texObject)
Gets resource descriptor of a texture object.
hipError_t hipTexObjectGetResourceViewDesc(HIP_RESOURCE_VIEW_DESC *pResViewDesc, hipTextureObject_t texObject)
Gets resource view descriptor of a texture object.
hipError_t hipDestroyTextureObject(hipTextureObject_t textureObject)
Destroys a texture object.
hipError_t hipMallocMipmappedArray(hipMipmappedArray_t *mipmappedArray, const struct hipChannelFormatDesc *desc, struct hipExtent extent, unsigned int numLevels, unsigned int flags __dparm(0))
Allocate a mipmapped array on the device.
hipError_t hipGetTextureObjectTextureDesc(hipTextureDesc *pTexDesc, hipTextureObject_t textureObject)
Gets texture descriptor for the texture object.
hipError_t hipMemAddressFree(void *devPtr, size_t size)
Frees an address range reservation made via hipMemAddressReserve.
hipError_t hipMemExportToShareableHandle(void *shareableHandle, hipMemGenericAllocationHandle_t handle, hipMemAllocationHandleType handleType, unsigned long long flags)
Exports an allocation to a requested shareable handle type.
hipError_t hipMemGetAccess(unsigned long long *flags, const hipMemLocation *location, void *ptr)
Get the access flags set for the given location and ptr.
hipError_t hipMemSetAccess(void *ptr, size_t size, const hipMemAccessDesc *desc, size_t count)
Set the access flags for each location specified in desc for the given virtual address range.
hipError_t hipMemRelease(hipMemGenericAllocationHandle_t handle)
Release a memory handle representing a memory allocation which was previously allocated through hipMe...
hipError_t hipMemUnmap(void *ptr, size_t size)
Unmap memory allocation of a given address range.
hipError_t hipMemGetAllocationGranularity(size_t *granularity, const hipMemAllocationProp *prop, hipMemAllocationGranularity_flags option)
Calculates either the minimal or recommended granularity.
hipError_t hipMemRetainAllocationHandle(hipMemGenericAllocationHandle_t *handle, void *addr)
Returns the allocation handle of the backing memory allocation given the address.
hipError_t hipMemCreate(hipMemGenericAllocationHandle_t *handle, size_t size, const hipMemAllocationProp *prop, unsigned long long flags)
Creates a memory allocation described by the properties and size.
hipError_t hipMemMap(void *ptr, size_t size, size_t offset, hipMemGenericAllocationHandle_t handle, unsigned long long flags)
Maps an allocation handle to a reserved virtual address range.
hipError_t hipMemAddressReserve(void **ptr, size_t size, size_t alignment, void *addr, unsigned long long flags)
Reserves an address range.
hipError_t hipMemGetAllocationPropertiesFromHandle(hipMemAllocationProp *prop, hipMemGenericAllocationHandle_t handle)
Retrieve the property structure of the given handle.
hipError_t hipMemMapArrayAsync(hipArrayMapInfo *mapInfoList, unsigned int count, hipStream_t stream)
Maps or unmaps subregions of sparse HIP arrays and sparse HIP mipmapped arrays.
hipError_t hipMemImportFromShareableHandle(hipMemGenericAllocationHandle_t *handle, void *osHandle, hipMemAllocationHandleType shHandleType)
Imports an allocation from a requested shareable handle type.
hipErrorIllegalAddress
Definition hip_runtime_api.h:278
hipErrorMemoryAllocation
Memory allocation error.
Definition hip_runtime_api.h:224
hipErrorNotFound
Not found.
Definition hip_runtime_api.h:273
hipErrorProfilerDisabled
Definition hip_runtime_api.h:229
hipErrorInvalidDevicePointer
Invalid Device Pointer.
Definition hip_runtime_api.h:236
hipErrorRuntimeMemory
Definition hip_runtime_api.h:327
hipErrorInvalidImage
Invalid image.
Definition hip_runtime_api.h:244
hipErrorPeerAccessUnsupported
Definition hip_runtime_api.h:261
hipErrorArrayIsMapped
Definition hip_runtime_api.h:251
hipErrorStreamCaptureIsolation
Definition hip_runtime_api.h:308
hipErrorInvalidKernelFile
In CUDA DRV, it is CUDA_ERROR_INVALID_PTX.
Definition hip_runtime_api.h:262
hipErrorNotMapped
Definition hip_runtime_api.h:255
hipErrorGraphExecUpdateFailure
Definition hip_runtime_api.h:321
hipErrorNotInitialized
Invalid not initialized.
Definition hip_runtime_api.h:225
hipErrorInvalidResourceHandle
Resource handle (hipEvent_t or hipStream_t) invalid.
Definition hip_runtime_api.h:271
hipErrorSharedObjectInitFailed
Failed to initialize shared object.
Definition hip_runtime_api.h:267
hipErrorInvalidPitchValue
Invalid pitch value.
Definition hip_runtime_api.h:234
hipErrorNoBinaryForGpu
Definition hip_runtime_api.h:253
hipErrorInvalidContext
Produced when input context is invalid.
Definition hip_runtime_api.h:245
hipErrorStreamCaptureUnsupported
Definition hip_runtime_api.h:299
hipErrorPeerAccessNotEnabled
Peer access was never enabled from the current device.
Definition hip_runtime_api.h:283
hipErrorOperatingSystem
Not the correct operating system.
Definition hip_runtime_api.h:268
hipErrorStreamCaptureUnjoined
Definition hip_runtime_api.h:306
hipErrorInitializationError
Definition hip_runtime_api.h:227
hipErrorNotMappedAsPointer
Definition hip_runtime_api.h:257
hipErrorPriorLaunchFailure
Definition hip_runtime_api.h:240
hipErrorContextIsDestroyed
The context is already destroyed.
Definition hip_runtime_api.h:286
hipErrorInvalidValue
Definition hip_runtime_api.h:220
hipErrorLaunchTimeOut
Timeout for the launch.
Definition hip_runtime_api.h:280
hipErrorUnmapFailed
Definition hip_runtime_api.h:250
hipErrorSetOnActiveProcess
The process is active.
Definition hip_runtime_api.h:285
hipErrorFileNotFound
the file is not found.
Definition hip_runtime_api.h:265
hipErrorLaunchFailure
An exception occurred on the device while executing a kernel.
Definition hip_runtime_api.h:292
hipErrorInvalidDevice
DeviceID must be in range from 0 to compute-devices.
Definition hip_runtime_api.h:243
hipErrorUnsupportedLimit
Unsupported limit.
Definition hip_runtime_api.h:259
hipErrorMapBufferObjectFailed
Produced when the IPC memory attach failed from ROCr.
Definition hip_runtime_api.h:249
hipErrorHostMemoryAlreadyRegistered
Produced when trying to lock a page-locked memory.
Definition hip_runtime_api.h:288
hipErrorMissingConfiguration
Definition hip_runtime_api.h:239
hipErrorSharedObjectSymbolNotFound
Definition hip_runtime_api.h:266
hipErrorECCNotCorrectable
Definition hip_runtime_api.h:258
hipErrorStreamCaptureWrongThread
Definition hip_runtime_api.h:317
hipErrorAssert
Produced when the kernel calls assert.
Definition hip_runtime_api.h:287
hipErrorContextAlreadyInUse
The context is already in use.
Definition hip_runtime_api.h:260
hipErrorNotReady
Definition hip_runtime_api.h:274
hipErrorInvalidSource
Invalid source.
Definition hip_runtime_api.h:264
hipErrorLaunchOutOfResources
Out of resources error.
Definition hip_runtime_api.h:279
hipErrorInvalidDeviceFunction
Invalid device function.
Definition hip_runtime_api.h:241
hipErrorCooperativeLaunchTooLarge
Definition hip_runtime_api.h:294
hipErrorAlreadyMapped
Definition hip_runtime_api.h:252
hipErrorProfilerNotInitialized
Definition hip_runtime_api.h:230
hipErrorProfilerAlreadyStarted
Definition hip_runtime_api.h:231
hipErrorNoDevice
Call to hipGetDeviceCount returned 0 devices.
Definition hip_runtime_api.h:242
@ HIP_ERROR_LAUNCH_OUT_OF_RESOURCES
Definition hip_runtime_api.h:42
@ HIP_ERROR_INVALID_VALUE
Definition hip_runtime_api.h:40
@ HIP_SUCCESS
Definition hip_runtime_api.h:39
@ HIP_ERROR_NOT_INITIALIZED
Definition hip_runtime_api.h:41
hipErrorMapFailed
Definition hip_runtime_api.h:247
hipErrorStreamCaptureInvalidated
Definition hip_runtime_api.h:301
hipErrorProfilerAlreadyStopped
Definition hip_runtime_api.h:232
hipErrorDeinitialized
Deinitialized.
Definition hip_runtime_api.h:228
hipErrorPeerAccessAlreadyEnabled
Peer access was already enabled from the current device.
Definition hip_runtime_api.h:281
hipErrorNotMappedAsArray
Definition hip_runtime_api.h:256
hipErrorNotSupported
Produced when the hip API is not supported/implemented.
Definition hip_runtime_api.h:298
hipErrorInvalidConfiguration
Invalide configuration.
Definition hip_runtime_api.h:233
hipErrorStreamCaptureImplicit
Definition hip_runtime_api.h:312
hipErrorStreamCaptureMerge
Definition hip_runtime_api.h:303
hipErrorInvalidGraphicsContext
Definition hip_runtime_api.h:263
hipErrorIllegalState
Resource required is not in a valid state to perform operation.
Definition hip_runtime_api.h:272
hipErrorOutOfMemory
out of memory range.
Definition hip_runtime_api.h:222
hipErrorInvalidSymbol
Invalid symbol.
Definition hip_runtime_api.h:235
hipSuccess
Successful completion.
Definition hip_runtime_api.h:219
hipErrorUnknown
Unknown error.
Definition hip_runtime_api.h:325
hipErrorInvalidHandle
Invalide handle.
Definition hip_runtime_api.h:269
hipErrorStreamCaptureUnmatched
The capture was not initiated in this stream.
Definition hip_runtime_api.h:305
hipErrorCapturedEvent
Definition hip_runtime_api.h:315
hipErrorAlreadyAcquired
Definition hip_runtime_api.h:254
hipErrorInsufficientDriver
Definition hip_runtime_api.h:238
hipErrorHostMemoryNotRegistered
Produced when trying to unlock a non-page-locked memory.
Definition hip_runtime_api.h:290
hipErrorRuntimeOther
Definition hip_runtime_api.h:329
hipErrorContextAlreadyCurrent
Definition hip_runtime_api.h:246
hipErrorInvalidMemcpyDirection
Invalid memory copy direction.
Definition hip_runtime_api.h:237
Definition driver_types.h:76
Definition driver_types.h:70
Definition driver_types.h:384
Definition driver_types.h:287
Definition driver_types.h:333
Definition driver_types.h:170
Definition hip_runtime_api.h:948
uint32_t y
y
Definition hip_runtime_api.h:950
uint32_t z
z
Definition hip_runtime_api.h:951
uint32_t x
x
Definition hip_runtime_api.h:949
Definition driver_types.h:98
Definition hip_runtime_api.h:1178
hipAccessProperty hitProp
Definition hip_runtime_api.h:1180
hipAccessProperty missProp
Definition hip_runtime_api.h:1182
float hitRatio
Definition hip_runtime_api.h:1181
void * base_ptr
Definition hip_runtime_api.h:1179
size_t num_bytes
Definition hip_runtime_api.h:1183
Definition driver_types.h:84
Definition hip_runtime_api.h:1346
unsigned int level
For mipmapped arrays must be a valid mipmap level. For arrays must be zero.
Definition hip_runtime_api.h:1355
struct hipArrayMapInfo::@26::@28 sparseLevel
unsigned int flags
flags for future use, must be zero now.
Definition hip_runtime_api.h:1377
union hipArrayMapInfo::@25 resource
hipArraySparseSubresourceType subresourceType
Sparse subresource type.
Definition hip_runtime_api.h:1352
unsigned int extentDepth
Depth in elements.
Definition hip_runtime_api.h:1362
unsigned int offsetX
X offset in elements.
Definition hip_runtime_api.h:1357
hipArray_t array
Definition hip_runtime_api.h:1350
hipResourceType resourceType
Resource type.
Definition hip_runtime_api.h:1347
hipMemOperationType memOperationType
Memory operation type.
Definition hip_runtime_api.h:1370
unsigned int reserved[2]
Reserved for future use, must be zero now.
Definition hip_runtime_api.h:1378
unsigned int extentWidth
Width in elements.
Definition hip_runtime_api.h:1360
union hipArrayMapInfo::@26 subresource
unsigned int extentHeight
Height in elements.
Definition hip_runtime_api.h:1361
unsigned int offsetY
Y offset in elements.
Definition hip_runtime_api.h:1358
hipMemHandleType memHandleType
Memory handle type.
Definition hip_runtime_api.h:1371
struct hipArrayMapInfo::@26::@29 miptail
hipMemGenericAllocationHandle_t memHandle
Definition hip_runtime_api.h:1373
hipMipmappedArray mipmap
Definition hip_runtime_api.h:1349
unsigned int layer
For layered arrays must be a valid layer index. Otherwise, must be zero.
Definition hip_runtime_api.h:1356
unsigned int deviceBitMask
Device ordinal bit mask.
Definition hip_runtime_api.h:1376
unsigned long long size
Extent in bytes.
Definition hip_runtime_api.h:1367
unsigned long long offset
Offset within mip tail.
Definition hip_runtime_api.h:1366
unsigned int offsetZ
Z offset in elements.
Definition hip_runtime_api.h:1359
Definition driver_types.h:49
Definition hip_runtime_api.h:54
unsigned hasSharedInt64Atomics
64-bit integer atomics for shared memory.
Definition hip_runtime_api.h:64
unsigned hasSharedInt32Atomics
32-bit integer atomics for shared memory.
Definition hip_runtime_api.h:58
unsigned hasFloatAtomicAdd
32-bit float atomic add in global and shared memory.
Definition hip_runtime_api.h:60
unsigned hasDoubles
Double-precision floating point.
Definition hip_runtime_api.h:67
unsigned hasWarpVote
Warp vote instructions (__any, __all).
Definition hip_runtime_api.h:70
unsigned hasWarpShuffle
Warp shuffle operations. (__shfl_*).
Definition hip_runtime_api.h:72
unsigned hasGlobalInt32Atomics
32-bit integer atomics for global memory.
Definition hip_runtime_api.h:56
unsigned hasGlobalFloatAtomicExch
32-bit float atomic exch for global memory.
Definition hip_runtime_api.h:57
unsigned hasDynamicParallelism
Dynamic parallelism.
Definition hip_runtime_api.h:82
unsigned hasSurfaceFuncs
Surface functions.
Definition hip_runtime_api.h:80
unsigned has3dGrid
Grid and group dims are 3D (rather than 2D).
Definition hip_runtime_api.h:81
unsigned hasFunnelShift
Funnel two words into one with shift&mask caps.
Definition hip_runtime_api.h:73
unsigned hasThreadFenceSystem
__threadfence_system.
Definition hip_runtime_api.h:76
unsigned hasGlobalInt64Atomics
64-bit integer atomics for global memory.
Definition hip_runtime_api.h:63
unsigned hasSyncThreadsExt
__syncthreads_count, syncthreads_and, syncthreads_or.
Definition hip_runtime_api.h:77
unsigned hasWarpBallot
Warp ballot instructions (__ballot).
Definition hip_runtime_api.h:71
unsigned hasSharedFloatAtomicExch
32-bit float atomic exch for shared memory.
Definition hip_runtime_api.h:59
Definition hip_runtime_api.h:96
int ECCEnabled
Device has ECC support enabled.
Definition hip_runtime_api.h:144
int maxTexture1DLinear
Maximum size for 1D textures bound to linear memory.
Definition hip_runtime_api.h:134
int pciBusID
PCI Bus ID.
Definition hip_runtime_api.h:124
int maxTexture1D
Maximum number of elements in 1D images.
Definition hip_runtime_api.h:135
int memoryBusWidth
Global memory bus width in bits.
Definition hip_runtime_api.h:107
int clockRate
Max clock frequency of the multiProcessors in khz.
Definition hip_runtime_api.h:105
int maxThreadsPerMultiProcessor
Maximum resident threads per multi-processor.
Definition hip_runtime_api.h:117
int l2CacheSize
L2 cache size.
Definition hip_runtime_api.h:116
size_t totalConstMem
Size of shared memory region (in bytes).
Definition hip_runtime_api.h:108
size_t memPitch
Maximum pitch in bytes allowed by memory copies.
Definition hip_runtime_api.h:140
size_t sharedMemPerBlock
Size of shared memory region (in bytes).
Definition hip_runtime_api.h:99
int cooperativeMultiDeviceLaunch
HIP device supports cooperative launch on multiple devices.
Definition hip_runtime_api.h:133
int pageableMemoryAccess
Definition hip_runtime_api.h:159
char name[256]
Device name.
Definition hip_runtime_api.h:97
size_t textureAlignment
Alignment requirement for textures.
Definition hip_runtime_api.h:141
int memoryClockRate
Max global memory clock frequency in khz.
Definition hip_runtime_api.h:106
int clockInstructionRate
Definition hip_runtime_api.h:119
int regsPerBlock
Registers per block.
Definition hip_runtime_api.h:100
int maxTexture2D[2]
Maximum dimensions (width, height) of 2D images, in image elements.
Definition hip_runtime_api.h:136
int cooperativeLaunch
HIP device supports cooperative launch.
Definition hip_runtime_api.h:132
int maxTexture3D[3]
Maximum dimensions (width, height, depth) of 3D images, in image elements.
Definition hip_runtime_api.h:137
int directManagedMemAccessFromHost
Host can directly access managed memory on the device without migration.
Definition hip_runtime_api.h:157
int cooperativeMultiDeviceUnmatchedFunc
Definition hip_runtime_api.h:146
int pageableMemoryAccessUsesHostPageTables
Device accesses pageable memory via the host's page tables.
Definition hip_runtime_api.h:161
unsigned int * hdpRegFlushCntl
Addres of HDP_REG_COHERENCY_FLUSH_CNTL register.
Definition hip_runtime_api.h:139
int maxThreadsDim[3]
Max number of threads in each dimension (XYZ) of a block.
Definition hip_runtime_api.h:103
int isMultiGpuBoard
1 if device is on a multi-GPU board, 0 if not.
Definition hip_runtime_api.h:127
int pciDomainID
PCI Domain ID.
Definition hip_runtime_api.h:123
size_t maxSharedMemoryPerMultiProcessor
Maximum Shared Memory Per Multiprocessor.
Definition hip_runtime_api.h:126
int cooperativeMultiDeviceUnmatchedGridDim
Definition hip_runtime_api.h:148
int minor
Definition hip_runtime_api.h:112
char gcnArchName[256]
AMD GCN Arch Name.
Definition hip_runtime_api.h:130
int kernelExecTimeoutEnabled
Run time limit for kernels executed on the device.
Definition hip_runtime_api.h:143
unsigned int * hdpMemFlushCntl
Addres of HDP_MEM_COHERENCY_FLUSH_CNTL register.
Definition hip_runtime_api.h:138
int concurrentManagedAccess
Device can coherently access managed memory concurrently with the CPU.
Definition hip_runtime_api.h:158
int integrated
APU vs dGPU.
Definition hip_runtime_api.h:131
int canMapHostMemory
Check whether HIP can map host memory.
Definition hip_runtime_api.h:128
int asicRevision
Revision of the GPU in this device.
Definition hip_runtime_api.h:155
size_t totalGlobalMem
Size of global memory region (in bytes).
Definition hip_runtime_api.h:98
int cooperativeMultiDeviceUnmatchedBlockDim
Definition hip_runtime_api.h:150
int concurrentKernels
Device can possibly execute multiple kernels concurrently.
Definition hip_runtime_api.h:122
int isLargeBar
1: if it is a large PCI bar device, else 0
Definition hip_runtime_api.h:154
int gcnArch
DEPRECATED: use gcnArchName instead.
Definition hip_runtime_api.h:129
int multiProcessorCount
Number of multi-processors (compute units).
Definition hip_runtime_api.h:115
int maxGridSize[3]
Max grid dimensions (XYZ).
Definition hip_runtime_api.h:104
int pciDeviceID
PCI Device ID.
Definition hip_runtime_api.h:125
int computeMode
Compute mode.
Definition hip_runtime_api.h:118
int major
Definition hip_runtime_api.h:109
int warpSize
Warp size.
Definition hip_runtime_api.h:101
int tccDriver
1:If device is Tesla device using TCC driver, else 0
Definition hip_runtime_api.h:145
int cooperativeMultiDeviceUnmatchedSharedMem
Definition hip_runtime_api.h:152
int maxThreadsPerBlock
Max work items per work group or workgroup max size.
Definition hip_runtime_api.h:102
hipDeviceArch_t arch
Architectural feature flags. New for HIP.
Definition hip_runtime_api.h:121
int managedMemory
Device supports allocating managed memory on this system.
Definition hip_runtime_api.h:156
size_t texturePitchAlignment
Pitch alignment requirement for texture references bound to pitched memory.
Definition hip_runtime_api.h:142
Definition driver_types.h:363
Definition hip_runtime_api.h:1003
unsigned int flags
Definition hip_runtime_api.h:1006
unsigned long long offset
Definition hip_runtime_api.h:1004
unsigned long long size
Definition hip_runtime_api.h:1005
Definition hip_runtime_api.h:991
unsigned int flags
Definition hip_runtime_api.h:1001
const void * name
Definition hip_runtime_api.h:997
int fd
Definition hip_runtime_api.h:994
unsigned long long size
Definition hip_runtime_api.h:1000
hipExternalMemoryHandleType type
Definition hip_runtime_api.h:992
void * handle
Definition hip_runtime_api.h:996
Definition hip_runtime_api.h:1015
unsigned int flags
Definition hip_runtime_api.h:1024
const void * name
Definition hip_runtime_api.h:1021
hipExternalSemaphoreHandleType type
Definition hip_runtime_api.h:1016
void * handle
Definition hip_runtime_api.h:1020
int fd
Definition hip_runtime_api.h:1018
Definition hip_runtime_api.h:1286
const hipExternalSemaphoreSignalParams * paramsArray
Total number of handles and parameters contained in extSemArray and paramsArray.
Definition hip_runtime_api.h:1290
hipExternalSemaphore_t * extSemArray
< Array containing external semaphore handles.
Definition hip_runtime_api.h:1288
unsigned int numExtSems
Definition hip_runtime_api.h:1292
Definition hip_runtime_api.h:1027
unsigned long long value
Definition hip_runtime_api.h:1030
unsigned int flags
Definition hip_runtime_api.h:1037
unsigned long long key
Definition hip_runtime_api.h:1033
Definition hip_runtime_api.h:1298
unsigned int numExtSems
Definition hip_runtime_api.h:1304
const hipExternalSemaphoreWaitParams * paramsArray
Total number of handles and parameters contained in extSemArray and paramsArray.
Definition hip_runtime_api.h:1302
hipExternalSemaphore_t * extSemArray
< Array containing external semaphore handles.
Definition hip_runtime_api.h:1300
Definition hip_runtime_api.h:1043
unsigned long long value
Definition hip_runtime_api.h:1046
unsigned int timeoutMs
Definition hip_runtime_api.h:1050
unsigned long long key
Definition hip_runtime_api.h:1049
unsigned int flags
Definition hip_runtime_api.h:1054
Definition hip_runtime_api.h:545
int cacheModeCA
Definition hip_runtime_api.h:547
int binaryVersion
Definition hip_runtime_api.h:546
size_t constSizeBytes
Definition hip_runtime_api.h:548
int preferredShmemCarveout
Definition hip_runtime_api.h:553
size_t sharedSizeBytes
Definition hip_runtime_api.h:555
int maxDynamicSharedSizeBytes
Definition hip_runtime_api.h:550
int ptxVersion
Definition hip_runtime_api.h:554
int numRegs
Definition hip_runtime_api.h:552
size_t localSizeBytes
Definition hip_runtime_api.h:549
int maxThreadsPerBlock
Definition hip_runtime_api.h:551
Definition hip_runtime_api.h:970
unsigned int gridDimX
Width(X) of grid in blocks.
Definition hip_runtime_api.h:972
unsigned int blockDimX
X dimension of each thread block.
Definition hip_runtime_api.h:975
hipFunction_t function
Kernel to launch.
Definition hip_runtime_api.h:971
hipStream_t hStream
Stream identifier.
Definition hip_runtime_api.h:979
unsigned int blockDimY
Y dimension of each thread block.
Definition hip_runtime_api.h:976
unsigned int gridDimY
Height(Y) of grid in blocks.
Definition hip_runtime_api.h:973
unsigned int gridDimZ
Depth(Z) of grid in blocks.
Definition hip_runtime_api.h:974
unsigned int sharedMemBytes
Shared memory.
Definition hip_runtime_api.h:978
unsigned int blockDimZ
Z dimension of each thread block.
Definition hip_runtime_api.h:977
void ** kernelParams
Kernel parameters.
Definition hip_runtime_api.h:980
Definition hip_runtime_api.h:1135
hipHostFn_t fn
Definition hip_runtime_api.h:1136
void * userData
Definition hip_runtime_api.h:1137
Definition hip_runtime_api.h:535
Definition hip_runtime_api.h:532
Definition hip_runtime_api.h:1139
dim3 gridDim
Definition hip_runtime_api.h:1143
unsigned int sharedMemBytes
Definition hip_runtime_api.h:1145
dim3 blockDim
Definition hip_runtime_api.h:1140
void ** kernelParams
Definition hip_runtime_api.h:1144
void ** extra
Definition hip_runtime_api.h:1141
void * func
Definition hip_runtime_api.h:1142
Definition hip_runtime_api.h:959
size_t sharedMem
Shared memory.
Definition hip_runtime_api.h:964
void ** args
Arguments.
Definition hip_runtime_api.h:963
dim3 blockDim
Block dimentions.
Definition hip_runtime_api.h:962
dim3 gridDim
Grid dimentions.
Definition hip_runtime_api.h:961
void * func
Device function symbol.
Definition hip_runtime_api.h:960
hipStream_t stream
Stream identifier.
Definition hip_runtime_api.h:965
Definition hip_runtime_api.h:850
hipMemAccessFlags flags
Accessibility flags to set.
Definition hip_runtime_api.h:852
hipMemLocation location
Location on which the accessibility has to change.
Definition hip_runtime_api.h:851
Definition hip_runtime_api.h:1156
const hipMemAccessDesc * accessDescs
Definition hip_runtime_api.h:1159
size_t bytesize
The size of the requested allocation in bytes.
Definition hip_runtime_api.h:1162
hipMemPoolProps poolProps
Definition hip_runtime_api.h:1157
size_t accessDescCount
The number of access descriptors.
Definition hip_runtime_api.h:1161
void * dptr
Returned device address of the allocation.
Definition hip_runtime_api.h:1163
Definition hip_runtime_api.h:1271
void * win32HandleMetaData
Metadata for Win32 handles.
Definition hip_runtime_api.h:1275
hipMemAllocationType type
Memory allocation type.
Definition hip_runtime_api.h:1272
unsigned char gpuDirectRDMACapable
RDMA capable.
Definition hip_runtime_api.h:1278
struct hipMemAllocationProp::@24 allocFlags
unsigned short usage
Usage.
Definition hip_runtime_api.h:1279
hipMemAllocationHandleType requestedHandleType
Requested handle type.
Definition hip_runtime_api.h:1273
hipMemLocation location
Memory location.
Definition hip_runtime_api.h:1274
unsigned char compressionType
Compression type.
Definition hip_runtime_api.h:1277
Definition hip_runtime_api.h:834
int id
Identifier for the provided location type hipMemLocationType.
Definition hip_runtime_api.h:836
hipMemLocationType type
Specifies the location type, which describes the meaning of id.
Definition hip_runtime_api.h:835
Definition hip_runtime_api.h:878
hipMemLocation location
Location where allocations should reside.
Definition hip_runtime_api.h:881
hipMemAllocationHandleType handleTypes
Handle types that will be supported by allocations from the pool.
Definition hip_runtime_api.h:880
hipMemAllocationType allocType
Allocation type. Currently must be specified as hipMemAllocationTypePinned.
Definition hip_runtime_api.h:879
unsigned char reserved[64]
Reserved for future use, must be 0.
Definition hip_runtime_api.h:886
void * win32SecurityAttributes
Definition hip_runtime_api.h:885
Definition hip_runtime_api.h:891
unsigned char reserved[64]
Definition hip_runtime_api.h:892
Definition driver_types.h:374
Definition hip_runtime_api.h:1147
void * dst
Definition hip_runtime_api.h:1148
unsigned int value
Definition hip_runtime_api.h:1152
unsigned int elementSize
Definition hip_runtime_api.h:1149
size_t width
Definition hip_runtime_api.h:1153
size_t pitch
Definition hip_runtime_api.h:1151
size_t height
Definition hip_runtime_api.h:1150
Definition driver_types.h:120
Definition driver_types.h:357
Definition hip_runtime_api.h:189
int device
Definition hip_runtime_api.h:195
enum hipMemoryType memoryType
Definition hip_runtime_api.h:192
void * hostPointer
Definition hip_runtime_api.h:197
enum hipMemoryType type
Definition hip_runtime_api.h:193
int isManaged
Definition hip_runtime_api.h:198
void * devicePointer
Definition hip_runtime_api.h:196
unsigned allocationFlags
Definition hip_runtime_api.h:199
Definition driver_types.h:263
Definition driver_types.h:319
Definition texture_types.h:118
Definition hip_runtime_api.h:85
Definition texture_types.h:97
Defines surface types for HIP runtime.
struct __hip_surface * hipSurfaceObject_t
Definition surface_types.h:41
struct __hip_texture * hipTextureObject_t
Definition texture_types.h:72
hipTextureAddressMode
Definition texture_types.h:77
hipTextureFilterMode
Definition texture_types.h:87
Definition hip_runtime_api.h:1185
int cooperative
Definition hip_runtime_api.h:1187
hipAccessPolicyWindow accessPolicyWindow
Definition hip_runtime_api.h:1186