Global enum and defines

Global enum and defines#

enum hipMemoryType#

hipMemoryType (for pointer attributes)

Note

hipMemoryType enum values are combination of cudaMemoryType and cuMemoryType and AMD specific enum values.

Values:

enumerator hipMemoryTypeUnregistered#: Unregistered memory.

enumerator hipMemoryTypeHost#: Memory is physically located on host.

enumerator hipMemoryTypeDevice#: Memory is physically located on device. (see deviceId for specific device)

enumerator hipMemoryTypeManaged#: Managed memory, automaticallly managed by the unified memory system place holder for new values.

enumerator hipMemoryTypeArray#: Array memory, physically located on device. (see deviceId for specific device)

enumerator hipMemoryTypeUnified#: unified address space

enum hipDeviceAttribute_t#

hipDeviceAttribute_t hipDeviceAttributeUnused number: 5

Values:

enumerator hipDeviceAttributeCudaCompatibleBegin#

enumerator hipDeviceAttributeEccEnabled#: Whether ECC support is enabled.

enumerator hipDeviceAttributeAccessPolicyMaxWindowSize#: Cuda only. The maximum size of the window policy in bytes.

enumerator hipDeviceAttributeAsyncEngineCount#: Asynchronous engines number.

enumerator hipDeviceAttributeCanMapHostMemory#: Whether host memory can be mapped into device address space.

enumerator hipDeviceAttributeCanUseHostPointerForRegisteredMem#: Device can access host registered memory at the same virtual address as the CPU

enumerator hipDeviceAttributeClockRate#: Peak clock frequency in kilohertz.

enumerator hipDeviceAttributeComputeMode#: Compute mode that device is currently in.

enumerator hipDeviceAttributeComputePreemptionSupported#: Device supports Compute Preemption.

enumerator hipDeviceAttributeConcurrentKernels#: Device can possibly execute multiple kernels concurrently.

enumerator hipDeviceAttributeConcurrentManagedAccess#: Device can coherently access managed memory concurrently with the CPU.

enumerator hipDeviceAttributeCooperativeLaunch#: Support cooperative launch.

enumerator hipDeviceAttributeCooperativeMultiDeviceLaunch#: Support cooperative launch on multiple devices.

enumerator hipDeviceAttributeDeviceOverlap#: Device can concurrently copy memory and execute a kernel. Deprecated. Use instead asyncEngineCount.

enumerator hipDeviceAttributeDirectManagedMemAccessFromHost#: Host can directly access managed memory on the device without migration

enumerator hipDeviceAttributeGlobalL1CacheSupported#: Device supports caching globals in L1.

enumerator hipDeviceAttributeHostNativeAtomicSupported#: Link between the device and the host supports native atomic operations.

enumerator hipDeviceAttributeIntegrated#: Device is integrated GPU.

enumerator hipDeviceAttributeIsMultiGpuBoard#: Multiple GPU devices.

enumerator hipDeviceAttributeKernelExecTimeout#: Run time limit for kernels executed on the device.

enumerator hipDeviceAttributeL2CacheSize#: Size of L2 cache in bytes. 0 if the device doesn’t have L2 cache.

enumerator hipDeviceAttributeLocalL1CacheSupported#: caching locals in L1 is supported

enumerator hipDeviceAttributeLuid#: 8-byte locally unique identifier in 8 bytes. Undefined on TCC and non-Windows platforms

enumerator hipDeviceAttributeLuidDeviceNodeMask#: Luid device node mask. Undefined on TCC and non-Windows platforms.

enumerator hipDeviceAttributeComputeCapabilityMajor#: Major compute capability version number.

enumerator hipDeviceAttributeManagedMemory#: Device supports allocating managed memory on this system.

enumerator hipDeviceAttributeMaxBlocksPerMultiProcessor#: Max block size per multiprocessor.

enumerator hipDeviceAttributeMaxBlockDimX#: Max block size in width.

enumerator hipDeviceAttributeMaxBlockDimY#: Max block size in height.

enumerator hipDeviceAttributeMaxBlockDimZ#: Max block size in depth.

enumerator hipDeviceAttributeMaxGridDimX#: Max grid size in width.

enumerator hipDeviceAttributeMaxGridDimY#: Max grid size in height.

enumerator hipDeviceAttributeMaxGridDimZ#: Max grid size in depth.

enumerator hipDeviceAttributeMaxSurface1D#: Maximum size of 1D surface.

enumerator hipDeviceAttributeMaxSurface1DLayered#: Cuda only. Maximum dimensions of 1D layered surface.

enumerator hipDeviceAttributeMaxSurface2D#: Maximum dimension (width, height) of 2D surface.

enumerator hipDeviceAttributeMaxSurface2DLayered#: Cuda only. Maximum dimensions of 2D layered surface.

enumerator hipDeviceAttributeMaxSurface3D#: Maximum dimension (width, height, depth) of 3D surface.

enumerator hipDeviceAttributeMaxSurfaceCubemap#: Cuda only. Maximum dimensions of Cubemap surface.

enumerator hipDeviceAttributeMaxSurfaceCubemapLayered#: Cuda only. Maximum dimension of Cubemap layered surface.

enumerator hipDeviceAttributeMaxTexture1DWidth#: Maximum size of 1D texture.

enumerator hipDeviceAttributeMaxTexture1DLayered#: Maximum dimensions of 1D layered texture.

enumerator hipDeviceAttributeMaxTexture1DLinear#: Maximum number of elements allocatable in a 1D linear texture. Use cudaDeviceGetTexture1DLinearMaxWidth() instead on Cuda.

enumerator hipDeviceAttributeMaxTexture1DMipmap#: Maximum size of 1D mipmapped texture.

enumerator hipDeviceAttributeMaxTexture2DWidth#: Maximum dimension width of 2D texture.

enumerator hipDeviceAttributeMaxTexture2DHeight#: Maximum dimension hight of 2D texture.

enumerator hipDeviceAttributeMaxTexture2DGather#: Maximum dimensions of 2D texture if gather operations performed.

enumerator hipDeviceAttributeMaxTexture2DLayered#: Maximum dimensions of 2D layered texture.

enumerator hipDeviceAttributeMaxTexture2DLinear#: Maximum dimensions (width, height, pitch) of 2D textures bound to pitched memory.

enumerator hipDeviceAttributeMaxTexture2DMipmap#: Maximum dimensions of 2D mipmapped texture.

enumerator hipDeviceAttributeMaxTexture3DWidth#: Maximum dimension width of 3D texture.

enumerator hipDeviceAttributeMaxTexture3DHeight#: Maximum dimension height of 3D texture.

enumerator hipDeviceAttributeMaxTexture3DDepth#: Maximum dimension depth of 3D texture.

enumerator hipDeviceAttributeMaxTexture3DAlt#: Maximum dimensions of alternate 3D texture.

enumerator hipDeviceAttributeMaxTextureCubemap#: Maximum dimensions of Cubemap texture.

enumerator hipDeviceAttributeMaxTextureCubemapLayered#: Maximum dimensions of Cubemap layered texture.

enumerator hipDeviceAttributeMaxThreadsDim#: Maximum dimension of a block.

enumerator hipDeviceAttributeMaxThreadsPerBlock#: Maximum number of threads per block.

enumerator hipDeviceAttributeMaxThreadsPerMultiProcessor#: Maximum resident threads per multiprocessor.

enumerator hipDeviceAttributeMaxPitch#: Maximum pitch in bytes allowed by memory copies.

enumerator hipDeviceAttributeMemoryBusWidth#: Global memory bus width in bits.

enumerator hipDeviceAttributeMemoryClockRate#: Peak memory clock frequency in kilohertz.

enumerator hipDeviceAttributeComputeCapabilityMinor#: Minor compute capability version number.

enumerator hipDeviceAttributeMultiGpuBoardGroupID#: Unique ID of device group on the same multi-GPU board.

enumerator hipDeviceAttributeMultiprocessorCount#: Number of multiprocessors on the device.

enumerator hipDeviceAttributeUnused1#: Previously hipDeviceAttributeName.

enumerator hipDeviceAttributePageableMemoryAccess#: Device supports coherently accessing pageable memory without calling hipHostRegister on it

enumerator hipDeviceAttributePageableMemoryAccessUsesHostPageTables#: Device accesses pageable memory via the host’s page tables.

enumerator hipDeviceAttributePciBusId#: PCI Bus ID.

enumerator hipDeviceAttributePciDeviceId#: PCI Device ID.

enumerator hipDeviceAttributePciDomainID#: PCI Domain ID.

enumerator hipDeviceAttributePersistingL2CacheMaxSize#: Maximum l2 persisting lines capacity in bytes.

enumerator hipDeviceAttributeMaxRegistersPerBlock#: 32-bit registers available to a thread block. This number is shared by all thread blocks simultaneously resident on a multiprocessor.

enumerator hipDeviceAttributeMaxRegistersPerMultiprocessor#: 32-bit registers available per block.

enumerator hipDeviceAttributeReservedSharedMemPerBlock#: Shared memory reserved by CUDA driver per block.

enumerator hipDeviceAttributeMaxSharedMemoryPerBlock#: Maximum shared memory available per block in bytes.

enumerator hipDeviceAttributeSharedMemPerBlockOptin#: Maximum shared memory per block usable by special opt in.

enumerator hipDeviceAttributeSharedMemPerMultiprocessor#: Shared memory available per multiprocessor.

enumerator hipDeviceAttributeSingleToDoublePrecisionPerfRatio#: Cuda only. Performance ratio of single precision to double precision.

enumerator hipDeviceAttributeStreamPrioritiesSupported#: Whether to support stream priorities.

enumerator hipDeviceAttributeSurfaceAlignment#: Alignment requirement for surfaces.

enumerator hipDeviceAttributeTccDriver#: Cuda only. Whether device is a Tesla device using TCC driver.

enumerator hipDeviceAttributeTextureAlignment#: Alignment requirement for textures.

enumerator hipDeviceAttributeTexturePitchAlignment#: Pitch alignment requirement for 2D texture references bound to pitched memory;.

enumerator hipDeviceAttributeTotalConstantMemory#: Constant memory size in bytes.

enumerator hipDeviceAttributeTotalGlobalMem#: Global memory available on devicice.

enumerator hipDeviceAttributeUnifiedAddressing#: Cuda only. An unified address space shared with the host.

enumerator hipDeviceAttributeUnused2#: Previously hipDeviceAttributeUuid.

enumerator hipDeviceAttributeWarpSize#: Warp size in threads.

enumerator hipDeviceAttributeMemoryPoolsSupported#: Device supports HIP Stream Ordered Memory Allocator.

enumerator hipDeviceAttributeVirtualMemoryManagementSupported#: Device supports HIP virtual memory management.

enumerator hipDeviceAttributeHostRegisterSupported#: Can device support host memory registration via hipHostRegister.

enumerator hipDeviceAttributeMemoryPoolSupportedHandleTypes#: Supported handle mask for HIP Stream Ordered Memory Allocator.

enumerator hipDeviceAttributeCudaCompatibleEnd#

enumerator hipDeviceAttributeAmdSpecificBegin#

enumerator hipDeviceAttributeClockInstructionRate#: Frequency in khz of the timer used by the device-side “clock*”.

enumerator hipDeviceAttributeUnused3#: Previously hipDeviceAttributeArch.

enumerator hipDeviceAttributeMaxSharedMemoryPerMultiprocessor#: Maximum Shared Memory PerMultiprocessor.

enumerator hipDeviceAttributeUnused4#: Previously hipDeviceAttributeGcnArch.

enumerator hipDeviceAttributeUnused5#: Previously hipDeviceAttributeGcnArchName.

enumerator hipDeviceAttributeHdpMemFlushCntl#: Address of the HDP_MEM_COHERENCY_FLUSH_CNTL register.

enumerator hipDeviceAttributeHdpRegFlushCntl#: Address of the HDP_REG_COHERENCY_FLUSH_CNTL register.

enumerator hipDeviceAttributeCooperativeMultiDeviceUnmatchedFunc#: Supports cooperative launch on multiple devices with unmatched functions

enumerator hipDeviceAttributeCooperativeMultiDeviceUnmatchedGridDim#: Supports cooperative launch on multiple devices with unmatched grid dimensions

enumerator hipDeviceAttributeCooperativeMultiDeviceUnmatchedBlockDim#: Supports cooperative launch on multiple devices with unmatched block dimensions

enumerator hipDeviceAttributeCooperativeMultiDeviceUnmatchedSharedMem#: Supports cooperative launch on multiple devices with unmatched shared memories

enumerator hipDeviceAttributeIsLargeBar#: Whether it is LargeBar.

enumerator hipDeviceAttributeAsicRevision#: Revision of the GPU in this device.

enumerator hipDeviceAttributeCanUseStreamWaitValue#: ‘1’ if Device supports hipStreamWaitValue32() and hipStreamWaitValue64(), ‘0’ otherwise.

enumerator hipDeviceAttributeImageSupport#: ‘1’ if Device supports image, ‘0’ otherwise.

enumerator hipDeviceAttributePhysicalMultiProcessorCount#: All available physical compute units for the device

enumerator hipDeviceAttributeFineGrainSupport#: ‘1’ if Device supports fine grain, ‘0’ otherwise

enumerator hipDeviceAttributeWallClockRate#: Constant frequency of wall clock in kilohertz.

enumerator hipDeviceAttributeAmdSpecificEnd#

enumerator hipDeviceAttributeVendorSpecificBegin#

enum hipDriverProcAddressQueryResult#

Values:

enumerator HIP_GET_PROC_ADDRESS_SUCCESS#

enumerator HIP_GET_PROC_ADDRESS_SYMBOL_NOT_FOUND#

enumerator HIP_GET_PROC_ADDRESS_VERSION_NOT_SUFFICIENT#

enum hipComputeMode#

Values:

enumerator hipComputeModeDefault#

enumerator hipComputeModeExclusive#

enumerator hipComputeModeProhibited#

enumerator hipComputeModeExclusiveProcess#

enum hipFlushGPUDirectRDMAWritesOptions#

Values:

enumerator hipFlushGPUDirectRDMAWritesOptionHost#

enumerator hipFlushGPUDirectRDMAWritesOptionMemOps#

enum hipGPUDirectRDMAWritesOrdering#

Values:

enumerator hipGPUDirectRDMAWritesOrderingNone#

enumerator hipGPUDirectRDMAWritesOrderingOwner#

enumerator hipGPUDirectRDMAWritesOrderingAllDevices#

enum hipDeviceP2PAttr#

Values:

enumerator hipDevP2PAttrPerformanceRank#

enumerator hipDevP2PAttrAccessSupported#

enumerator hipDevP2PAttrNativeAtomicSupported#

enumerator hipDevP2PAttrHipArrayAccessSupported#

enum hipLimit_t#

hipLimit

Note

In HIP device limit-related APIs, any input limit value other than those defined in the enum is treated as “UnsupportedLimit” by default.

Values:

enumerator hipLimitStackSize#: Limit of stack size in bytes on the current device, per thread. The size is in units of 256 dwords, up to the limit of (128K - 16)

enumerator hipLimitPrintfFifoSize#: Size limit in bytes of fifo used by printf call on the device. Currently not supported

enumerator hipLimitMallocHeapSize#: Limit of heap size in bytes on the current device, should be less than the global memory size on the device

enumerator hipLimitRange#: Supported limit range.

enum hipMemoryAdvise#

HIP Memory Advise values

Note

This memory advise enumeration is used on Linux, not Windows.

Values:

enumerator hipMemAdviseSetReadMostly#: Data will mostly be read and only occassionally be written to

enumerator hipMemAdviseUnsetReadMostly#: Undo the effect of hipMemAdviseSetReadMostly.

enumerator hipMemAdviseSetPreferredLocation#: Set the preferred location for the data as the specified device

enumerator hipMemAdviseUnsetPreferredLocation#: Clear the preferred location for the data.

enumerator hipMemAdviseSetAccessedBy#: Data will be accessed by the specified device so prevent page faults as much as possible

enumerator hipMemAdviseUnsetAccessedBy#: Let HIP to decide on the page faulting policy for the specified device

enumerator hipMemAdviseSetCoarseGrain#: The default memory model is fine-grain. That allows coherent operations between host and device, while executing kernels. The coarse-grain can be used for data that only needs to be coherent at dispatch boundaries for better performance

enumerator hipMemAdviseUnsetCoarseGrain#: Restores cache coherency policy back to fine-grain.

enum hipMemRangeCoherencyMode#

HIP Coherency Mode

Values:

enumerator hipMemRangeCoherencyModeFineGrain#: Updates to memory with this attribute can be done coherently from all devices

enumerator hipMemRangeCoherencyModeCoarseGrain#: Writes to memory with this attribute can be performed by a single device at a time

enumerator hipMemRangeCoherencyModeIndeterminate#: Memory region queried contains subregions with both hipMemRangeCoherencyModeFineGrain and hipMemRangeCoherencyModeCoarseGrain attributes

enum hipMemRangeAttribute#

HIP range attributes

Values:

enumerator hipMemRangeAttributeReadMostly#: Whether the range will mostly be read and only occassionally be written to

enumerator hipMemRangeAttributePreferredLocation#: The preferred location of the range.

enumerator hipMemRangeAttributeAccessedBy#: Memory range has hipMemAdviseSetAccessedBy set for the specified device

enumerator hipMemRangeAttributeLastPrefetchLocation#

prefetched

The last location to where the range was

enumerator hipMemRangeAttributeCoherencyMode#: Returns coherency mode hipMemRangeCoherencyMode for the range

enum hipMemPoolAttr#

HIP memory pool attributes

Values:

enumerator hipMemPoolReuseFollowEventDependencies#: (value type = int) Allow hipMemAllocAsync to use memory asynchronously freed in another streams as long as a stream ordering dependency of the allocating stream on the free action exists. hip events and null stream interactions can create the required stream ordered dependencies. (default enabled)

enumerator hipMemPoolReuseAllowOpportunistic#: (value type = int) Allow reuse of already completed frees when there is no dependency between the free and allocation. (default enabled)

enumerator hipMemPoolReuseAllowInternalDependencies#: (value type = int) Allow hipMemAllocAsync to insert new stream dependencies in order to establish the stream ordering required to reuse a piece of memory released by cuFreeAsync (default enabled).

enumerator hipMemPoolAttrReleaseThreshold#: (value type = uint64_t) Amount of reserved memory in bytes to hold onto before trying to release memory back to the OS. When more than the release threshold bytes of memory are held by the memory pool, the allocator will try to release memory back to the OS on the next call to stream, event or context synchronize. (default 0)

enumerator hipMemPoolAttrReservedMemCurrent#: (value type = uint64_t) Amount of backing memory currently allocated for the mempool.

enumerator hipMemPoolAttrReservedMemHigh#: (value type = uint64_t) High watermark of backing memory allocated for the mempool since the last time it was reset. High watermark can only be reset to zero.

enumerator hipMemPoolAttrUsedMemCurrent#: (value type = uint64_t) Amount of memory from the pool that is currently in use by the application.

enumerator hipMemPoolAttrUsedMemHigh#: (value type = uint64_t) High watermark of the amount of memory from the pool that was in use by the application since the last time it was reset. High watermark can only be reset to zero.

enum hipMemLocationType#

Specifies the type of location

Values:

enumerator hipMemLocationTypeInvalid#

enumerator hipMemLocationTypeDevice#: Device location, thus it’s HIP device ID.

enum hipMemAccessFlags#

Specifies the memory protection flags for mapping

Values:

enumerator hipMemAccessFlagsProtNone#: Default, make the address range not accessible.

enumerator hipMemAccessFlagsProtRead#: Set the address range read accessible.

enumerator hipMemAccessFlagsProtReadWrite#: Set the address range read-write accessible.

enum hipMemAllocationType#

Defines the allocation types

Values:

enumerator hipMemAllocationTypeInvalid#

enumerator hipMemAllocationTypePinned#: This allocation type is ‘pinned’, i.e. cannot migrate from its current location while the application is actively using it

enumerator hipMemAllocationTypeMax#

enum hipMemAllocationHandleType#

Flags for specifying handle types for memory pool allocations

Values:

enumerator hipMemHandleTypeNone#: Does not allow any export mechanism.

enumerator hipMemHandleTypePosixFileDescriptor#: Allows a file descriptor for exporting. Permitted only on POSIX systems.

enumerator hipMemHandleTypeWin32#: Allows a Win32 NT handle for exporting. (HANDLE)

enumerator hipMemHandleTypeWin32Kmt#: Allows a Win32 KMT handle for exporting. (D3DKMT_HANDLE)

enum hipJitOption#

hipJitOption

Values:

enumerator hipJitOptionMaxRegisters#

enumerator hipJitOptionThreadsPerBlock#

enumerator hipJitOptionWallTime#

enumerator hipJitOptionInfoLogBuffer#

enumerator hipJitOptionInfoLogBufferSizeBytes#

enumerator hipJitOptionErrorLogBuffer#

enumerator hipJitOptionErrorLogBufferSizeBytes#

enumerator hipJitOptionOptimizationLevel#

enumerator hipJitOptionTargetFromContext#

enumerator hipJitOptionTarget#

enumerator hipJitOptionFallbackStrategy#

enumerator hipJitOptionGenerateDebugInfo#

enumerator hipJitOptionLogVerbose#

enumerator hipJitOptionGenerateLineInfo#

enumerator hipJitOptionCacheMode#

enumerator hipJitOptionSm3xOpt#

enumerator hipJitOptionFastCompile#

enumerator hipJitOptionNumOptions#

enum hipFuncAttribute#

Warning

On AMD devices and some Nvidia devices, these hints and controls are ignored.

Values:

enumerator hipFuncAttributeMaxDynamicSharedMemorySize#

enumerator hipFuncAttributePreferredSharedMemoryCarveout#

enumerator hipFuncAttributeMax#

enum hipFuncCache_t#

Warning

On AMD devices and some Nvidia devices, these hints and controls are ignored.

Values:

enumerator hipFuncCachePreferNone#: no preference for shared memory or L1 (default)

enumerator hipFuncCachePreferShared#: prefer larger shared memory and smaller L1 cache

enumerator hipFuncCachePreferL1#: prefer larger L1 cache and smaller shared memory

enumerator hipFuncCachePreferEqual#: prefer equal size L1 cache and shared memory

enum hipSharedMemConfig#

Warning

On AMD devices and some Nvidia devices, these hints and controls are ignored.

Values:

enumerator hipSharedMemBankSizeDefault#: The compiler selects a device-specific value for the banking.

enumerator hipSharedMemBankSizeFourByte#: Shared mem is banked at 4-bytes intervals and performs best when adjacent threads access data 4 bytes apart.

enumerator hipSharedMemBankSizeEightByte#: Shared mem is banked at 8-byte intervals and performs best when adjacent threads access data 4 bytes apart.

enum hipExternalMemoryHandleType#

Values:

enumerator hipExternalMemoryHandleTypeOpaqueFd#

enumerator hipExternalMemoryHandleTypeOpaqueWin32#

enumerator hipExternalMemoryHandleTypeOpaqueWin32Kmt#

enumerator hipExternalMemoryHandleTypeD3D12Heap#

enumerator hipExternalMemoryHandleTypeD3D12Resource#

enumerator hipExternalMemoryHandleTypeD3D11Resource#

enumerator hipExternalMemoryHandleTypeD3D11ResourceKmt#

enumerator hipExternalMemoryHandleTypeNvSciBuf#

enum hipExternalSemaphoreHandleType#

Values:

enumerator hipExternalSemaphoreHandleTypeOpaqueFd#

enumerator hipExternalSemaphoreHandleTypeOpaqueWin32#

enumerator hipExternalSemaphoreHandleTypeOpaqueWin32Kmt#

enumerator hipExternalSemaphoreHandleTypeD3D12Fence#

enumerator hipExternalSemaphoreHandleTypeD3D11Fence#

enumerator hipExternalSemaphoreHandleTypeNvSciSync#

enumerator hipExternalSemaphoreHandleTypeKeyedMutex#

enumerator hipExternalSemaphoreHandleTypeKeyedMutexKmt#

enumerator hipExternalSemaphoreHandleTypeTimelineSemaphoreFd#

enumerator hipExternalSemaphoreHandleTypeTimelineSemaphoreWin32#

enum hipGraphicsRegisterFlags#

HIP Access falgs for Interop resources.

Values:

enumerator hipGraphicsRegisterFlagsNone#

enumerator hipGraphicsRegisterFlagsReadOnly#: HIP will not write to this registered resource.

enumerator hipGraphicsRegisterFlagsWriteDiscard#: HIP will only write and will not read from this registered resource.

enumerator hipGraphicsRegisterFlagsSurfaceLoadStore#: HIP will bind this resource to a surface.

enumerator hipGraphicsRegisterFlagsTextureGather#: HIP will perform texture gather operations on this registered resource.

enum hipGraphNodeType#

hipGraphNodeType

Values:

enumerator hipGraphNodeTypeKernel#: GPU kernel node.

enumerator hipGraphNodeTypeMemcpy#: Memcpy node.

enumerator hipGraphNodeTypeMemset#: Memset node.

enumerator hipGraphNodeTypeHost#: Host (executable) node.

enumerator hipGraphNodeTypeGraph#: Node which executes an embedded graph.

enumerator hipGraphNodeTypeEmpty#: Empty (no-op) node.

enumerator hipGraphNodeTypeWaitEvent#: External event wait node.

enumerator hipGraphNodeTypeEventRecord#: External event record node.

enumerator hipGraphNodeTypeExtSemaphoreSignal#: External Semaphore signal node.

enumerator hipGraphNodeTypeExtSemaphoreWait#: External Semaphore wait node.

enumerator hipGraphNodeTypeMemAlloc#: Memory alloc node.

enumerator hipGraphNodeTypeMemFree#: Memory free node.

enumerator hipGraphNodeTypeMemcpyFromSymbol#: MemcpyFromSymbol node.

enumerator hipGraphNodeTypeMemcpyToSymbol#: MemcpyToSymbol node.

enumerator hipGraphNodeTypeCount#

enum hipAccessProperty#

Values:

enumerator hipAccessPropertyNormal#

enumerator hipAccessPropertyStreaming#

enumerator hipAccessPropertyPersisting#

enum hipLaunchAttributeID#

Launch Attribute ID

Values:

enumerator hipLaunchAttributeAccessPolicyWindow#: Valid for Streams, graph nodes, launches

enumerator hipLaunchAttributeCooperative#: Valid for graph nodes, launches

enumerator hipLaunchAttributePriority#: Valid for graph node, streams, launches

enum hipGraphExecUpdateResult#

Graph execution update result

Values:

enumerator hipGraphExecUpdateSuccess#: The update succeeded.

enumerator hipGraphExecUpdateError#: The update failed for an unexpected reason which is described in the return value of the function

enumerator hipGraphExecUpdateErrorTopologyChanged#: The update failed because the topology changed.

enumerator hipGraphExecUpdateErrorNodeTypeChanged#: The update failed because a node type changed.

enumerator hipGraphExecUpdateErrorFunctionChanged#: The update failed because the function of a kernel node changed.

enumerator hipGraphExecUpdateErrorParametersChanged#: The update failed because the parameters changed in a way that is not supported.

enumerator hipGraphExecUpdateErrorNotSupported#: The update failed because something about the node is not supported.

enumerator hipGraphExecUpdateErrorUnsupportedFunctionChange#

enum hipStreamCaptureMode#

Values:

enumerator hipStreamCaptureModeGlobal#

enumerator hipStreamCaptureModeThreadLocal#

enumerator hipStreamCaptureModeRelaxed#

enum hipStreamCaptureStatus#

Values:

enumerator hipStreamCaptureStatusNone#: Stream is not capturing.

enumerator hipStreamCaptureStatusActive#: Stream is actively capturing.

enumerator hipStreamCaptureStatusInvalidated#: Stream is part of a capture sequence that has been invalidated, but not terminated

enum hipStreamUpdateCaptureDependenciesFlags#

Values:

enumerator hipStreamAddCaptureDependencies#: Add new nodes to the dependency set.

enumerator hipStreamSetCaptureDependencies#: Replace the dependency set with the new nodes.

enum hipGraphMemAttributeType#

Values:

enumerator hipGraphMemAttrUsedMemCurrent#: Amount of memory, in bytes, currently associated with graphs.

enumerator hipGraphMemAttrUsedMemHigh#: High watermark of memory, in bytes, associated with graphs since the last time.

enumerator hipGraphMemAttrReservedMemCurrent#: Amount of memory, in bytes, currently allocated for graphs.

enumerator hipGraphMemAttrReservedMemHigh#: High watermark of memory, in bytes, currently allocated for graphs.

enum hipUserObjectFlags#

Values:

enumerator hipUserObjectNoDestructorSync#: Destructor execution is not synchronized.

enum hipUserObjectRetainFlags#

Values:

enumerator hipGraphUserObjectMove#: Add new reference or retain.

enum hipGraphInstantiateFlags#

Values:

enumerator hipGraphInstantiateFlagAutoFreeOnLaunch#: Automatically free memory allocated in a graph before relaunching.

enumerator hipGraphInstantiateFlagUpload#: Automatically upload the graph after instantiation.

enumerator hipGraphInstantiateFlagDeviceLaunch#: Instantiate the graph to be launched from the device.

enumerator hipGraphInstantiateFlagUseNodePriority#: Run the graph using the per-node priority attributes rather than the priority of the stream it is launched into.

enum hipGraphDebugDotFlags#

Values:

enumerator hipGraphDebugDotFlagsVerbose#: Output all debug data as if every debug flag is enabled

enumerator hipGraphDebugDotFlagsKernelNodeParams#: Adds hipKernelNodeParams to output

enumerator hipGraphDebugDotFlagsMemcpyNodeParams#: Adds hipMemcpy3DParms to output

enumerator hipGraphDebugDotFlagsMemsetNodeParams#: Adds hipMemsetParams to output

enumerator hipGraphDebugDotFlagsHostNodeParams#: Adds hipHostNodeParams to output

enumerator hipGraphDebugDotFlagsEventNodeParams#: Adds hipEvent_t handle from record and wait nodes to output

enumerator hipGraphDebugDotFlagsExtSemasSignalNodeParams#: Adds hipExternalSemaphoreSignalNodeParams values to output

enumerator hipGraphDebugDotFlagsExtSemasWaitNodeParams#: Adds hipExternalSemaphoreWaitNodeParams to output

enumerator hipGraphDebugDotFlagsKernelNodeAttributes#: Adds hipKernelNodeAttrID values to output

enumerator hipGraphDebugDotFlagsHandles#: Adds node handles and every kernel function handle to output

enum hipGraphInstantiateResult#

hipGraphInstantiateWithParams results

Values:

enumerator hipGraphInstantiateSuccess#: Instantiation Success

enumerator hipGraphInstantiateError#: Instantiation failed for an unexpected reason which is described in the return value of the function

enumerator hipGraphInstantiateInvalidStructure#: Instantiation failed due to invalid structure, such as cycles

enumerator hipGraphInstantiateNodeOperationNotSupported#: Instantiation for device launch failed because the graph contained an unsupported operation

enumerator hipGraphInstantiateMultipleDevicesNotSupported#: Instantiation for device launch failed due to the nodes belonging to different contexts

enum hipMemAllocationGranularity_flags#

Flags for granularity

Values:

enumerator hipMemAllocationGranularityMinimum#: Minimum granularity.

enumerator hipMemAllocationGranularityRecommended#: Recommended granularity for performance.

enum hipMemHandleType#

Memory handle type

Values:

enumerator hipMemHandleTypeGeneric#: Generic handle type.

enum hipMemOperationType#

Memory operation types

Values:

enumerator hipMemOperationTypeMap#: Map operation.

enumerator hipMemOperationTypeUnmap#: Unmap operation.

enum hipArraySparseSubresourceType#

Subresource types for sparse arrays

Values:

enumerator hipArraySparseSubresourceTypeSparseLevel#: Sparse level.

enumerator hipArraySparseSubresourceTypeMiptail#: Miptail.

enum hipGraphDependencyType#

Values:

enumerator hipGraphDependencyTypeDefault#

enumerator hipGraphDependencyTypeProgrammatic#

enum hiprtcResult#

hiprtc error code

Values:

enumerator HIPRTC_SUCCESS#: Success.

enumerator HIPRTC_ERROR_OUT_OF_MEMORY#: Out of memory.

enumerator HIPRTC_ERROR_PROGRAM_CREATION_FAILURE#: Failed to create program.

enumerator HIPRTC_ERROR_INVALID_INPUT#: Invalid input.

enumerator HIPRTC_ERROR_INVALID_PROGRAM#: Invalid program.

enumerator HIPRTC_ERROR_INVALID_OPTION#: Invalid option.

enumerator HIPRTC_ERROR_COMPILATION#: Compilation error.

enumerator HIPRTC_ERROR_BUILTIN_OPERATION_FAILURE#: Failed in builtin operation.

enumerator HIPRTC_ERROR_NO_NAME_EXPRESSIONS_AFTER_COMPILATION#: No name expression after compilation.

enumerator HIPRTC_ERROR_NO_LOWERED_NAMES_BEFORE_COMPILATION#: No lowered names before compilation.

enumerator HIPRTC_ERROR_NAME_EXPRESSION_NOT_VALID#: Invalid name expression.

enumerator HIPRTC_ERROR_INTERNAL_ERROR#: Internal error.

enumerator HIPRTC_ERROR_LINKING#: Error in linking.

enum hiprtcJIT_option#

hiprtc JIT option

Values:

enumerator HIPRTC_JIT_MAX_REGISTERS#: CUDA Only Maximum registers may be used in a thread, passed to compiler.

enumerator HIPRTC_JIT_THREADS_PER_BLOCK#: CUDA Only Number of thread per block.

enumerator HIPRTC_JIT_WALL_TIME#: CUDA Only Value for total wall clock time.

enumerator HIPRTC_JIT_INFO_LOG_BUFFER#: CUDA Only Pointer to the buffer with logged information.

enumerator HIPRTC_JIT_INFO_LOG_BUFFER_SIZE_BYTES#: CUDA Only Size of the buffer in bytes for logged info.

enumerator HIPRTC_JIT_ERROR_LOG_BUFFER#: CUDA Only Pointer to the buffer with logged error(s)

enumerator HIPRTC_JIT_ERROR_LOG_BUFFER_SIZE_BYTES#: CUDA Only Size of the buffer in bytes for logged error(s)

enumerator HIPRTC_JIT_OPTIMIZATION_LEVEL#: Value of optimization level for generated codes, acceptable options -O0, -O1, -O2, -O3.

enumerator HIPRTC_JIT_TARGET_FROM_HIPCONTEXT#: CUDA Only The target context, which is the default.

enumerator HIPRTC_JIT_TARGET#: CUDA Only JIT target.

enumerator HIPRTC_JIT_FALLBACK_STRATEGY#: CUDA Only Fallback strategy.

enumerator HIPRTC_JIT_GENERATE_DEBUG_INFO#: CUDA Only Generate debug information.

enumerator HIPRTC_JIT_LOG_VERBOSE#: CUDA Only Generate log verbose.

enumerator HIPRTC_JIT_GENERATE_LINE_INFO#: CUDA Only Generate line number information.

enumerator HIPRTC_JIT_CACHE_MODE#: CUDA Only Set cache mode.

enumerator HIPRTC_JIT_NEW_SM3X_OPT#

Deprecated:: CUDA Only New SM3X option.

enumerator HIPRTC_JIT_FAST_COMPILE#: CUDA Only Set fast compile.

enumerator HIPRTC_JIT_GLOBAL_SYMBOL_NAMES#: CUDA Only Array of device symbol names to be relocated to the host.

enumerator HIPRTC_JIT_GLOBAL_SYMBOL_ADDRESS#: CUDA Only Array of host addresses to be relocated to the device.

enumerator HIPRTC_JIT_GLOBAL_SYMBOL_COUNT#: CUDA Only Number of symbol count.

enumerator HIPRTC_JIT_LTO#

Deprecated:: CUDA Only Enable link-time optimization for device code

enumerator HIPRTC_JIT_FTZ#

Deprecated:: CUDA Only Set single-precision denormals.

enumerator HIPRTC_JIT_PREC_DIV#

Deprecated:: CUDA Only Set single-precision floating-point division and reciprocals

enumerator HIPRTC_JIT_PREC_SQRT#

Deprecated:: CUDA Only Set single-precision floating-point square root

enumerator HIPRTC_JIT_FMA#

Deprecated:: CUDA Only Enable floating-point multiplies and adds/subtracts operations

enumerator HIPRTC_JIT_NUM_OPTIONS#: Number of options.

enumerator HIPRTC_JIT_IR_TO_ISA_OPT_EXT#: Linker options to be passed on to compiler

Note

Only supported for the AMD platform.

enumerator HIPRTC_JIT_IR_TO_ISA_OPT_COUNT_EXT#: Count of linker options to be passed on to compiler

Note

Only supported for the AMD platform

enum hiprtcJITInputType#

hiprtc JIT input type

Values:

enumerator HIPRTC_JIT_INPUT_CUBIN#: Input cubin.

enumerator HIPRTC_JIT_INPUT_PTX#: Input PTX.

enumerator HIPRTC_JIT_INPUT_FATBINARY#: Input fat binary.

enumerator HIPRTC_JIT_INPUT_OBJECT#: Input object.

enumerator HIPRTC_JIT_INPUT_LIBRARY#: Input library.

enumerator HIPRTC_JIT_INPUT_NVVM#: Input NVVM.

enumerator HIPRTC_JIT_NUM_LEGACY_INPUT_TYPES#: Number of legacy input type.

enumerator HIPRTC_JIT_INPUT_LLVM_BITCODE#: LLVM bitcode or IR assembly.

enumerator HIPRTC_JIT_INPUT_LLVM_BUNDLED_BITCODE#: LLVM bundled bitcode.

enumerator HIPRTC_JIT_INPUT_LLVM_ARCHIVES_OF_BUNDLED_BITCODE#: LLVM archives of boundled bitcode.

enumerator HIPRTC_JIT_NUM_INPUT_TYPES#

enum hipGLDeviceList#

HIP Devices used by current OpenGL Context.

Values:

enumerator hipGLDeviceListAll#: All hip devices used by current OpenGL context.

enumerator hipGLDeviceListCurrentFrame#

frame

Hip devices used by current OpenGL context in current

enumerator hipGLDeviceListNextFrame#

frame.

Hip devices used by current OpenGL context in next

enum __HIP_NODISCARD hipError_t hipError_t: HIP error type

typedef struct ihipCtx_t *hipCtx_t#

typedef int hipDevice_t#

typedef struct ihipStream_t *hipStream_t#

typedef struct ihipModule_t *hipModule_t#

typedef struct ihipModuleSymbol_t *hipFunction_t#

typedef struct ihipMemPoolHandle_t *hipMemPool_t#: HIP memory pool

typedef struct ihipEvent_t *hipEvent_t#

typedef void *hipExternalMemory_t#

typedef void *hipExternalSemaphore_t#

typedef struct _hipGraphicsResource hipGraphicsResource#

typedef hipGraphicsResource *hipGraphicsResource_t#

typedef struct ihipGraph *hipGraph_t#: An opaque value that represents a hip graph

typedef struct hipGraphNode *hipGraphNode_t#: An opaque value that represents a hip graph node

typedef struct hipGraphExec *hipGraphExec_t#: An opaque value that represents a hip graph Exec

typedef struct hipUserObject *hipUserObject_t#: An opaque value that represents a user obj

typedef void (*hipHostFn_t)(void *userData)#

typedef struct ihipMemGenericAllocationHandle *hipMemGenericAllocationHandle_t#: Generic handle for memory allocation

typedef unsigned int GLuint#: GLuint as uint.

typedef unsigned int GLenum#: GLenum as uint.

hipGetDeviceProperties#

hipDeviceProp_t#

hipChooseDevice#

__HIP_NODISCARD#

GENERIC_GRID_LAUNCH#

DEPRECATED(msg)#

DEPRECATED_MSG#

HIP_LAUNCH_PARAM_BUFFER_POINTER#

HIP_LAUNCH_PARAM_BUFFER_SIZE#

HIP_LAUNCH_PARAM_END#

__dparm(x)#

hipIpcMemLazyEnablePeerAccess#

HIP_IPC_HANDLE_SIZE#

hipStreamDefault#: Flags that can be used with hipStreamCreateWithFlags. Default stream creation flags. These are used with hipStreamCreate().

hipStreamNonBlocking#: Stream does not implicitly synchronize with null stream.

hipEventDefault#: Default flags.

hipEventBlockingSync#: Waiting will yield CPU. Power-friendly and usage-friendly but may increase latency.

hipEventDisableTiming#: Disable event’s capability to record timing information. May improve performance.

hipEventInterprocess#: Event can support IPC. hipEventDisableTiming also must be set.

hipEventDisableSystemFence#: Disable performing a system scope sequentially consistent memory fence when the event transitions from recording to recorded. This can be used for events that are only being used to measure timing, and do not require the event inspection operations (see hipEventSynchronize, hipEventQuery, and hipEventElapsedTime) to synchronize-with the work on which the recorded event (see hipEventRecord) is waiting. On some AMD GPU devices this can improve the accuracy of timing measurements by avoiding the cost of cache writeback and invalidation, and the performance impact of those actions on the execution of following work.

hipEventReleaseToDevice#: Use a device-scope release when recording this event. This flag is useful to obtain more precise timings of commands between events. The flag is a no-op on CUDA platforms.

hipEventReleaseToSystem#: Use a system-scope release when recording this event. This flag is useful to make non-coherent host memory visible to the host. The flag is a no-op on CUDA platforms.

hipHostMallocDefault#: Default pinned memory allocation on the host.

hipHostMallocPortable#: Memory is considered allocated by all contexts.

hipHostMallocMapped#: Map the allocation into the address space for the current device. The device pointer can be obtained with hipHostGetDevicePointer.

hipHostMallocWriteCombined#: Allocates the memory as write-combined. On some system configurations, write-combined allocation may be transferred faster across the PCI Express bus, however, could have low read efficiency by most CPUs. It’s a good option for data transfer from host to device via mapped pinned memory.

hipHostMallocNumaUser#: Host memory allocation will follow numa policy set by user.

Note

This numa allocation flag is applicable on Linux, under development on Windows.

hipHostMallocCoherent#: Allocate coherent memory. Overrides HIP_HOST_COHERENT for specific allocation.

hipHostMallocNonCoherent#: Allocate non-coherent memory. Overrides HIP_HOST_COHERENT for specific allocation.

hipMemAttachGlobal#: Memory can be accessed by any stream on any device

hipMemAttachHost#: Memory cannot be accessed by any stream on any device.

hipMemAttachSingle#: Memory can only be accessed by a single stream on the associated device.

hipDeviceMallocDefault#

hipDeviceMallocFinegrained#: Memory is allocated in fine grained region of device.

hipMallocSignalMemory#: Memory represents a HSA signal.

hipDeviceMallocUncached#: Memory allocated will be uncached.

hipDeviceMallocContiguous#: Memory allocated will be contiguous.

hipHostRegisterDefault#: Memory is Mapped and Portable.

hipHostRegisterPortable#: Memory is considered registered by all contexts.

hipHostRegisterMapped#: Map the allocation into the address space for the current device. The device pointer can be obtained with hipHostGetDevicePointer.

hipHostRegisterIoMemory#: Not supported.

hipHostRegisterReadOnly#: This flag is ignored On AMD devices.

hipExtHostRegisterCoarseGrained#: Coarse Grained host memory lock.

hipDeviceScheduleAuto#: Automatically select between Spin and Yield.

hipDeviceScheduleSpin#: Dedicate a CPU core to spin-wait. Provides lowest latency, but burns a CPU core and may consume more power.

hipDeviceScheduleYield#: Yield the CPU to the operating system when waiting. May increase latency, but lowers power and is friendlier to other threads in the system.

hipDeviceScheduleBlockingSync#

hipDeviceScheduleMask#

hipDeviceMapHost#

hipDeviceLmemResizeToMax#

hipArrayDefault#: Default HIP array allocation flag.

hipArrayLayered#

hipArraySurfaceLoadStore#

hipArrayCubemap#

hipArrayTextureGather#

hipOccupancyDefault#

hipOccupancyDisableCachingOverride#

hipCooperativeLaunchMultiDeviceNoPreSync#

hipCooperativeLaunchMultiDeviceNoPostSync#

hipCpuDeviceId#

hipInvalidDeviceId#

hipExtAnyOrderLaunch#: AnyOrderLaunch of kernels.

hipStreamWaitValueGte#

hipStreamWaitValueEq#

hipStreamWaitValueAnd#

hipStreamWaitValueNor#

hipStreamPerThread#: Implicit stream per application thread.

hipStreamLegacy#

hipExternalMemoryDedicated#

hipKernelNodeAttrID#: Kernel node attributeID

hipKernelNodeAttributeAccessPolicyWindow#

hipKernelNodeAttributeCooperative#

hipKernelNodeAttributePriority#

hipKernelNodeAttrValue#: Kernel node attribute value

hipGraphKernelNodePortDefault#: This port activates when the kernel has finished executing.

hipGraphKernelNodePortLaunchCompletion#: This port activates when all blocks of the kernel have begun execution.

hipGraphKernelNodePortProgrammatic#: This port activates when all blocks of the kernel have performed hipTriggerProgrammaticLaunchCompletion() or have terminated. It must be used with edge type hipGraphDependencyTypeProgrammatic.

struct hipDeviceArch_t#: #include <hip_runtime_api.h>

hipDeviceArch_t

struct hipUUID#: #include <hip_runtime_api.h>

struct hipDeviceProp_t#: #include <hip_runtime_api.h>

hipDeviceProp

struct hipPointerAttribute_t#: #include <hip_runtime_api.h>

Pointer attributes

struct hipIpcMemHandle_t#: #include <hip_runtime_api.h>

struct hipIpcEventHandle_t#: #include <hip_runtime_api.h>

struct hipFuncAttributes#: #include <hip_runtime_api.h>

struct hipMemLocation#

#include <hip_runtime_api.h>

Specifies a memory location.

To specify a gpu, set type = hipMemLocationTypeDevice and set id = the gpu’s device ID

struct hipMemAccessDesc#: #include <hip_runtime_api.h>

Memory access descriptor

struct hipMemPoolProps#: #include <hip_runtime_api.h>

Specifies the properties of allocations made from the pool.

struct hipMemPoolPtrExportData#: #include <hip_runtime_api.h>

Opaque data structure for exporting a pool allocation

struct dim3#: #include <hip_runtime_api.h>

Struct for data in 3D

struct hipLaunchParams#: #include <hip_runtime_api.h>

struct hipLaunchParams_t

struct hipFunctionLaunchParams#: #include <hip_runtime_api.h>

struct hipFunctionLaunchParams_t

struct hipExternalMemoryHandleDesc#: #include <hip_runtime_api.h>

struct hipExternalMemoryBufferDesc#: #include <hip_runtime_api.h>

struct hipExternalMemoryMipmappedArrayDesc#: #include <hip_runtime_api.h>

struct hipExternalSemaphoreHandleDesc#: #include <hip_runtime_api.h>

struct hipExternalSemaphoreSignalParams#: #include <hip_runtime_api.h>

struct hipExternalSemaphoreWaitParams#: #include <hip_runtime_api.h>

External semaphore wait parameters, compatible with driver type

struct hipHostNodeParams#: #include <hip_runtime_api.h>

struct hipKernelNodeParams#: #include <hip_runtime_api.h>

struct hipMemsetParams#: #include <hip_runtime_api.h>

struct hipMemAllocNodeParams#: #include <hip_runtime_api.h>

struct hipAccessPolicyWindow#: #include <hip_runtime_api.h>

union hipLaunchAttributeValue#

#include <hip_runtime_api.h>

Launch Attribute Value

Public Members

hipAccessPolicyWindow accessPolicyWindow#: Value of launch attribute:: hipLaunchAttributePolicyWindow.

int cooperative#: Value of launch attribute hipLaunchAttributeCooperative

int priority#: Value of launch attribute :: hipLaunchAttributePriority. Execution priority of kernel.

struct HIP_MEMSET_NODE_PARAMS#: #include <hip_runtime_api.h>

Memset node params

struct hipGraphInstantiateParams#: #include <hip_runtime_api.h>

Graph Instantiation parameters

struct hipMemAllocationProp#: #include <hip_runtime_api.h>

Memory allocation properties

struct hipExternalSemaphoreSignalNodeParams#: #include <hip_runtime_api.h>

External semaphore signal node parameters

struct hipExternalSemaphoreWaitNodeParams#: #include <hip_runtime_api.h>

External semaphore wait node parameters

struct hipArrayMapInfo#: #include <hip_runtime_api.h>

Map info for arrays

struct hipMemcpyNodeParams#: #include <hip_runtime_api.h>

Memcpy node params

struct hipChildGraphNodeParams#: #include <hip_runtime_api.h>

Child graph node params

struct hipEventWaitNodeParams#: #include <hip_runtime_api.h>

Event record node params

struct hipEventRecordNodeParams#: #include <hip_runtime_api.h>

Event record node params

struct hipMemFreeNodeParams#: #include <hip_runtime_api.h>

Memory free node params

struct hipGraphNodeParams#: #include <hip_runtime_api.h>

Params for different graph nodes

struct hipGraphEdgeData#: #include <hip_runtime_api.h>

Global enum and defines

Contents

Global enum and defines#