RPP common definitions

RPP common definitions#

enum RppStatus#

RPP RppStatus type enums.

Values:

enumerator RPP_SUCCESS#: No error.

enumerator RPP_ERROR#: Unspecified error.

enumerator RPP_ERROR_INVALID_ARGUMENTS#: One or more arguments invalid. (Needs to adhere to function specification.) group_rppdefs.

enumerator RPP_ERROR_LOW_OFFSET#: Low tensor offsetInBytes provided for src/dst tensor.

enumerator RPP_ERROR_ZERO_DIVISION#: Arguments provided will result in zero division error.

enumerator RPP_ERROR_HIGH_SRC_DIMENSION#: Src tensor / src ROI dimension too high. (Needs to adhere to function specification.)

enumerator RPP_ERROR_NOT_IMPLEMENTED#: Function variant requested is not implemented / unsupported.

enumerator RPP_ERROR_INVALID_SRC_CHANNELS#: Invalid src tensor number of channels. (Needs to adhere to function specification.)

enumerator RPP_ERROR_INVALID_DST_CHANNELS#: Invalid dst tensor number of channels. (Needs to adhere to function specification.)

enumerator RPP_ERROR_INVALID_SRC_LAYOUT#: Invalid src tensor layout. (Needs to adhere to function specification.) group_rppdefs.

enumerator RPP_ERROR_INVALID_DST_LAYOUT#: Invalid dst tensor layout. (Needs to adhere to function specification.) group_rppdefs.

enumerator RPP_ERROR_INVALID_SRC_DATATYPE#: Invalid src tensor datatype. (Needs to adhere to function specification.) group_rppdefs.

enumerator RPP_ERROR_INVALID_DST_DATATYPE#: Invalid dst tensor datatype. (Needs to adhere to function specification.) group_rppdefs.

enumerator RPP_ERROR_INVALID_SRC_OR_DST_DATATYPE#: Invalid src/dst tensor datatype. (Needs to adhere to function specification.)

enumerator RPP_ERROR_INSUFFICIENT_DST_BUFFER_LENGTH#: Insufficient dst buffer length provided. (Needs to adhere to function specification.)

enumerator RPP_ERROR_INVALID_PARAMETER_DATATYPE#: Invalid datatype.

enumerator RPP_ERROR_NOT_ENOUGH_MEMORY#: Not enough memory to write outputs, as per dim-lengths and strides set in descriptor.

enumerator RPP_ERROR_OUT_OF_BOUND_SRC_ROI#: Out of bound source ROI.

enumerator RPP_ERROR_LAYOUT_MISMATCH#: src and dst layout mismatch

enumerator RPP_ERROR_INVALID_CHANNELS#: Number of channels is invalid. (Needs to adhere to function specification.) group_rppdefs.

enumerator RPP_ERROR_INVALID_OUTPUT_TILE_LENGTH#: Invalid output tile length (Needs to adhere to function specification.) group_rppdefs.

enumerator RPP_ERROR_OUT_OF_BOUND_SHARED_MEMORY_SIZE#: Shared memory size needed is beyond the bounds (Needs to adhere to function specification.)

enumerator RPP_ERROR_OUT_OF_BOUND_SCRATCH_MEMORY_SIZE#: Scratch memory size needed is beyond the bounds (Needs to adhere to function specification.)

enumerator RPP_ERROR_INVALID_SRC_DIMS#: Number of src dims is invalid. (Needs to adhere to function specification.) group_rppdefs.

enumerator RPP_ERROR_INVALID_DST_DIMS#: Number of dst dims is invalid. (Needs to adhere to function specification.) group_rppdefs.

enumerator RPP_ERROR_INVALID_DIM_LENGTHS#: Length in some dimensions are invalid. (Needs to adhere to function specification.)

enumerator RPP_ERROR_INVALID_AXIS#: The specified axis is invalid or out of range. (Needs to adhere to function specification.)

enumerator RPP_ERROR_INCOMPATIBLE_BACKEND#: The user specified backend is not compatible with the initialized handle group_rppdefs.

enumerator RPP_ERROR_HIP_LAUNCH#: HIP/GPU runtime or kernel launch error.

enumerator RPP_ERROR_HIP_RUNTIME#: Internal HIP/GPU runtime error.

enum RppBackend#

RPP RppBackend type enums.

Values:

enumerator RPP_HOST_BACKEND#

enumerator RPP_HIP_BACKEND#

enum rppStatus_t#

RPP rppStatus_t type enums.

Values:

enumerator rppStatusSuccess#

enumerator rppStatusBadParm#

enumerator rppStatusUnknownError#

enumerator rppStatusNotInitialized#

enumerator rppStatusInvalidValue#

enumerator rppStatusAllocFailed#

enumerator rppStatusInternalError#

enumerator rppStatusNotImplemented#

enumerator rppStatusUnsupportedOp#

enum RpptOp#

RPP Tensor Operations type enum.

Values:

enumerator RPP_TENSOR_OP_ADD#

enumerator RPP_TENSOR_OP_SUBTRACT#

enumerator RPP_TENSOR_OP_MULTIPLY#

enumerator RPP_TENSOR_OP_DIVIDE#

enum RpptBitwiseOp#

RPPT Broadcast Mode type enum.

Values:

enumerator RPP_TENSOR_OP_AND#

enumerator RPP_TENSOR_OP_OR#

enumerator RPP_TENSOR_OP_XOR#

enum RpptDataType#

RPPT Tensor datatype enum.

Values:

enumerator U8#

enumerator F32#

enumerator F16#

enumerator I8#

enumerator I16#

enumerator U16#

enumerator I32#

enumerator U32#

enum RpptLayout#

RPPT Tensor layout type enum.

Values:

enumerator NCHW#

enumerator NHWC#

enumerator NCDHW#

enumerator NDHWC#

enumerator NHW#

enumerator NFT#

enumerator NTF#

enum RpptColorStandard#

Color space standard for NV12 to RGB (YUV/YCbCr to RGB matrix).

Selects which industry color space standard defines the luma coefficients (wr, wb).

Values:

enumerator RpptColorStandard_BT709#: ITU-R BT.709 (default wr=0.2126, wb=0.0722)

enumerator RpptColorStandard_FCC#: FCC

enumerator RpptColorStandard_BT470BG#: ITU-R BT.470 System B, G

enumerator RpptColorStandard_BT601#: ITU-R BT.601 / SMPTE 170M

enumerator RpptColorStandard_SMPTE240M#: SMPTE 240M

enumerator RpptColorStandard_BT2020_NCL#: ITU-R BT.2020 non-constant luminance

enumerator RpptColorStandard_BT2020_CL#: ITU-R BT.2020 constant luminance

enum RpptColorRange#

Color range (luma/chroma legal levels) for NV12 to RGB.

Color range specifies how 8-bit Y (and scaling of the matrix) is interpreted: limited (studio / TV / MPEG) vs full (JPEG / PC). RpptColorRange_STUDIO: luma nominally 16–235, chroma centered with Y bias 16. RpptColorRange_FULL: luma 0–255, Y bias 0.

Values:

enumerator RpptColorRange_STUDIO#: Limited / MPEG / TV range

enumerator RpptColorRange_FULL#: Full / JPEG / PC range

enum RpptRoiType#

RPPT Tensor 2D ROI type enum.

Values:

enumerator LTRB#

enumerator XYWH#

enum RpptRoi3DType#

RPPT Tensor 3D ROI type enum.

Values:

enumerator LTFRBB#

enumerator XYZWHD#

enum RpptSubpixelLayout#

RPPT Tensor subpixel layout type enum.

Values:

enumerator RGBtype#

enumerator BGRtype#

enum RpptInterpolationType#

RPPT Tensor interpolation type enum.

Values:

enumerator NEAREST_NEIGHBOR#

enumerator BILINEAR#

enumerator BICUBIC#

enumerator LANCZOS#

enumerator GAUSSIAN#

enumerator TRIANGULAR#

enum RpptAudioBorderType#

RPPT Audio Border Type.

Values:

enumerator ZERO#

enumerator CLAMP#

enumerator REFLECT#

enum RpptImageBorderEdge#

RPPT Image Border Edge type enum.

Values:

enumerator TOP_EDGE#

enumerator BOTTOM_EDGE#

enumerator LEFT_EDGE#

enumerator RIGHT_EDGE#

enum RpptImageBorderType#

RPPT Image Border Type.

Values:

enumerator REPLICATE#

enumerator CONSTANT#

enumerator REFLECT_NO_EDGE#

enum RpptMelScaleFormula#

RPPT Mel Scale Formula.

Values:

enumerator SLANEY#

enumerator HTK#

typedef unsigned char Rpp8u#: 8 bit unsigned char

typedef signed char Rpp8s#: 8 bit signed char

typedef unsigned short Rpp16u#: 16 bit unsigned short

typedef short Rpp16s#: 16 bit signed short

typedef unsigned int Rpp32u#: 32 bit unsigned int

typedef int Rpp32s#: 32 bit signed int

typedef unsigned long long Rpp64u#: 64 bit unsigned long long

typedef long long Rpp64s#: 64 bit long long

typedef float Rpp32f#: 32 bit float

typedef double Rpp64f#: 64 bit double

typedef void *RppPtr_t#: void pointer

typedef size_t RppSize_t#: size_t

typedef struct GenericFilter GenericFilter#: RPPT Tensor GenericFilter type struct.

typedef struct RpptResamplingWindow RpptResamplingWindow#: RPPT Tensor RpptResamplingWindow type struct.

inline const char *rppGetStatusString(RppStatus status) noexcept#

Get the string representation of a RPP status code.

Function to convert a RppStatus value returned by any RPP tensor API function into its enumerator name, for logging and error reporting.

Parameters:: status – [in] A RppStatus enumeration.
Returns:: A null-terminated string with static storage duration holding the enumerator name (for example “RPP_ERROR_INVALID_SRC_LAYOUT”), or “RPP_STATUS_UNKNOWN” if the value does not name an enumerator. The caller must not free the returned pointer.

RPP_MIN_8U#: 8 bit unsigned char minimum \

RPP_MAX_8U#: 8 bit unsigned char maximum \

RPPT_MAX_DIMS#: RPP maximum dimensions in tensor .

RPPT_MAX_AUDIO_CHANNELS#: RPP maximum channels in audio tensor .

RPP_RETURN_IF_ERROR(expr)#: If a nested RPP call fails, return its RppStatus from the enclosing function. Use when forwarding the result of a *_host_tensor or hip_exec_* helper.

RPP_HIP_RETURN_IF_ERROR(expr)#: If a HIP API call fails, return RPP_ERROR_HIP_RUNTIME from the enclosing function and print the error message to stderr.

HIP_CHECK_LAUNCH_RETURN()#: Check last HIP error after kernel launch; return RPP_ERROR_HIP_LAUNCH on failure. Use after hipLaunchKernelGGL.

struct RppLayoutParams#: #include <rppdefs.h>

RPP layout params.

struct Rpp32f6#: #include <rppdefs.h>

RPP 6 float vector.

struct Rpp32f9#: #include <rppdefs.h>

RPP 9 float vector.

struct Rpp32s24#: #include <rppdefs.h>

RPP 24 signed int vector.

struct Rpp32f24#: #include <rppdefs.h>

RPP 24 float vector.

struct RpptPoint2D#: #include <rppdefs.h>

RPPT Image 2D cartesian point type struct.

struct RpptPoint3D#: #include <rppdefs.h>

RPPT Image 3D point type struct.

struct RpptRoiLtrb#: #include <rppdefs.h>

RPPT Tensor 2D ROI LTRB struct.

struct RpptChannelOffsets#: #include <rppdefs.h>

RPPT Tensor Channel Offsets struct.

struct RpptRoiLtfrbb#: #include <rppdefs.h>

RPPT Tensor 3D ROI LTFRBB struct.

struct RpptRoiXywh#: #include <rppdefs.h>

RPPT Tensor 2D ROI XYWH struct.

struct RpptRoiXyzwhd#: #include <rppdefs.h>

RPPT Tensor 3D ROI XYZWHD struct.

union RpptROI#

#include <rppdefs.h>

RPPT Tensor 2D ROI union.

Public Members

RpptRoiLtrb ltrbROI#

RpptRoiXywh xywhROI#

union RpptROI3D#

#include <rppdefs.h>

RPPT Tensor 3D ROI union.

Public Members

RpptRoiLtfrbb ltfrbbROI#

RpptRoiXyzwhd xyzwhdROI#

struct RpptStrides#: #include <rppdefs.h>

RPPT Tensor strides type struct.

struct RpptDesc#: #include <rppdefs.h>

RPPT Tensor descriptor type struct.

struct RpptGenericDesc#: #include <rppdefs.h>

RPPT Tensor Generic descriptor type struct.

struct RpptRGB#: #include <rppdefs.h>

RPPT Tensor 8-bit uchar RGB type struct.

struct RpptFloatRGB#: #include <rppdefs.h>

RPPT Tensor 32-bit float RGB type struct.

struct RpptUintVector2D#: #include <rppdefs.h>

RPPT Tensor 2D 32-bit uint vector type struct.

struct RpptFloatVector2D#: #include <rppdefs.h>

RPPT Tensor 2D 32-bit float vector type struct.

struct RpptImagePatch#: #include <rppdefs.h>

RPPT Tensor 2D image patch dimensions type struct.

struct RpptXorwowState#: #include <rppdefs.h>

RPPT Tensor random number generator state (xorwow state) type struct.

struct RpptXorwowStateBoxMuller#: #include <rppdefs.h>

RPPT Tensor random number generator state (xorwow box muller state) type struct.

struct RpptBilinearNbhoodLocsVecLen8#: #include <rppdefs.h>

RPPT Tensor 2D bilinear neighborhood 32-bit signed int 8-length-vectors type struct.

struct RpptBilinearNbhoodValsVecLen8#: #include <rppdefs.h>

RPPT Tensor 2D bilinear neighborhood 32-bit float 8-length-vectors type struct.

struct GenericFilter: #include <rppdefs.h>

RPPT Tensor GenericFilter type struct.

struct RpptResamplingWindow: #include <rppdefs.h>

RPPT Tensor RpptResamplingWindow type struct.

struct BaseMelScale#

#include <rppdefs.h>

Base class for Mel scale conversions.

Subclassed by HtkMelScale, SlaneyMelScale

struct HtkMelScale : public BaseMelScale #: #include <rppdefs.h>

Derived class for HTK Mel scale conversions.

struct SlaneyMelScale : public BaseMelScale #: #include <rppdefs.h>

Derived class for Slaney Mel scale conversions.

struct memCPU#: #include <rppdefs.h>

RPP HOST memory type struct.

struct hipMemRpp32f#: #include <rppdefs.h>

RPP HIP 32-bit float memory.

struct memGPU#: #include <rppdefs.h>

RPP HIP memory management type struct.

struct memMgmt#: #include <rppdefs.h>

RPP HIP-HOST memory management.

struct InitHandle#: #include <rppdefs.h>

RPP initialize handle.

RPP common definitions

Contents

RPP common definitions#