HIP: Heterogenous-computing Interface for Portability
|
Go to the documentation of this file.
30 #ifndef HIP_INCLUDE_HIP_HIP_RUNTIME_API_H
31 #define HIP_INCLUDE_HIP_HIP_RUNTIME_API_H
35 #include <hip/hip_version.h>
36 #include <hip/hip_common.h>
40 HIP_ERROR_INVALID_VALUE,
41 HIP_ERROR_NOT_INITIALIZED,
42 HIP_ERROR_LAUNCH_OUT_OF_RESOURCES
175 unsigned allocationFlags;
189 #if __cplusplus >= 201703L
190 #define __HIP_NODISCARD [[nodiscard]]
192 #define __HIP_NODISCARD
203 typedef enum __HIP_NODISCARD hipError_t {
206 hipErrorOutOfMemory = 2,
210 hipErrorNotInitialized = 3,
212 hipErrorInitializationError = 3,
213 hipErrorDeinitialized = 4,
214 hipErrorProfilerDisabled = 5,
215 hipErrorProfilerNotInitialized = 6,
216 hipErrorProfilerAlreadyStarted = 7,
217 hipErrorProfilerAlreadyStopped = 8,
218 hipErrorInvalidConfiguration = 9,
219 hipErrorInvalidPitchValue = 12,
220 hipErrorInvalidSymbol = 13,
223 hipErrorInsufficientDriver = 35,
224 hipErrorMissingConfiguration = 52,
225 hipErrorPriorLaunchFailure = 53,
226 hipErrorInvalidDeviceFunction = 98,
229 hipErrorInvalidImage = 200,
231 hipErrorContextAlreadyCurrent = 202,
232 hipErrorMapFailed = 205,
235 hipErrorUnmapFailed = 206,
236 hipErrorArrayIsMapped = 207,
237 hipErrorAlreadyMapped = 208,
238 hipErrorNoBinaryForGpu = 209,
239 hipErrorAlreadyAcquired = 210,
240 hipErrorNotMapped = 211,
241 hipErrorNotMappedAsArray = 212,
242 hipErrorNotMappedAsPointer = 213,
243 hipErrorECCNotCorrectable = 214,
244 hipErrorUnsupportedLimit = 215,
245 hipErrorContextAlreadyInUse = 216,
246 hipErrorPeerAccessUnsupported = 217,
248 hipErrorInvalidGraphicsContext = 219,
249 hipErrorInvalidSource = 300,
250 hipErrorFileNotFound = 301,
251 hipErrorSharedObjectSymbolNotFound = 302,
252 hipErrorSharedObjectInitFailed = 303,
253 hipErrorOperatingSystem = 304,
254 hipErrorInvalidHandle = 400,
258 hipErrorNotFound = 500,
260 hipErrorIllegalAddress = 700,
265 hipErrorLaunchTimeOut = 702,
270 hipErrorSetOnActiveProcess = 708,
271 hipErrorContextIsDestroyed = 709,
307 hipErrorUnknown = 999,
319 #undef __HIP_NODISCARD
327 hipDeviceAttributeCudaCompatibleBegin = 0,
424 hipDeviceAttributeCudaCompatibleEnd = 9999,
425 hipDeviceAttributeAmdSpecificBegin = 10000,
448 hipDeviceAttributeAmdSpecificEnd = 19999,
450 hipDeviceAttributeVendorSpecificBegin = 20000,
454 enum hipComputeMode {
455 hipComputeModeDefault = 0,
456 hipComputeModeExclusive = 1,
457 hipComputeModeProhibited = 2,
458 hipComputeModeExclusiveProcess = 3
465 #if (defined(__HIP_PLATFORM_HCC__) || defined(__HIP_PLATFORM_AMD__)) && !(defined(__HIP_PLATFORM_NVCC__) || defined(__HIP_PLATFORM_NVIDIA__))
469 #ifndef GENERIC_GRID_LAUNCH
470 #define GENERIC_GRID_LAUNCH 1
472 #include <hip/amd_detail/host_defines.h>
473 #include <hip/driver_types.h>
474 #include <hip/texture_types.h>
476 #if defined(_MSC_VER)
477 #define DEPRECATED(msg) __declspec(deprecated(msg))
478 #else // !defined(_MSC_VER)
479 #define DEPRECATED(msg) __attribute__ ((deprecated(msg)))
480 #endif // !defined(_MSC_VER)
481 #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"
482 #define HIP_LAUNCH_PARAM_BUFFER_POINTER ((void*)0x01)
483 #define HIP_LAUNCH_PARAM_BUFFER_SIZE ((void*)0x02)
484 #define HIP_LAUNCH_PARAM_END ((void*)0x03)
492 #pragma GCC visibility push (default)
496 hipError_t hip_init();
505 typedef struct ihipCtx_t* hipCtx_t;
507 typedef int hipDevice_t;
508 typedef enum hipDeviceP2PAttr {
509 hipDevP2PAttrPerformanceRank = 0,
510 hipDevP2PAttrAccessSupported,
511 hipDevP2PAttrNativeAtomicSupported,
512 hipDevP2PAttrHipArrayAccessSupported
514 typedef struct ihipStream_t* hipStream_t;
515 #define hipIpcMemLazyEnablePeerAccess 0
516 #define HIP_IPC_HANDLE_SIZE 64
518 char reserved[HIP_IPC_HANDLE_SIZE];
521 char reserved[HIP_IPC_HANDLE_SIZE];
523 typedef struct ihipModule_t* hipModule_t;
524 typedef struct ihipModuleSymbol_t* hipFunction_t;
528 size_t constSizeBytes;
529 size_t localSizeBytes;
530 int maxDynamicSharedSizeBytes;
531 int maxThreadsPerBlock;
533 int preferredShmemCarveout;
535 size_t sharedSizeBytes;
537 typedef struct ihipEvent_t* hipEvent_t;
539 hipLimitPrintfFifoSize = 0x01,
540 hipLimitMallocHeapSize = 0x02,
548 #define hipStreamDefault 0x00
551 #define hipStreamNonBlocking 0x01
555 #define hipEventDefault 0x0
558 #define hipEventBlockingSync 0x1
561 #define hipEventDisableTiming 0x2
564 #define hipEventInterprocess 0x4
568 #define hipEventReleaseToDevice 0x40000000
572 #define hipEventReleaseToSystem 0x80000000
576 #define hipHostMallocDefault 0x0
579 #define hipHostMallocPortable 0x1
583 #define hipHostMallocMapped 0x2
588 #define hipHostMallocWriteCombined 0x4
591 #define hipHostMallocNumaUser 0x20000000
594 #define hipHostMallocCoherent 0x40000000
597 #define hipHostMallocNonCoherent 0x80000000
600 #define hipMemAttachGlobal 0x01
603 #define hipMemAttachHost 0x02
606 #define hipMemAttachSingle 0x04
608 #define hipDeviceMallocDefault 0x0
611 #define hipDeviceMallocFinegrained 0x1
614 #define hipMallocSignalMemory 0x2
618 #define hipHostRegisterDefault 0x0
621 #define hipHostRegisterPortable 0x1
625 #define hipHostRegisterMapped 0x2
628 #define hipHostRegisterIoMemory 0x4
631 #define hipExtHostRegisterCoarseGrained 0x8
634 #define hipDeviceScheduleAuto 0x0
638 #define hipDeviceScheduleSpin 0x1
642 #define hipDeviceScheduleYield 0x2
643 #define hipDeviceScheduleBlockingSync 0x4
644 #define hipDeviceScheduleMask 0x7
645 #define hipDeviceMapHost 0x8
646 #define hipDeviceLmemResizeToMax 0x16
648 #define hipArrayDefault 0x00
649 #define hipArrayLayered 0x01
650 #define hipArraySurfaceLoadStore 0x02
651 #define hipArrayCubemap 0x04
652 #define hipArrayTextureGather 0x08
653 #define hipOccupancyDefault 0x00
654 #define hipCooperativeLaunchMultiDeviceNoPreSync 0x01
655 #define hipCooperativeLaunchMultiDeviceNoPostSync 0x02
656 #define hipCpuDeviceId ((int)-1)
657 #define hipInvalidDeviceId ((int)-2)
660 #define hipExtAnyOrderLaunch 0x01
662 #define hipStreamWaitValueGte 0x0
663 #define hipStreamWaitValueEq 0x1
664 #define hipStreamWaitValueAnd 0x2
665 #define hipStreamWaitValueNor 0x3
668 #define hipStreamPerThread ((hipStream_t)2)
727 typedef enum hipJitOption {
728 hipJitOptionMaxRegisters = 0,
729 hipJitOptionThreadsPerBlock,
730 hipJitOptionWallTime,
731 hipJitOptionInfoLogBuffer,
732 hipJitOptionInfoLogBufferSizeBytes,
733 hipJitOptionErrorLogBuffer,
734 hipJitOptionErrorLogBufferSizeBytes,
735 hipJitOptionOptimizationLevel,
736 hipJitOptionTargetFromContext,
738 hipJitOptionFallbackStrategy,
739 hipJitOptionGenerateDebugInfo,
740 hipJitOptionLogVerbose,
741 hipJitOptionGenerateLineInfo,
742 hipJitOptionCacheMode,
744 hipJitOptionFastCompile,
745 hipJitOptionNumOptions
751 hipFuncAttributeMaxDynamicSharedMemorySize = 8,
752 hipFuncAttributePreferredSharedMemoryCarveout = 9,
783 constexpr __host__ __device__
dim3(uint32_t _x = 1, uint32_t _y = 1, uint32_t _z = 1) :
x(_x),
y(_y),
z(_z){};
794 typedef enum hipExternalMemoryHandleType_enum {
795 hipExternalMemoryHandleTypeOpaqueFd = 1,
796 hipExternalMemoryHandleTypeOpaqueWin32 = 2,
797 hipExternalMemoryHandleTypeOpaqueWin32Kmt = 3,
798 hipExternalMemoryHandleTypeD3D12Heap = 4,
799 hipExternalMemoryHandleTypeD3D12Resource = 5,
800 hipExternalMemoryHandleTypeD3D11Resource = 6,
801 hipExternalMemoryHandleTypeD3D11ResourceKmt = 7,
802 } hipExternalMemoryHandleType;
804 hipExternalMemoryHandleType type;
812 unsigned long long size;
816 unsigned long long offset;
817 unsigned long long size;
820 typedef void* hipExternalMemory_t;
821 typedef enum hipExternalSemaphoreHandleType_enum {
822 hipExternalSemaphoreHandleTypeOpaqueFd = 1,
823 hipExternalSemaphoreHandleTypeOpaqueWin32 = 2,
824 hipExternalSemaphoreHandleTypeOpaqueWin32Kmt = 3,
825 hipExternalSemaphoreHandleTypeD3D12Fence = 4
826 } hipExternalSemaphoreHandleType;
828 hipExternalSemaphoreHandleType type;
838 typedef void* hipExternalSemaphore_t;
842 unsigned long long value;
845 unsigned long long key;
847 unsigned int reserved[12];
850 unsigned int reserved[16];
858 unsigned long long value;
861 unsigned long long key;
862 unsigned int timeoutMs;
864 unsigned int reserved[10];
867 unsigned int reserved[16];
870 #if __HIP_HAS_GET_PCH
876 void __hipGetPCH(
const char** pch,
unsigned int*size);
898 hipGraphicsRegisterFlagsNone = 0,
907 typedef struct _hipGraphicsResource hipGraphicsResource;
909 typedef hipGraphicsResource* hipGraphicsResource_t;
1011 int srcDevice,
int dstDevice);
1720 unsigned int flags);
1765 uint32_t mask __dparm(0xFFFFFFFF));
1799 uint64_t mask __dparm(0xFFFFFFFFFFFFFFFF));
1914 hipError_t
hipEventRecord(hipEvent_t event, hipStream_t stream = NULL);
2048 hipDeviceptr_t ptr);
2066 void** data, hipDeviceptr_t ptr);
2093 unsigned int numExtSems, hipStream_t stream);
2108 unsigned int numExtSems, hipStream_t stream);
2194 DEPRECATED(
"use hipHostMalloc instead")
2259 hipStream_t stream __dparm(0));
2289 const
void* dev_ptr,
2308 size_t num_attributes,
2309 const
void* dev_ptr,
2325 size_t length __dparm(0),
2454 hipError_t
hipMemAllocPitch(hipDeviceptr_t* dptr,
size_t* pitch,
size_t widthInBytes,
size_t height,
unsigned int elementSizeBytes);
2523 hipError_t
hipMemcpy(
void* dst, const
void* src,
size_t sizeBytes, hipMemcpyKind kind);
2525 hipError_t hipMemcpyWithStream(
void* dst, const
void* src,
size_t sizeBytes,
2526 hipMemcpyKind kind, hipStream_t stream);
2580 hipError_t
hipMemcpyDtoD(hipDeviceptr_t dst, hipDeviceptr_t src,
size_t sizeBytes);
2635 hipStream_t stream);
2652 hipModule_t hmod, const
char* name);
2695 size_t sizeBytes,
size_t offset __dparm(0),
2696 hipMemcpyKind kind __dparm(hipMemcpyHostToDevice));
2712 size_t sizeBytes,
size_t offset,
2713 hipMemcpyKind kind, hipStream_t stream __dparm(0));
2728 size_t sizeBytes,
size_t offset __dparm(0),
2729 hipMemcpyKind kind __dparm(hipMemcpyDeviceToHost));
2745 size_t sizeBytes,
size_t offset,
2747 hipStream_t stream __dparm(0));
2776 hipError_t
hipMemcpyAsync(
void* dst, const
void* src,
size_t sizeBytes, hipMemcpyKind kind,
2777 hipStream_t stream __dparm(0));
2787 hipError_t
hipMemset(
void* dst,
int value,
size_t sizeBytes);
2797 hipError_t
hipMemsetD8(hipDeviceptr_t dest,
unsigned char value,
size_t count);
2813 hipError_t
hipMemsetD8Async(hipDeviceptr_t dest,
unsigned char value,
size_t count, hipStream_t stream __dparm(0));
2823 hipError_t
hipMemsetD16(hipDeviceptr_t dest,
unsigned short value,
size_t count);
2839 hipError_t
hipMemsetD16Async(hipDeviceptr_t dest,
unsigned short value,
size_t count, hipStream_t stream __dparm(0));
2865 hipError_t
hipMemsetAsync(
void* dst,
int value,
size_t sizeBytes, hipStream_t stream __dparm(0));
2882 hipStream_t stream __dparm(0));
2893 hipError_t
hipMemset2D(
void* dst,
size_t pitch,
int value,
size_t width,
size_t height);
2905 hipError_t
hipMemset2DAsync(
void* dst,
size_t pitch,
int value,
size_t width,
size_t height,hipStream_t stream __dparm(0));
2935 hipError_t hipMemPtrGetInfo(
void* ptr,
size_t* size);
2949 size_t height __dparm(0),
unsigned int flags __dparm(
hipArrayDefault));
2951 hipError_t hipArrayDestroy(
hipArray* array);
2983 struct
hipExtent extent,
unsigned int flags);
2999 unsigned int numLevels,
3000 unsigned int flags __dparm(0));
3013 unsigned int level);
3030 hipError_t
hipMemcpy2D(
void* dst,
size_t dpitch, const
void* src,
size_t spitch,
size_t width,
3031 size_t height, hipMemcpyKind kind);
3070 hipError_t
hipMemcpy2DAsync(
void* dst,
size_t dpitch, const
void* src,
size_t spitch,
size_t width,
3071 size_t height, hipMemcpyKind kind, hipStream_t stream __dparm(0));
3090 size_t spitch,
size_t width,
size_t height, hipMemcpyKind kind);
3110 size_t spitch,
size_t width,
size_t height, hipMemcpyKind kind,
3111 hipStream_t stream __dparm(0));
3127 DEPRECATED(DEPRECATED_MSG)
3129 size_t count, hipMemcpyKind kind);
3145 DEPRECATED(DEPRECATED_MSG)
3147 size_t count, hipMemcpyKind kind);
3330 #ifndef USE_PEER_NON_UNIFIED
3331 #define USE_PEER_NON_UNIFIED 1
3333 #if USE_PEER_NON_UNIFIED == 1
3345 hipError_t
hipMemcpyPeer(
void* dst,
int dstDeviceId,
const void* src,
int srcDeviceId,
3360 size_t sizeBytes, hipStream_t stream __dparm(0));
3392 DEPRECATED(DEPRECATED_MSG)
3393 hipError_t
hipCtxCreate(hipCtx_t* ctx,
unsigned int flags, hipDevice_t device);
3404 DEPRECATED(DEPRECATED_MSG)
3416 DEPRECATED(DEPRECATED_MSG)
3428 DEPRECATED(DEPRECATED_MSG)
3440 DEPRECATED(DEPRECATED_MSG)
3452 DEPRECATED(DEPRECATED_MSG)
3464 DEPRECATED(DEPRECATED_MSG)
3483 DEPRECATED(DEPRECATED_MSG)
3498 DEPRECATED(DEPRECATED_MSG)
3513 DEPRECATED(DEPRECATED_MSG)
3528 DEPRECATED(DEPRECATED_MSG)
3543 DEPRECATED(DEPRECATED_MSG)
3556 DEPRECATED(DEPRECATED_MSG)
3568 DEPRECATED(DEPRECATED_MSG)
3589 DEPRECATED(DEPRECATED_MSG)
3607 DEPRECATED(DEPRECATED_MSG)
3770 hipJitOption* options,
void** optionValues);
3800 unsigned int gridDimZ,
unsigned int blockDimX,
3801 unsigned int blockDimY,
unsigned int blockDimZ,
3802 unsigned int sharedMemBytes, hipStream_t stream,
3803 void** kernelParams,
void** extra);
3823 void** kernelParams,
unsigned int sharedMemBytes,
3824 hipStream_t stream);
3836 int numDevices,
unsigned int flags);
3849 int numDevices,
unsigned int flags);
3877 hipFunction_t f,
size_t dynSharedMemPerBlk,
3878 int blockSizeLimit);
3896 hipFunction_t f,
size_t dynSharedMemPerBlk,
3897 int blockSizeLimit,
unsigned int flags);
3907 int* numBlocks, hipFunction_t f,
int blockSize,
size_t dynSharedMemPerBlk);
3918 int* numBlocks, hipFunction_t f,
int blockSize,
size_t dynSharedMemPerBlk,
unsigned int flags);
3928 int* numBlocks, const
void* f,
int blockSize,
size_t dynSharedMemPerBlk);
3939 int* numBlocks, const
void* f,
int blockSize,
size_t dynSharedMemPerBlk,
unsigned int flags __dparm(hipOccupancyDefault));
3955 const
void* f,
size_t dynSharedMemPerBlk,
3956 int blockSizeLimit);
3977 DEPRECATED("use roctracer/rocTX instead")
3984 DEPRECATED("use roctracer/rocTX instead")
4052 size_t sharedMem __dparm(0),
4053 hipStream_t stream __dparm(0));
4076 hipStream_t *stream);
4096 size_t sharedMemBytes __dparm(0),
4097 hipStream_t stream __dparm(0));
4128 void** args,
size_t sharedMemBytes, hipStream_t stream,
4129 hipEvent_t startEvent, hipEvent_t stopEvent,
int flags);
4168 const
void* symbol);
4182 hipTextureObject_t* pTexObject,
4221 hipTextureObject_t textureObject);
4234 hipTextureObject_t textureObject);
4247 hipTextureObject_t textureObject);
4252 hipError_t hipTexRefSetAddressMode(
4255 enum hipTextureAddressMode am);
4256 hipError_t hipTexRefSetArray(
4259 unsigned int flags);
4260 hipError_t hipTexRefSetFilterMode(
4262 enum hipTextureFilterMode fm);
4263 hipError_t hipTexRefSetFlags(
4265 unsigned int Flags);
4266 hipError_t hipTexRefSetFormat(
4268 hipArray_Format fmt,
4269 int NumPackedComponents);
4270 hipError_t hipTexObjectCreate(
4271 hipTextureObject_t* pTexObject,
4275 hipError_t hipTexObjectDestroy(
4276 hipTextureObject_t texObject);
4277 hipError_t hipTexObjectGetResourceDesc(
4279 hipTextureObject_t texObject);
4280 hipError_t hipTexObjectGetResourceViewDesc(
4282 hipTextureObject_t texObject);
4283 hipError_t hipTexObjectGetTextureDesc(
4285 hipTextureObject_t texObject);
4294 DEPRECATED(DEPRECATED_MSG)
4295 hipError_t hipBindTexture(
4300 size_t size __dparm(UINT_MAX));
4301 DEPRECATED(DEPRECATED_MSG)
4302 hipError_t hipBindTexture2D(
4310 DEPRECATED(DEPRECATED_MSG)
4311 hipError_t hipBindTextureToArray(
4315 DEPRECATED(DEPRECATED_MSG)
4316 hipError_t hipGetTextureAlignmentOffset(
4319 DEPRECATED(DEPRECATED_MSG)
4321 DEPRECATED(DEPRECATED_MSG)
4322 hipError_t hipTexRefGetAddress(
4323 hipDeviceptr_t* dev_ptr,
4325 DEPRECATED(DEPRECATED_MSG)
4326 hipError_t hipTexRefGetAddressMode(
4327 enum hipTextureAddressMode* pam,
4330 DEPRECATED(DEPRECATED_MSG)
4331 hipError_t hipTexRefGetFilterMode(
4332 enum hipTextureFilterMode* pfm,
4334 DEPRECATED(DEPRECATED_MSG)
4335 hipError_t hipTexRefGetFlags(
4336 unsigned int* pFlags,
4338 DEPRECATED(DEPRECATED_MSG)
4339 hipError_t hipTexRefGetFormat(
4340 hipArray_Format* pFormat,
4343 DEPRECATED(DEPRECATED_MSG)
4344 hipError_t hipTexRefGetMaxAnisotropy(
4347 DEPRECATED(DEPRECATED_MSG)
4348 hipError_t hipTexRefGetMipmapFilterMode(
4349 enum hipTextureFilterMode* pfm,
4351 DEPRECATED(DEPRECATED_MSG)
4352 hipError_t hipTexRefGetMipmapLevelBias(
4355 DEPRECATED(DEPRECATED_MSG)
4356 hipError_t hipTexRefGetMipmapLevelClamp(
4357 float* pminMipmapLevelClamp,
4358 float* pmaxMipmapLevelClamp,
4360 DEPRECATED(DEPRECATED_MSG)
4361 hipError_t hipTexRefGetMipMappedArray(
4364 DEPRECATED(DEPRECATED_MSG)
4365 hipError_t hipTexRefSetAddress(
4368 hipDeviceptr_t dptr,
4370 DEPRECATED(DEPRECATED_MSG)
4371 hipError_t hipTexRefSetAddress2D(
4374 hipDeviceptr_t dptr,
4376 DEPRECATED(DEPRECATED_MSG)
4377 hipError_t hipTexRefSetMaxAnisotropy(
4379 unsigned int maxAniso);
4394 DEPRECATED(DEPRECATED_MSG)
4395 hipError_t hipTexRefSetBorderColor(
4397 float* pBorderColor);
4398 hipError_t hipTexRefSetMipmapFilterMode(
4400 enum hipTextureFilterMode fm);
4401 hipError_t hipTexRefSetMipmapLevelBias(
4404 hipError_t hipTexRefSetMipmapLevelClamp(
4406 float minMipMapLevelClamp,
4407 float maxMipMapLevelClamp);
4408 hipError_t hipTexRefSetMipmappedArray(
4411 unsigned int Flags);
4412 hipError_t hipMipmappedArrayCreate(
4415 unsigned int numMipmapLevels);
4416 hipError_t hipMipmappedArrayDestroy(
4418 hipError_t hipMipmappedArrayGetLevel(
4421 unsigned int level);
4452 hipError_t hipRegisterApiCallback(uint32_t
id,
void* fun,
void* arg);
4453 hipError_t hipRemoveApiCallback(uint32_t
id);
4454 hipError_t hipRegisterActivityCallback(uint32_t
id,
void* fun,
void* arg);
4455 hipError_t hipRemoveActivityCallback(uint32_t
id);
4456 const
char* hipApiName(uint32_t
id);
4457 const
char* hipKernelNameRef(const hipFunction_t f);
4458 const
char* hipKernelNameRefByPtr(const
void* hostFunction, hipStream_t stream);
4459 int hipGetStreamDeviceId(hipStream_t stream);
4503 hipGraphNodeTypeCount
4506 typedef void (*hipHostFn_t)(
void* userData);
4516 void** kernelParams;
4517 unsigned int sharedMemBytes;
4521 unsigned int elementSize;
4545 hipGraphExecUpdateErrorUnsupportedFunctionChange = 0x7
4548 typedef enum hipStreamCaptureMode {
4549 hipStreamCaptureModeGlobal = 0,
4550 hipStreamCaptureModeThreadLocal,
4551 hipStreamCaptureModeRelaxed
4552 } hipStreamCaptureMode;
4608 unsigned long long* pId);
4627 unsigned long long* id_out __dparm(0),
4630 size_t* numDependencies_out __dparm(0));
4659 size_t numDependencies,
4660 unsigned int flags __dparm(0));
4776 size_t* pNumRootNodes);
4795 size_t* pNumDependencies);
4815 size_t* pNumDependentNodes);
4887 hipGraphNode_t* pErrorNode,
char* pLogBuffer,
size_t bufferSize);
4902 unsigned long long flags);
5064 void* dst,
const void* src,
size_t count, hipMemcpyKind kind);
5079 size_t count, hipMemcpyKind kind);
5096 void* dst,
const void* src,
size_t count,
5097 hipMemcpyKind kind);
5117 size_t numDependencies,
void* dst,
const void* symbol,
5118 size_t count,
size_t offset, hipMemcpyKind kind);
5134 size_t count,
size_t offset, hipMemcpyKind kind);
5152 void* dst,
const void* symbol,
size_t count,
5153 size_t offset, hipMemcpyKind kind);
5173 size_t numDependencies,
const void* symbol,
5174 const void* src,
size_t count,
size_t offset,
5175 hipMemcpyKind kind);
5191 const void* src,
size_t count,
size_t offset,
5192 hipMemcpyKind kind);
5210 const void* symbol,
const void* src,
5211 size_t count,
size_t offset, hipMemcpyKind kind);
5486 typedef unsigned int GLuint;
5487 typedef unsigned int GLenum;
5490 hipError_t hipGLGetDevices(
unsigned int* pHipDeviceCount,
int* pHipDevices,
5493 hipError_t hipGraphicsGLRegisterBuffer(hipGraphicsResource** resource, GLuint buffer,
5494 unsigned int flags);
5496 hipError_t hipGraphicsGLRegisterImage(hipGraphicsResource** resource, GLuint image,
5497 GLenum target,
unsigned int flags);
5499 hipError_t hipGraphicsMapResources(
int count, hipGraphicsResource_t* resources,
5500 hipStream_t stream __dparm(0) );
5502 hipError_t hipGraphicsSubResourceGetMappedArray(
hipArray_t* array, hipGraphicsResource_t resource,
5503 unsigned int arrayIndex,
unsigned int mipLevel);
5505 hipError_t hipGraphicsResourceGetMappedPointer(
void** devPtr,
size_t* size,
5506 hipGraphicsResource_t resource);
5508 hipError_t hipGraphicsUnmapResources(
int count, hipGraphicsResource_t* resources,
5509 hipStream_t stream __dparm(0));
5511 hipError_t hipGraphicsUnregisterResource(hipGraphicsResource_t resource);
5522 #if defined(__clang__) && defined(__HIP__)
5523 template <
typename T>
5525 T f,
size_t dynSharedMemPerBlk = 0,
int blockSizeLimit = 0) {
5528 template <
typename T>
5529 static hipError_t __host__
inline hipOccupancyMaxPotentialBlockSizeWithFlags(
int* gridSize,
int* blockSize,
5530 T f,
size_t dynSharedMemPerBlk = 0,
int blockSizeLimit = 0,
unsigned int flags = 0 ) {
5533 #endif // defined(__clang__) && defined(__HIP__)
5534 template <
typename T>
5538 template <
typename T>
5542 template <
typename T>
5543 hipError_t
hipMemcpyToSymbol(
const T& symbol,
const void* src,
size_t sizeBytes,
5544 size_t offset __dparm(0),
5545 hipMemcpyKind kind __dparm(hipMemcpyHostToDevice)) {
5548 template <
typename T>
5550 hipMemcpyKind kind, hipStream_t stream __dparm(0)) {
5553 template <
typename T>
5555 size_t sizeBytes,
size_t offset __dparm(0),
5556 hipMemcpyKind kind __dparm(hipMemcpyDeviceToHost)) {
5559 template <
typename T>
5561 hipMemcpyKind kind, hipStream_t stream __dparm(0)) {
5566 int* numBlocks, T f,
int blockSize,
size_t dynSharedMemPerBlk) {
5568 numBlocks,
reinterpret_cast<const void*
>(f), blockSize, dynSharedMemPerBlk);
5572 int* numBlocks, T f,
int blockSize,
size_t dynSharedMemPerBlk,
unsigned int flags) {
5574 numBlocks,
reinterpret_cast<const void*
>(f), blockSize, dynSharedMemPerBlk, flags);
5576 template <
typename F>
5578 F kernel,
size_t dynSharedMemPerBlk, uint32_t blockSizeLimit) {
5583 void** kernelParams,
unsigned int sharedMemBytes, hipStream_t stream) {
5585 blockDim, kernelParams, sharedMemBytes, stream);
5589 unsigned int numDevices,
unsigned int flags = 0) {
5594 unsigned int numDevices,
unsigned int flags = 0) {
5597 hipError_t hipCreateSurfaceObject(hipSurfaceObject_t* pSurfObject,
const hipResourceDesc* pResDesc);
5598 hipError_t hipDestroySurfaceObject(hipSurfaceObject_t surfaceObject);
5599 template <
class T,
int dim, enum hipTextureReadMode readMode>
5600 DEPRECATED(DEPRECATED_MSG)
5601 static inline hipError_t hipBindTexture(
size_t* offset,
const struct texture<T, dim, readMode>& tex,
5602 const void* devPtr,
size_t size = UINT_MAX) {
5603 return hipBindTexture(offset, &tex, devPtr, &tex.channelDesc, size);
5605 template <
class T,
int dim, enum hipTextureReadMode readMode>
5606 DEPRECATED(DEPRECATED_MSG)
5607 static inline hipError_t
5608 hipBindTexture(
size_t* offset,
const struct texture<T, dim, readMode>& tex,
const void* devPtr,
5610 return hipBindTexture(offset, &tex, devPtr, &desc, size);
5612 template<
class T,
int dim, enum hipTextureReadMode readMode>
5613 DEPRECATED(DEPRECATED_MSG)
5614 static inline hipError_t hipBindTexture2D(
5616 const struct texture<T, dim, readMode> &tex,
5622 return hipBindTexture2D(offset, &tex, devPtr, &tex.channelDesc, width, height, pitch);
5624 template<
class T,
int dim, enum hipTextureReadMode readMode>
5625 DEPRECATED(DEPRECATED_MSG)
5626 static inline hipError_t hipBindTexture2D(
5628 const struct texture<T, dim, readMode> &tex,
5635 return hipBindTexture2D(offset, &tex, devPtr, &desc, width, height, pitch);
5637 template<
class T,
int dim, enum hipTextureReadMode readMode>
5638 DEPRECATED(DEPRECATED_MSG)
5639 static inline hipError_t hipBindTextureToArray(
5640 const struct texture<T, dim, readMode> &tex,
5645 return (err ==
hipSuccess) ? hipBindTextureToArray(&tex, array, &desc) : err;
5647 template<
class T,
int dim, enum hipTextureReadMode readMode>
5648 DEPRECATED(DEPRECATED_MSG)
5649 static inline hipError_t hipBindTextureToArray(
5650 const struct texture<T, dim, readMode> &tex,
5654 return hipBindTextureToArray(&tex, array, &desc);
5656 template<
class T,
int dim, enum hipTextureReadMode readMode>
5658 const struct texture<T, dim, readMode> &tex,
5670 template<
class T,
int dim, enum hipTextureReadMode readMode>
5672 const struct texture<T, dim, readMode> &tex,
5678 template<
class T,
int dim, enum hipTextureReadMode readMode>
5679 DEPRECATED(DEPRECATED_MSG)
5680 static inline hipError_t hipUnbindTexture(
5681 const struct texture<T, dim, readMode> &tex)
5683 return hipUnbindTexture(&tex);
5687 #endif // __cplusplus
5690 #pragma GCC visibility pop
5697 #elif !(defined(__HIP_PLATFORM_HCC__) || defined(__HIP_PLATFORM_AMD__)) && (defined(__HIP_PLATFORM_NVCC__) || defined(__HIP_PLATFORM_NVIDIA__))
5698 #include "hip/nvidia_detail/nvidia_hip_runtime_api.h"
5700 #error("Must define exactly one of __HIP_PLATFORM_AMD__ or __HIP_PLATFORM_NVIDIA__");
5715 #if defined(__cplusplus) && !defined(__HIP_DISABLE_CPP_FUNCTIONS__)
5717 static inline hipError_t
hipMalloc(T** devPtr,
size_t size) {
5724 static inline hipError_t
hipHostMalloc(T** ptr,
size_t size,
5738 #include <hip/amd_detail/hip_prof_str.h>
hipError_t hipGraphExecMemcpyNodeSetParams(hipGraphExec_t hGraphExec, hipGraphNode_t node, hipMemcpy3DParms *pNodeParams)
Sets the parameters for a memcpy node in the given graphExec.
Definition: hip_runtime_api.h:525
Definition: hip_runtime_api.h:815
hipError_t hipPointerGetAttributes(hipPointerAttribute_t *attributes, const void *ptr)
Return attributes for the specified pointer.
@ hipGraphExecUpdateErrorFunctionChanged
The update failed because the function of a kernel node changed.
Definition: hip_runtime_api.h:4539
@ hipDeviceAttributeMaxPitch
Maximum pitch in bytes allowed by memory copies.
Definition: hip_runtime_api.h:391
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.
@ hipDeviceAttributeAsyncEngineCount
Cuda only. Asynchronous engines number.
Definition: hip_runtime_api.h:331
@ hipGraphExecUpdateErrorTopologyChanged
The update failed because the topology changed.
Definition: hip_runtime_api.h:4537
hipError_t hipMemcpy3D(const struct hipMemcpy3DParms *p)
Copies data between host and device.
hipErrorStreamCaptureImplicit
Definition: hip_runtime_api.h:297
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 hipCtxEnablePeerAccess(hipCtx_t peerCtx, unsigned int flags)
Enables direct access to memory allocations in a peer context.
hipGraphicsRegisterFlags
Definition: hip_runtime_api.h:897
int regsPerBlock
Registers per block.
Definition: hip_runtime_api.h:88
hipError_t hipMallocPitch(void **ptr, size_t *pitch, size_t width, size_t height)
hipError_t hipSetDevice(int deviceId)
Set default device to be used for subsequent hip API calls from this thread.
@ hipDeviceAttributeManagedMemory
Device supports allocating managed memory on this system.
Definition: hip_runtime_api.h:356
hipError_t hipGetTextureObjectResourceViewDesc(struct hipResourceViewDesc *pResViewDesc, hipTextureObject_t textureObject)
Gets resource view descriptor for the texture object.
hipError_t hipDeviceGetP2PAttribute(int *value, hipDeviceP2PAttr attr, int srcDevice, int dstDevice)
Returns a value for attr of link between two devices.
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 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.
@ hipDeviceAttributeMemoryBusWidth
Global memory bus width in bits.
Definition: hip_runtime_api.h:392
@ hipDeviceAttributePageableMemoryAccessUsesHostPageTables
Device accesses pageable memory via the host's page tables.
Definition: hip_runtime_api.h:400
hipError_t hipDestroyExternalMemory(hipExternalMemory_t extMem)
Destroys an external memory object.
hipError_t hipGraphRemoveDependencies(hipGraph_t graph, const hipGraphNode_t *from, const hipGraphNode_t *to, size_t numDependencies)
Removes dependency edges from a graph.
@ hipDeviceAttributeSingleToDoublePrecisionPerfRatio
Cuda only. Performance ratio of single precision to double precision.
Definition: hip_runtime_api.h:412
struct hipGraphNode * hipGraphNode_t
Definition: hip_runtime_api.h:4480
const char * hipGetErrorString(hipError_t hipError)
Return handy text string message to explain the error which occurred.
hipError_t hipGraphLaunch(hipGraphExec_t graphExec, hipStream_t stream)
launches an executable graph in a stream
hipError_t hipGetDeviceFlags(unsigned int *flags)
Gets the flags set for current device.
hipError_t hipDeviceGetByPCIBusId(int *device, const char *pciBusId)
Returns a handle to a compute device.
hipErrorInvalidMemcpyDirection
Invalid memory copy direction.
Definition: hip_runtime_api.h:222
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 hipMalloc3DArray(hipArray **array, const struct hipChannelFormatDesc *desc, struct hipExtent extent, unsigned int flags)
Allocate an array on the device.
unsigned hasGlobalInt64Atomics
64-bit integer atomics for global memory.
Definition: hip_runtime_api.h:54
@ hipDeviceAttributeSharedMemPerMultiprocessor
Cuda only. Shared memory available per multiprocessor.
Definition: hip_runtime_api.h:411
int minor
Definition: hip_runtime_api.h:100
@ hipDeviceAttributeMaxBlockDimX
Max block size in width.
Definition: hip_runtime_api.h:358
hipErrorInvalidDevicePointer
Invalid Device Pointer.
Definition: hip_runtime_api.h:221
hipError_t hipChooseDevice(int *device, const hipDeviceProp_t *prop)
Device which matches hipDeviceProp_t is returned.
hipError_t hipGraphGetRootNodes(hipGraph_t graph, hipGraphNode_t *pRootNodes, size_t *pNumRootNodes)
Returns graph's root nodes.
hipError_t hipGraphClone(hipGraph_t *pGraphClone, hipGraph_t originalGraph)
Clones a graph.
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 hipStreamBeginCapture(hipStream_t stream, hipStreamCaptureMode mode)
Begins graph capture on a stream.
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 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 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...
hipGLDeviceList
Definition: hip_runtime_api.h:884
hipError_t hipGraphCreate(hipGraph_t *pGraph, unsigned int flags)
Creates a graph.
hipErrorMapBufferObjectFailed
Produced when the IPC memory attach failed from ROCr.
Definition: hip_runtime_api.h:234
size_t texturePitchAlignment
Pitch alignment requirement for texture references bound to pitched memory.
Definition: hip_runtime_api.h:130
@ hipDeviceAttributeMaxGridDimX
Max grid size in width.
Definition: hip_runtime_api.h:361
hipError_t hipGraphExecHostNodeSetParams(hipGraphExec_t hGraphExec, hipGraphNode_t node, const hipHostNodeParams *pNodeParams)
Sets the parameters for a host node in the given graphExec.
unsigned hasThreadFenceSystem
__threadfence_system.
Definition: hip_runtime_api.h:67
hipError_t hipStreamCreate(hipStream_t *stream)
Create an asynchronous stream.
@ hipGraphNodeTypeGraph
Node which executes an embedded graph.
Definition: hip_runtime_api.h:4496
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.
int pageableMemoryAccessUsesHostPageTables
Device accesses pageable memory via the host's page tables.
Definition: hip_runtime_api.h:149
hipError_t hipDeviceGetStreamPriorityRange(int *leastPriority, int *greatestPriority)
Returns numerical values that correspond to the least and greatest stream priority.
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 hipWaitExternalSemaphoresAsync(const hipExternalSemaphore_t *extSemArray, const hipExternalSemaphoreWaitParams *paramsArray, unsigned int numExtSems, hipStream_t stream)
Waits on a set of external semaphore objects.
hipError_t hipGraphHostNodeGetParams(hipGraphNode_t node, hipHostNodeParams *pNodeParams)
Returns a host node's parameters.
@ hipDeviceAttributeMaxThreadsDim
Maximum dimension of a block.
Definition: hip_runtime_api.h:388
Definition: hip_runtime_api.h:520
hipError_t hipStreamGetCaptureInfo(hipStream_t stream, hipStreamCaptureStatus *pCaptureStatus, unsigned long long *pId)
Get capture status of a stream.
@ hipDeviceAttributeImageSupport
'1' if Device supports image, '0' otherwise.
Definition: hip_runtime_api.h:446
@ hipDeviceAttributeLuid
Cuda only. 8-byte locally unique identifier in 8 bytes. Undefined on TCC and non-Windows platforms.
Definition: hip_runtime_api.h:353
@ hipMemAdviseSetPreferredLocation
Definition: hip_runtime_api.h:678
int maxTexture3D[3]
Maximum dimensions (width, height, depth) of 3D images, in image elements.
Definition: hip_runtime_api.h:125
hipError_t hipGetTextureObjectTextureDesc(hipTextureDesc *pTexDesc, hipTextureObject_t textureObject)
Gets texture descriptor for the texture object.
hipError_t hipGraphNodeGetDependencies(hipGraphNode_t node, hipGraphNode_t *pDependencies, size_t *pNumDependencies)
Returns a node's dependencies.
hipError_t hipStreamCreateWithPriority(hipStream_t *stream, unsigned int flags, int priority)
Create an asynchronous stream with the specified priority.
@ hipMemRangeAttributeCoherencyMode
Definition: hip_runtime_api.h:719
hipFuncCache_t
Definition: hip_runtime_api.h:758
hipError_t hipGraphAddDependencies(hipGraph_t graph, const hipGraphNode_t *from, const hipGraphNode_t *to, size_t numDependencies)
Adds dependency edges to a graph.
int cooperativeMultiDeviceUnmatchedBlockDim
Definition: hip_runtime_api.h:138
hipError_t hipPeekAtLastError(void)
Return last error returned by any HIP runtime API call.
struct ihipGraph * hipGraph_t
Definition: hip_runtime_api.h:4476
@ hipDeviceAttributeMaxSurfaceCubemapLayered
Cuda only. Maximum dimension of Cubemap layered surface.
Definition: hip_runtime_api.h:370
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 hipModuleGetTexRef(textureReference **texRef, hipModule_t hmod, const char *name)
returns the handle of the texture reference with the name from the module.
hipError_t hipMemcpy3DAsync(const struct hipMemcpy3DParms *p, hipStream_t stream __dparm(0))
Copies data between host and device asynchronously.
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 hipGraphDestroy(hipGraph_t graph)
Destroys a graph.
hipError_t hipHostGetFlags(unsigned int *flagsPtr, void *hostPtr)
Return flags associated with host pointer.
hipErrorHostMemoryNotRegistered
Produced when trying to unlock a non-page-locked memory.
Definition: hip_runtime_api.h:275
@ hipGraphNodeTypeKernel
GPU kernel node.
Definition: hip_runtime_api.h:4492
hipErrorRuntimeOther
Definition: hip_runtime_api.h:314
@ hipDeviceAttributeClockRate
Peak clock frequency in kilohertz.
Definition: hip_runtime_api.h:335
hipError_t hipMemGetAddressRange(hipDeviceptr_t *pbase, size_t *psize, hipDeviceptr_t dptr)
Get information on memory allocations.
int asicRevision
Revision of the GPU in this device.
Definition: hip_runtime_api.h:143
@ hipDeviceAttributeDirectManagedMemAccessFromHost
Definition: hip_runtime_api.h:344
@ hipGraphNodeTypeMemcpyFromSymbol
MemcpyFromSymbol node.
Definition: hip_runtime_api.h:4501
hipError_t hipDestroyExternalSemaphore(hipExternalSemaphore_t extSem)
Destroys an external semaphore object and releases any references to the underlying resource....
hipError_t hipStreamWaitEvent(hipStream_t stream, hipEvent_t event, unsigned int flags)
Make the specified compute stream wait for an event.
@ hipFuncCachePreferEqual
prefer equal size L1 cache and shared memory
Definition: hip_runtime_api.h:762
hipError_t hipGraphInstantiate(hipGraphExec_t *pGraphExec, hipGraph_t graph, hipGraphNode_t *pErrorNode, char *pLogBuffer, size_t bufferSize)
Creates an executable graph from a graph.
@ hipDeviceAttributeMaxGridDimZ
Max grid size in depth.
Definition: hip_runtime_api.h:363
hipError_t hipGetDevice(int *deviceId)
Return the default device id for the calling host thread.
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 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.
@ hipGraphNodeTypeHost
Host (executable) node.
Definition: hip_runtime_api.h:4495
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 hipCtxSetCurrent(hipCtx_t ctx)
Set the passed context as current/default.
hipError_t hipGraphExecDestroy(hipGraphExec_t graphExec)
Destroys an executable graph.
@ hipMemRangeCoherencyModeCoarseGrain
Definition: hip_runtime_api.h:700
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.
Definition: driver_types.h:379
@ hipMemoryTypeDevice
Definition: hip_runtime_api.h:158
@ hipDeviceAttributeMaxRegistersPerBlock
Definition: hip_runtime_api.h:405
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 hipMemcpyDtoDAsync(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream)
Copy data from Device to Device asynchronously.
hipErrorNoDevice
Call to hipGetDeviceCount returned 0 devices.
Definition: hip_runtime_api.h:227
int isLargeBar
1: if it is a large PCI bar device, else 0
Definition: hip_runtime_api.h:142
@ hipDeviceAttributeLocalL1CacheSupported
caching locals in L1 is supported
Definition: hip_runtime_api.h:352
@ hipDeviceAttributeComputeCapabilityMinor
Minor compute capability version number.
Definition: hip_runtime_api.h:394
hipError_t hipGraphMemcpyNodeSetParams(hipGraphNode_t node, const hipMemcpy3DParms *pNodeParams)
Sets a memcpy node's parameters.
@ hipDeviceAttributeAsicRevision
Revision of the GPU in this device.
Definition: hip_runtime_api.h:443
hipError_t hipGraphEventWaitNodeGetEvent(hipGraphNode_t node, hipEvent_t *event_out)
Returns the event associated with an event wait node.
hipStreamCaptureStatus
Definition: hip_runtime_api.h:4553
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.
@ hipGraphExecUpdateError
Definition: hip_runtime_api.h:4535
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
int concurrentManagedAccess
Device can coherently access managed memory concurrently with the CPU.
Definition: hip_runtime_api.h:146
int l2CacheSize
L2 cache size.
Definition: hip_runtime_api.h:104
hipError_t hipDevicePrimaryCtxRelease(hipDevice_t dev)
Release the primary context on the GPU.
size_t textureAlignment
Alignment requirement for textures.
Definition: hip_runtime_api.h:129
@ hipDeviceAttributeMaxTextureCubemap
Cuda only. Maximum dimensions of Cubemap texture.
Definition: hip_runtime_api.h:386
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 hipHostMalloc(void **ptr, size_t size, unsigned int flags)
Allocate device accessible page locked host memory.
@ hipDeviceAttributeKernelExecTimeout
Run time limit for kernels executed on the device.
Definition: hip_runtime_api.h:350
uint32_t y
y
Definition: hip_runtime_api.h:780
@ hipDeviceAttributeL2CacheSize
Size of L2 cache in bytes. 0 if the device doesn't have L2 cache.
Definition: hip_runtime_api.h:351
@ hipDeviceAttributeCooperativeMultiDeviceUnmatchedGridDim
Definition: hip_runtime_api.h:436
hipError_t hipDeviceGetName(char *name, int len, hipDevice_t device)
Returns an identifer string for the device.
@ hipDeviceAttributeMaxTexture3DWidth
Maximum dimension width of 3D texture.
Definition: hip_runtime_api.h:382
unsigned hasSurfaceFuncs
Surface functions.
Definition: hip_runtime_api.h:71
@ hipDeviceAttributeIntegrated
Device is integrated GPU.
Definition: hip_runtime_api.h:348
int isMultiGpuBoard
1 if device is on a multi-GPU board, 0 if not.
Definition: hip_runtime_api.h:115
hipError_t hipMemcpyParam2DAsync(const hip_Memcpy2D *pCopy, hipStream_t stream __dparm(0))
Copies memory for 2D arrays.
@ hipDeviceAttributeMaxGridDimY
Max grid size in height.
Definition: hip_runtime_api.h:362
@ hipDeviceAttributeUnifiedAddressing
Cuda only. An unified address space shared with the host.
Definition: hip_runtime_api.h:420
@ hipMemAdviseUnsetCoarseGrain
Restores cache coherency policy back to fine-grain.
Definition: hip_runtime_api.h:690
@ hipMemoryTypeHost
Memory is physically located on host.
Definition: hip_runtime_api.h:157
@ hipDeviceAttributeSurfaceAlignment
Cuda only. Alignment requirement for surfaces.
Definition: hip_runtime_api.h:414
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 hipCtxSetCacheConfig(hipFuncCache_t cacheConfig)
Set L1/Shared cache partition.
hipError_t hipExternalMemoryGetMappedBuffer(void **devPtr, hipExternalMemory_t extMem, const hipExternalMemoryBufferDesc *bufferDesc)
Maps a buffer onto an imported memory object.
hipErrorInvalidContext
Produced when input context is invalid.
Definition: hip_runtime_api.h:230
unsigned hasSharedInt64Atomics
64-bit integer atomics for shared memory.
Definition: hip_runtime_api.h:55
int computeMode
Compute mode.
Definition: hip_runtime_api.h:106
@ hipGraphicsRegisterFlagsReadOnly
HIP will not write to this registered resource.
Definition: hip_runtime_api.h:899
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.
Definition: hip_runtime_api.h:827
hipError_t hipCtxPushCurrent(hipCtx_t ctx)
Push the context to be set as current/ default context.
@ hipDeviceAttributeIsMultiGpuBoard
Multiple GPU devices.
Definition: hip_runtime_api.h:349
hipSharedMemConfig
Definition: hip_runtime_api.h:767
hipError_t hipGraphGetNodes(hipGraph_t graph, hipGraphNode_t *nodes, size_t *numNodes)
Returns graph nodes.
@ hipDeviceAttributeMaxSurfaceCubemap
Cuda only. Maximum dimensions of Cubemap surface.
Definition: hip_runtime_api.h:369
hipError_t hipDrvMemcpy3D(const HIP_MEMCPY3D *pCopy)
Copies data between host and device.
@ hipDeviceAttributeMaxTexture2DMipmap
Cuda only. Maximum dimensions of 2D mipmapped texture.
Definition: hip_runtime_api.h:381
uint32_t x
x
Definition: hip_runtime_api.h:779
@ hipDeviceAttributeGcnArchName
Device gcnArch name in 256 bytes.
Definition: hip_runtime_api.h:431
int clockRate
Max clock frequency of the multiProcessors in khz.
Definition: hip_runtime_api.h:93
unsigned int * hdpRegFlushCntl
Addres of HDP_REG_COHERENCY_FLUSH_CNTL register.
Definition: hip_runtime_api.h:127
hipErrorPeerAccessNotEnabled
Peer access was never enabled from the current device.
Definition: hip_runtime_api.h:268
hipError_t hipDeviceComputeCapability(int *major, int *minor, hipDevice_t device)
Returns the compute capability of the device.
void(* hipStreamCallback_t)(hipStream_t stream, hipError_t status, void *userData)
Definition: hip_runtime_api.h:1703
int maxTexture1DLinear
Maximum size for 1D textures bound to linear memory.
Definition: hip_runtime_api.h:122
@ hipDeviceAttributeComputePreemptionSupported
Cuda only. Device supports Compute Preemption.
Definition: hip_runtime_api.h:337
hipError_t hipGraphKernelNodeSetParams(hipGraphNode_t node, const hipKernelNodeParams *pNodeParams)
Sets a kernel node's parameters.
hipError_t hipGraphGetEdges(hipGraph_t graph, hipGraphNode_t *from, hipGraphNode_t *to, size_t *numEdges)
Returns a graph's dependency edges.
unsigned hasDynamicParallelism
Dynamic parallelism.
Definition: hip_runtime_api.h:73
hipMemoryAdvise
Definition: hip_runtime_api.h:674
Definition: driver_types.h:95
hipError_t hipGetMipmappedArrayLevel(hipArray_t *levelArray, hipMipmappedArray_const_t mipmappedArray, unsigned int level)
Gets a mipmap level of a HIP mipmapped array.
hipError_t hipDestroyTextureObject(hipTextureObject_t textureObject)
Destroys a texture object.
Definition: hip_runtime_api.h:855
int canMapHostMemory
Check whether HIP can map host memory.
Definition: hip_runtime_api.h:116
size_t sharedMemPerBlock
Size of shared memory region (in bytes).
Definition: hip_runtime_api.h:87
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
@ hipDeviceAttributeMaxSurface2DLayered
Cuda only. Maximum dimensions of 2D layered surface.
Definition: hip_runtime_api.h:367
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 hipIpcCloseMemHandle(void *devPtr)
Close memory mapped with hipIpcOpenMemHandle.
@ hipDeviceAttributeConcurrentManagedAccess
Device can coherently access managed memory concurrently with the CPU.
Definition: hip_runtime_api.h:339
@ hipGraphExecUpdateErrorParametersChanged
The update failed because the parameters changed in a way that is not supported.
Definition: hip_runtime_api.h:4541
hipError_t hipDevicePrimaryCtxGetState(hipDevice_t dev, unsigned int *flags, int *active)
Get the state of the primary context.
hipError_t hipGetSymbolSize(size_t *size, const void *symbol)
Gets the size of the given symbol on the device.
@ hipDeviceAttributeCooperativeMultiDeviceLaunch
Support cooperative launch on multiple devices.
Definition: hip_runtime_api.h:341
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...
int maxThreadsPerMultiProcessor
Maximum resident threads per multi-processor.
Definition: hip_runtime_api.h:105
hipError_t hipDeviceSetCacheConfig(hipFuncCache_t cacheConfig)
Set L1/Shared cache partition.
@ hipDeviceAttributeMaxSurface2D
Maximum dimension (width, height) of 2D surface.
Definition: hip_runtime_api.h:366
@ hipDeviceAttributePhysicalMultiProcessorCount
Definition: hip_runtime_api.h:447
int major
Definition: hip_runtime_api.h:97
@ hipDeviceAttributeMaxSharedMemoryPerBlock
Maximum shared memory available per block in bytes.
Definition: hip_runtime_api.h:409
hipError_t hipMemcpyAtoH(void *dst, hipArray *srcArray, size_t srcOffset, size_t count)
Copies data between host and device.
hipError_t hipGetDeviceCount(int *count)
Return number of compute-capable devices.
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.
hipSuccess
Successful completion.
Definition: hip_runtime_api.h:204
struct hipExternalSemaphoreWaitParams_st hipExternalSemaphoreWaitParams
hipError_t hipModuleOccupancyMaxActiveBlocksPerMultiprocessorWithFlags(int *numBlocks, hipFunction_t f, int blockSize, size_t dynSharedMemPerBlk, unsigned int flags)
Returns occupancy for a device function.
hipError_t hipSetupArgument(const void *arg, size_t size, size_t offset)
Set a kernel argument.
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 hipHostUnregister(void *hostPtr)
Un-register host pointer.
hipErrorStreamCaptureInvalidated
Definition: hip_runtime_api.h:286
hipError_t hipStreamWriteValue32(hipStream_t stream, void *ptr, uint32_t value, unsigned int flags)
Enqueues a write command to the stream.[BETA].
@ hipDeviceAttributeMaxTexture3DAlt
Cuda only. Maximum dimensions of alternate 3D texture.
Definition: hip_runtime_api.h:385
hipError_t hipStreamGetFlags(hipStream_t stream, unsigned int *flags)
Return flags associated with this stream.
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...
@ hipDeviceAttributeMaxThreadsPerBlock
Maximum number of threads per block.
Definition: hip_runtime_api.h:389
hipError_t hipExtStreamCreateWithCUMask(hipStream_t *stream, uint32_t cuMaskSize, const uint32_t *cuMask)
Create an asynchronous stream with the specified CU mask.
@ hipGraphExecUpdateErrorNotSupported
The update failed because something about the node is not supported.
Definition: hip_runtime_api.h:4543
int gcnArch
DEPRECATED: use gcnArchName instead.
Definition: hip_runtime_api.h:117
hipError_t hipStreamSynchronize(hipStream_t stream)
Wait for all commands in stream to complete.
Definition: hip_runtime_api.h:45
const char * hipGetErrorName(hipError_t hip_error)
Return name of the specified error code in text form.
int kernelExecTimeoutEnabled
Run time limit for kernels executed on the device.
Definition: hip_runtime_api.h:131
hipError_t hipDeviceGet(hipDevice_t *device, int ordinal)
Returns a handle to a compute device.
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 hipMemcpyDtoD(hipDeviceptr_t dst, hipDeviceptr_t src, size_t sizeBytes)
Copy data from Device to Device.
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].
int maxTexture1D
Maximum number of elements in 1D images.
Definition: hip_runtime_api.h:123
Definition: driver_types.h:369
@ hipDeviceAttributeMaxBlockDimZ
Max block size in depth.
Definition: hip_runtime_api.h:360
@ hipStreamCaptureStatusInvalidated
Definition: hip_runtime_api.h:4556
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 hipMallocManaged(void **dev_ptr, size_t size, unsigned int flags __dparm(hipMemAttachGlobal))
Allocates memory that will be automatically managed by HIP.
hipError_t hipIpcGetMemHandle(hipIpcMemHandle_t *handle, void *devPtr)
Gets an interprocess memory handle for an existing device memory allocation.
hipError_t hipMemcpyHtoD(hipDeviceptr_t dst, void *src, size_t sizeBytes)
Copy data from Host to Device.
hipError_t hipDriverGetVersion(int *driverVersion)
Returns the approximate HIP driver version.
hipError_t hipGraphMemcpyNodeGetParams(hipGraphNode_t node, hipMemcpy3DParms *pNodeParams)
Gets a memcpy node's parameters.
unsigned hasDoubles
Double-precision floating point.
Definition: hip_runtime_api.h:58
hipErrorInvalidKernelFile
In CUDA DRV, it is CUDA_ERROR_INVALID_PTX.
Definition: hip_runtime_api.h:247
int maxThreadsPerBlock
Max work items per work group or workgroup max size.
Definition: hip_runtime_api.h:90
hipError_t hipCtxGetFlags(unsigned int *flags)
Return flags used for creating default context.
@ hipDeviceAttributeMaxTexture1DMipmap
Cuda only. Maximum size of 1D mipmapped texture.
Definition: hip_runtime_api.h:375
@ hipDeviceAttributeMaxBlockDimY
Max block size in height.
Definition: hip_runtime_api.h:359
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 hipMemAllocPitch(hipDeviceptr_t *dptr, size_t *pitch, size_t widthInBytes, size_t height, unsigned int elementSizeBytes)
Definition: hip_runtime_api.h:84
@ hipDeviceAttributeCooperativeMultiDeviceUnmatchedBlockDim
Definition: hip_runtime_api.h:438
hipErrorStreamCaptureUnjoined
Definition: hip_runtime_api.h:291
@ hipDeviceAttributeMaxTexture2DHeight
Maximum dimension hight of 2D texture.
Definition: hip_runtime_api.h:377
@ hipDeviceAttributeMaxSurface1D
Maximum size of 1D surface.
Definition: hip_runtime_api.h:364
unsigned hasSharedInt32Atomics
32-bit integer atomics for shared memory.
Definition: hip_runtime_api.h:49
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.
hipErrorInvalidValue
Definition: hip_runtime_api.h:205
size_t memPitch
Maximum pitch in bytes allowed by memory copies.
Definition: hip_runtime_api.h:128
@ hipDeviceAttributeDeviceOverlap
Definition: hip_runtime_api.h:342
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...
@ hipGraphNodeTypeMemcpy
Memcpy 3D node.
Definition: hip_runtime_api.h:4493
int pciBusID
PCI Bus ID.
Definition: hip_runtime_api.h:112
hipError_t hipStreamEndCapture(hipStream_t stream, hipGraph_t *pGraph)
Ends capture on a stream, returning the captured graph.
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 hipRuntimeGetVersion(int *runtimeVersion)
Returns the approximate HIP Runtime version.
@ hipDeviceAttributeComputeCapabilityMajor
Major compute capability version number.
Definition: hip_runtime_api.h:355
hipError_t hipConfigureCall(dim3 gridDim, dim3 blockDim, size_t sharedMem __dparm(0), hipStream_t stream __dparm(0))
Configure a kernel launch.
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 hipEventQuery(hipEvent_t event)
Query event status.
hipError_t hipGetChannelDesc(hipChannelFormatDesc *desc, hipArray_const_t array)
Gets the channel descriptor in an array.
@ hipDeviceAttributeMaxTexture3DDepth
Maximum dimension depth of 3D texture.
Definition: hip_runtime_api.h:384
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 hipGraphEventWaitNodeSetEvent(hipGraphNode_t node, hipEvent_t event)
Sets an event wait node's event.
hipErrorRuntimeMemory
Definition: hip_runtime_api.h:312
@ hipDeviceAttributeMaxThreadsPerMultiProcessor
Maximum resident threads per multiprocessor.
Definition: hip_runtime_api.h:390
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 hipStreamGetPriority(hipStream_t stream, int *priority)
Query the priority of a stream.
hipError_t hipOccupancyMaxActiveBlocksPerMultiprocessor(int *numBlocks, const void *f, int blockSize, size_t dynSharedMemPerBlk)
Returns occupancy for a device function.
hipDeviceArch_t arch
Architectural feature flags. New for HIP.
Definition: hip_runtime_api.h:109
hipError_t hipCtxGetApiVersion(hipCtx_t ctx, int *apiVersion)
Returns the approximate HIP api version.
@ hipSharedMemBankSizeFourByte
Definition: hip_runtime_api.h:769
hipError_t hipEventSynchronize(hipEvent_t event)
Wait for an event to complete.
@ hipFuncCachePreferNone
no preference for shared memory or L1 (default)
Definition: hip_runtime_api.h:759
@ hipDeviceAttributeMaxTexture1DLayered
Cuda only. Maximum dimensions of 1D layered texture.
Definition: hip_runtime_api.h:372
hipError_t hipCtxCreate(hipCtx_t *ctx, unsigned int flags, hipDevice_t device)
Create a context and set it as current/ default context.
@ hipDeviceAttributeHdpRegFlushCntl
Address of the HDP_REG_COHERENCY_FLUSH_CNTL register.
Definition: hip_runtime_api.h:433
hipErrorStreamCaptureUnmatched
The capture was not initiated in this stream.
Definition: hip_runtime_api.h:290
@ hipMemRangeCoherencyModeFineGrain
Definition: hip_runtime_api.h:698
hipError_t hipHostFree(void *ptr)
Free memory allocated by the hcc hip host memory allocation API This API performs an implicit hipDevi...
@ hipDeviceAttributePciBusId
PCI Bus ID.
Definition: hip_runtime_api.h:401
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...
int tccDriver
1:If device is Tesla device using TCC driver, else 0
Definition: hip_runtime_api.h:133
Definition: driver_types.h:117
@ hipDeviceAttributePersistingL2CacheMaxSize
Cuda11 only. Maximum l2 persisting lines capacity in bytes.
Definition: hip_runtime_api.h:404
@ hipMemRangeAttributeAccessedBy
Definition: hip_runtime_api.h:715
hipError_t hipDeviceGetLimit(size_t *pValue, enum hipLimit_t limit)
Get Resource limits of current device.
void ** args
Arguments.
Definition: hip_runtime_api.h:790
hipError_t hipMalloc(void **ptr, size_t size)
Allocate memory on the default accelerator.
@ hipGraphExecUpdateErrorNodeTypeChanged
The update failed because a node type changed.
Definition: hip_runtime_api.h:4538
@ hipMemoryTypeUnified
Not used currently.
Definition: hip_runtime_api.h:162
Definition: hip_runtime_api.h:4519
@ hipDeviceAttributeMaxTexture1DLinear
Definition: hip_runtime_api.h:373
Definition: hip_runtime_api.h:517
hipError_t hipEventElapsedTime(float *ms, hipEvent_t start, hipEvent_t stop)
Return the elapsed time between two events.
hipError_t hipGraphInstantiateWithFlags(hipGraphExec_t *pGraphExec, hipGraph_t graph, unsigned long long flags)
Creates an executable graph from a graph.
hipError_t hipGetLastError(void)
Return last error returned by any HIP runtime API call and resets the stored error code to hipSuccess...
hipError_t hipInit(unsigned int flags)
Explicitly initializes the HIP runtime.
@ hipGraphNodeTypeEmpty
Empty (no-op) node.
Definition: hip_runtime_api.h:4497
@ hipDeviceAttributeTexturePitchAlignment
Pitch alignment requirement for 2D texture references bound to pitched memory;.
Definition: hip_runtime_api.h:417
@ hipDeviceAttributeWarpSize
Warp size in threads.
Definition: hip_runtime_api.h:422
unsigned hasGlobalInt32Atomics
32-bit integer atomics for global memory.
Definition: hip_runtime_api.h:47
Definition: driver_types.h:165
hipError_t hipFuncSetCacheConfig(const void *func, hipFuncCache_t config)
Set Cache configuration for a specific function.
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 hipCtxPopCurrent(hipCtx_t *ctx)
Pop the current/default context and return the popped context.
Definition: driver_types.h:82
unsigned hasSyncThreadsExt
__syncthreads_count, syncthreads_and, syncthreads_or.
Definition: hip_runtime_api.h:68
hipError_t hipHostAlloc(void **ptr, size_t size, unsigned int flags)
Allocate device accessible page locked host memory [Deprecated].
@ hipMemoryTypeArray
Definition: hip_runtime_api.h:160
hipErrorInvalidDevice
DeviceID must be in range 0...#compute-devices.
Definition: hip_runtime_api.h:228
unsigned hasFunnelShift
Funnel two words into one with shift&mask caps.
Definition: hip_runtime_api.h:64
@ hipDeviceAttributeMaxTexture3DHeight
Maximum dimension height of 3D texture.
Definition: hip_runtime_api.h:383
@ hipDeviceAttributeLuidDeviceNodeMask
Cuda only. Luid device node mask. Undefined on TCC and non-Windows platforms.
Definition: hip_runtime_api.h:354
struct hipGraphExec * hipGraphExec_t
Definition: hip_runtime_api.h:4484
@ hipDeviceAttributeMemoryClockRate
Peak memory clock frequency in kilohertz.
Definition: hip_runtime_api.h:393
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.
hipErrorNotReady
Definition: hip_runtime_api.h:259
@ hipDeviceAttributeMaxTexture2DGather
Cuda only. Maximum dimensions of 2D texture if gather operations performed.
Definition: hip_runtime_api.h:378
hipError_t hipHostGetDevicePointer(void **devPtr, void *hstPtr, unsigned int flags)
Get Device pointer from Host Pointer allocated through hipHostMalloc.
Definition: hip_runtime_api.h:839
hipError_t hipMemGetInfo(size_t *free, size_t *total)
Query memory info. Return snapshot of free memory, and total allocatable memory on the device.
hipError_t hipEventDestroy(hipEvent_t event)
Destroy the specified event.
@ hipDeviceAttributeTccDriver
Cuda only. Whether device is a Tesla device using TCC driver.
Definition: hip_runtime_api.h:415
@ hipMemAdviseSetCoarseGrain
Definition: hip_runtime_api.h:685
struct hipPointerAttribute_t hipPointerAttribute_t
hipError_t hipDeviceSetSharedMemConfig(hipSharedMemConfig config)
The bank width of shared memory on current device is set.
@ hipDeviceAttributeUuid
Cuda only. Unique ID in 16 byte.
Definition: hip_runtime_api.h:421
hipError_t hipDeviceReset(void)
The state of current device is discarded and updated to a fresh state.
int maxGridSize[3]
Max grid dimensions (XYZ).
Definition: hip_runtime_api.h:92
@ hipDeviceAttributeComputeMode
Compute mode that device is currently in.
Definition: hip_runtime_api.h:336
@ hipDeviceAttributeName
Device name.
Definition: hip_runtime_api.h:397
hipError_t hipSetDeviceFlags(unsigned flags)
The current device behavior is changed according the flags passed.
hipError_t hipCtxGetCurrent(hipCtx_t *ctx)
Get the handle of the current/ default context.
hipErrorStreamCaptureIsolation
Definition: hip_runtime_api.h:293
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.
@ hipDeviceAttributePciDeviceId
PCI Device ID.
Definition: hip_runtime_api.h:402
hipError_t hipFuncGetAttributes(struct hipFuncAttributes *attr, const void *func)
Find out attributes for a given function.
@ hipDeviceAttributeAccessPolicyMaxWindowSize
Cuda only. The maximum size of the window policy in bytes.
Definition: hip_runtime_api.h:330
Definition: hip_runtime_api.h:4511
hipErrorIllegalState
Resource required is not in a valid state to perform operation.
Definition: hip_runtime_api.h:257
Definition: driver_types.h:74
hipError_t hipFuncGetAttribute(int *value, hipFunction_attribute attrib, hipFunction_t hfunc)
Find out a specific attribute for a given function.
@ hipMemAdviseUnsetReadMostly
Undo the effect of hipMemAdviseSetReadMostly.
Definition: hip_runtime_api.h:677
size_t maxSharedMemoryPerMultiProcessor
Maximum Shared Memory Per Multiprocessor.
Definition: hip_runtime_api.h:114
int clockInstructionRate
Definition: hip_runtime_api.h:107
hipError_t hipIpcOpenEventHandle(hipEvent_t *event, hipIpcEventHandle_t handle)
Opens an interprocess event handles.
Definition: hip_runtime_api.h:778
hipError_t hipStreamQuery(hipStream_t stream)
Return hipSuccess if all of the operations in the specified stream have completed,...
hipError_t hipLaunchByPtr(const void *func)
Launch a kernel.
@ hipStreamAddCaptureDependencies
Add new nodes to the dependency set.
Definition: hip_runtime_api.h:4561
hipError_t hipExtStreamGetCUMask(hipStream_t stream, uint32_t cuMaskSize, uint32_t *cuMask)
Get CU mask associated with an asynchronous stream.
hipError_t hipExtMallocWithFlags(void **ptr, size_t sizeBytes, unsigned int flags)
Allocate memory on the default accelerator.
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 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.
unsigned int * hdpMemFlushCntl
Addres of HDP_MEM_COHERENCY_FLUSH_CNTL register.
Definition: hip_runtime_api.h:126
hipError_t hipDevicePrimaryCtxSetFlags(hipDevice_t dev, unsigned int flags)
Set flags for the primary context.
hipErrorStreamCaptureMerge
Definition: hip_runtime_api.h:288
Definition: hip_runtime_api.h:169
@ hipGraphNodeTypeWaitEvent
External event wait node.
Definition: hip_runtime_api.h:4498
@ hipDeviceAttributeTotalConstantMemory
Constant memory size in bytes.
Definition: hip_runtime_api.h:418
hipError_t hipFree(void *ptr)
Free memory allocated by the hcc hip memory allocation API. This API performs an implicit hipDeviceSy...
hipError_t hipGraphMemsetNodeSetParams(hipGraphNode_t node, const hipMemsetParams *pNodeParams)
Sets a memset node's parameters.
unsigned hasWarpShuffle
Warp shuffle operations. (__shfl_*).
Definition: hip_runtime_api.h:63
void * func
Device function symbol.
Definition: hip_runtime_api.h:787
#define hipArrayDefault
Definition: hip_runtime_api.h:648
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.
@ hipDeviceAttributeGlobalL1CacheSupported
Cuda only. Device supports caching globals in L1.
Definition: hip_runtime_api.h:346
@ hipDeviceAttributeTotalGlobalMem
Global memory available on devicice.
Definition: hip_runtime_api.h:419
hipError_t hipDevicePrimaryCtxRetain(hipCtx_t *pctx, hipDevice_t dev)
Retain the primary context on the GPU.
@ hipGLDeviceListCurrentFrame
frame
Definition: hip_runtime_api.h:886
hipError_t hipGraphEventRecordNodeSetEvent(hipGraphNode_t node, hipEvent_t event)
Sets an event record node's event.
hipError_t hipCtxSynchronize(void)
Blocks until the default context has completed all preceding requested tasks.
hipError_t hipMemcpyHtoA(hipArray *dstArray, size_t dstOffset, const void *srcHost, size_t count)
Copies data between host and device.
hipError_t hipFreeHost(void *ptr)
Free memory allocated by the hcc hip host memory allocation API. [Deprecated].
Defines surface types for HIP runtime.
int memoryBusWidth
Global memory bus width in bits.
Definition: hip_runtime_api.h:95
@ hipGraphExecUpdateSuccess
The update succeeded.
Definition: hip_runtime_api.h:4534
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...
unsigned hasWarpVote
Warp vote instructions (__any, __all).
Definition: hip_runtime_api.h:61
char name[256]
Device name.
Definition: hip_runtime_api.h:85
@ hipDeviceAttributeMaxRegistersPerMultiprocessor
32-bit registers available per block.
Definition: hip_runtime_api.h:407
@ hipStreamSetCaptureDependencies
Replace the dependency set with the new nodes.
Definition: hip_runtime_api.h:4562
hipStream_t stream
Stream identifier.
Definition: hip_runtime_api.h:792
hipErrorGraphExecUpdateFailure
Definition: hip_runtime_api.h:306
hipError_t hipMemcpyDtoHAsync(void *dst, hipDeviceptr_t src, size_t sizeBytes, hipStream_t stream)
Copy data from Device to Host asynchronously.
unsigned hasGlobalFloatAtomicExch
32-bit float atomic exch for global memory.
Definition: hip_runtime_api.h:48
hipMemRangeCoherencyMode
Definition: hip_runtime_api.h:697
@ hipGraphNodeTypeEventRecord
External event record node.
Definition: hip_runtime_api.h:4499
hipErrorStreamCaptureUnsupported
Definition: hip_runtime_api.h:284
int concurrentKernels
Device can possibly execute multiple kernels concurrently.
Definition: hip_runtime_api.h:110
@ hipMemRangeAttributeReadMostly
Definition: hip_runtime_api.h:712
unsigned hasWarpBallot
Warp ballot instructions (__ballot).
Definition: hip_runtime_api.h:62
@ hipMemRangeAttributeLastPrefetchLocation
prefetched
Definition: hip_runtime_api.h:717
size_t totalGlobalMem
Size of global memory region (in bytes).
Definition: hip_runtime_api.h:86
@ hipDeviceAttributeTextureAlignment
Alignment requirement for textures.
Definition: hip_runtime_api.h:416
hipError_t hipDrvMemcpy3DAsync(const HIP_MEMCPY3D *pCopy, hipStream_t stream)
Copies data between host and device asynchronously.
hipError_t hipEventRecord(hipEvent_t event, hipStream_t stream)
Record an event in the specified stream.
Definition: hip_runtime_api.h:803
dim3 gridDim
Grid dimentions.
Definition: hip_runtime_api.h:788
hipError_t hipGraphNodeFindInClone(hipGraphNode_t *pNode, hipGraphNode_t originalNode, hipGraph_t clonedGraph)
Finds a cloned version of a node.
int directManagedMemAccessFromHost
Host can directly access managed memory on the device without migration.
Definition: hip_runtime_api.h:145
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.
Definition: driver_types.h:358
Definition: driver_types.h:352
#define hipHostMallocDefault
Definition: hip_runtime_api.h:576
char gcnArchName[256]
AMD GCN Arch Name.
Definition: hip_runtime_api.h:118
@ hipSharedMemBankSizeDefault
The compiler selects a device-specific value for the banking.
Definition: hip_runtime_api.h:768
@ hipDeviceAttributeMaxTextureCubemapLayered
Cuda only. Maximum dimensions of Cubemap layered texture.
Definition: hip_runtime_api.h:387
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 hipMemset3D(hipPitchedPtr pitchedDevPtr, int value, hipExtent extent)
Fills synchronously the memory area pointed to by pitchedDevPtr with the constant value.
hipError_t hipStreamCreateWithFlags(hipStream_t *stream, unsigned int flags)
Create an asynchronous stream.
@ hipDeviceAttributeMaxSurface3D
Maximum dimension (width, height, depth) of 3D surface.
Definition: hip_runtime_api.h:368
hipError_t hipDeviceGetAttribute(int *pi, hipDeviceAttribute_t attr, int deviceId)
Query for a specific device attribute.
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.
@ hipDeviceAttributeCanMapHostMemory
Whether host memory can be mapped into device address space.
Definition: hip_runtime_api.h:332
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.
int maxThreadsDim[3]
Max number of threads in each dimension (XYZ) of a block.
Definition: hip_runtime_api.h:91
@ hipGraphNodeTypeMemcpy1D
Memcpy 1D node.
Definition: hip_runtime_api.h:4500
hipError_t hipMemcpyPeerAsync(void *dst, int dstDeviceId, const void *src, int srcDevice, size_t sizeBytes, hipStream_t stream __dparm(0))
Copies memory from one device to memory on another device.
hipError_t hipDrvMemcpy2DUnaligned(const hip_Memcpy2D *pCopy)
@ hipDeviceAttributePciDomainID
PCI Domain ID.
Definition: hip_runtime_api.h:403
hipError_t hipMemcpyHtoDAsync(hipDeviceptr_t dst, void *src, size_t sizeBytes, hipStream_t stream)
Copy data from Host to Device asynchronously.
int cooperativeMultiDeviceLaunch
HIP device supports cooperative launch on multiple devices.
Definition: hip_runtime_api.h:121
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 hipMemcpyDtoH(void *dst, hipDeviceptr_t src, size_t sizeBytes)
Copy data from Device to Host.
unsigned has3dGrid
Grid and group dims are 3D (rather than 2D).
Definition: hip_runtime_api.h:72
hipError_t hipDeviceGetCacheConfig(hipFuncCache_t *cacheConfig)
Set Cache configuration for a specific function.
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 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 hipMemcpyPeer(void *dst, int dstDeviceId, const void *src, int srcDeviceId, size_t sizeBytes)
Copies memory from one device to memory on another device.
@ hipDeviceAttributeMaxTexture1DWidth
Maximum size of 1D texture.
Definition: hip_runtime_api.h:371
@ hipDeviceAttributeCooperativeLaunch
Support cooperative launch.
Definition: hip_runtime_api.h:340
hipError_t hipImportExternalSemaphore(hipExternalSemaphore_t *extSem_out, const hipExternalSemaphoreHandleDesc *semHandleDesc)
Imports an external semaphore.
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 hipFreeMipmappedArray(hipMipmappedArray_t mipmappedArray)
Frees a mipmapped array on the device.
hipError_t hipGraphNodeGetDependentNodes(hipGraphNode_t node, hipGraphNode_t *pDependentNodes, size_t *pNumDependentNodes)
Returns a node's dependent nodes.
hipGraphNodeType
Definition: hip_runtime_api.h:4491
@ hipDeviceAttributeMultiprocessorCount
Number of multiprocessors on the device.
Definition: hip_runtime_api.h:396
hipError_t hipGraphEventRecordNodeGetEvent(hipGraphNode_t node, hipEvent_t *event_out)
Returns the event associated with an event record node.
@ hipGLDeviceListAll
All hip devices used by current OpenGL context.
Definition: hip_runtime_api.h:885
hipError_t hipIpcGetEventHandle(hipIpcEventHandle_t *handle, hipEvent_t event)
Gets an opaque interprocess handle for an event.
int pciDeviceID
PCI Device ID.
Definition: hip_runtime_api.h:113
hipError_t hipGetDeviceProperties(hipDeviceProp_t *prop, int deviceId)
Returns device properties.
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 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.
@ hipDeviceAttributeArch
Device architecture.
Definition: hip_runtime_api.h:428
hipError_t hipMemcpy(void *dst, const void *src, size_t sizeBytes, hipMemcpyKind kind)
Copy data from src to dst.
@ hipDeviceAttributeIsLargeBar
Whether it is LargeBar.
Definition: hip_runtime_api.h:442
int memoryClockRate
Max global memory clock frequency in khz.
Definition: hip_runtime_api.h:94
hipError_t hipEventCreateWithFlags(hipEvent_t *event, unsigned flags)
Create an event with the specified flags.
@ hipMemAdviseUnsetAccessedBy
Definition: hip_runtime_api.h:683
hipErrorCooperativeLaunchTooLarge
Definition: hip_runtime_api.h:279
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 hipBindTextureToMipmappedArray(const textureReference *tex, hipMipmappedArray_const_t mipmappedArray, const hipChannelFormatDesc *desc)
Binds a mipmapped array to a texture.
int pageableMemoryAccess
Definition: hip_runtime_api.h:147
hipError_t hipGraphNodeGetType(hipGraphNode_t node, hipGraphNodeType *pType)
Returns a node's type.
int warpSize
Warp size.
Definition: hip_runtime_api.h:89
hipError_t hipDeviceTotalMem(size_t *bytes, hipDevice_t device)
Returns the total amount of memory on the device.
@ hipDeviceAttributeReservedSharedMemPerBlock
Cuda11 only. Shared memory reserved by CUDA driver per block.
Definition: hip_runtime_api.h:408
hipError_t hipFreeArray(hipArray *array)
Frees an array on the device.
hipErrorAssert
Produced when the kernel calls assert.
Definition: hip_runtime_api.h:272
hipError_t hipGraphHostNodeSetParams(hipGraphNode_t node, const hipHostNodeParams *pNodeParams)
Sets a host node's parameters.
Definition: texture_types.h:88
@ hipDeviceAttributeStreamPrioritiesSupported
Cuda only. Whether to support stream priorities.
Definition: hip_runtime_api.h:413
hipErrorCapturedEvent
Definition: hip_runtime_api.h:300
int cooperativeMultiDeviceUnmatchedFunc
Definition: hip_runtime_api.h:134
hipError_t hipCtxGetSharedMemConfig(hipSharedMemConfig *pConfig)
Get Shared memory bank configuration.
hipError_t hipGetSymbolAddress(void **devPtr, const void *symbol)
Gets device pointer associated with symbol on the device.
int cooperativeMultiDeviceUnmatchedGridDim
Definition: hip_runtime_api.h:136
hipError_t hipDeviceCanAccessPeer(int *canAccessPeer, int deviceId, int peerDeviceId)
Determine if a device can access a peer's memory.
@ hipMemAdviseUnsetPreferredLocation
Clear the preferred location for the data.
Definition: hip_runtime_api.h:680
hipError_t hipModuleGetFunction(hipFunction_t *function, hipModule_t module, const char *kname)
Function with kname will be extracted if present in module.
hipErrorStreamCaptureWrongThread
Definition: hip_runtime_api.h:302
@ hipDeviceAttributeMultiGpuBoardGroupID
Cuda only. Unique ID of device group on the same multi-GPU board.
Definition: hip_runtime_api.h:395
unsigned hasFloatAtomicAdd
32-bit float atomic add in global and shared memory.
Definition: hip_runtime_api.h:51
hipError_t hipStreamIsCapturing(hipStream_t stream, hipStreamCaptureStatus *pCaptureStatus)
Get stream's capture state.
@ hipDeviceAttributeClockInstructionRate
Frequency in khz of the timer used by the device-side "clock*".
Definition: hip_runtime_api.h:427
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.
@ hipMemRangeCoherencyModeIndeterminate
Definition: hip_runtime_api.h:702
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 hipGraphExecEventRecordNodeSetEvent(hipGraphExec_t hGraphExec, hipGraphNode_t hNode, hipEvent_t event)
Sets the event for an event record node in the given graphExec.
hipError_t hipCtxDisablePeerAccess(hipCtx_t peerCtx)
Disable direct access from current context's virtual address space to memory allocations physically l...
int cooperativeLaunch
HIP device supports cooperative launch.
Definition: hip_runtime_api.h:120
unsigned hasSharedFloatAtomicExch
32-bit float atomic exch for shared memory.
Definition: hip_runtime_api.h:50
@ hipDeviceAttributeCooperativeMultiDeviceUnmatchedSharedMem
Definition: hip_runtime_api.h:440
Definition: driver_types.h:282
Definition: texture_types.h:109
Definition: driver_types.h:314
int multiProcessorCount
Number of multi-processors (compute units).
Definition: hip_runtime_api.h:103
hipError_t hipGraphDestroyNode(hipGraphNode_t node)
Remove a node from the graph.
@ hipDeviceAttributeHostNativeAtomicSupported
Cuda only. Link between the device and the host supports native atomic operations.
Definition: hip_runtime_api.h:347
uint32_t z
z
Definition: hip_runtime_api.h:781
hipError_t hipGetTextureReference(const textureReference **texref, const void *symbol)
Gets the texture reference related with the symbol.
hipError_t hipCtxSetSharedMemConfig(hipSharedMemConfig config)
Set Shared memory bank configuration.
@ hipGraphNodeTypeMemset
Memset 1D node.
Definition: hip_runtime_api.h:4494
int integrated
APU vs dGPU.
Definition: hip_runtime_api.h:119
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 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.
Definition: hip_runtime_api.h:4507
int ECCEnabled
Device has ECC support enabled.
Definition: hip_runtime_api.h:132
Definition: driver_types.h:68
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 hipCtxGetDevice(hipDevice_t *device)
Get the handle of the device associated with current/default context.
@ hipDeviceAttributeCanUseStreamWaitValue
Definition: hip_runtime_api.h:444
dim3 blockDim
Block dimentions.
Definition: hip_runtime_api.h:789
size_t totalConstMem
Size of shared memory region (in bytes).
Definition: hip_runtime_api.h:96
int maxTexture2D[2]
Maximum dimensions (width, height) of 2D images, in image elements.
Definition: hip_runtime_api.h:124
hipError_t hipMemAllocHost(void **ptr, size_t size)
Allocate pinned host memory [Deprecated].
@ hipDeviceAttributeCanUseHostPointerForRegisteredMem
Definition: hip_runtime_api.h:333
#define hipMemAttachGlobal
Definition: hip_runtime_api.h:600
hipError_t hipCreateTextureObject(hipTextureObject_t *pTexObject, const hipResourceDesc *pResDesc, const hipTextureDesc *pTexDesc, const struct hipResourceViewDesc *pResViewDesc)
Creates a texture object.
Definition: hip_runtime_api.h:786
@ hipFuncCachePreferShared
prefer larger shared memory and smaller L1 cache
Definition: hip_runtime_api.h:760
hipError_t hipStreamWriteValue64(hipStream_t stream, void *ptr, uint64_t value, unsigned int flags)
Enqueues a write command to the stream.[BETA].
@ hipMemRangeAttributePreferredLocation
The preferred location of the range.
Definition: hip_runtime_api.h:714
hipErrorHostMemoryAlreadyRegistered
Produced when trying to lock a page-locked memory.
Definition: hip_runtime_api.h:273
@ hipDeviceAttributeMaxSurface1DLayered
Cuda only. Maximum dimensions of 1D layered surface.
Definition: hip_runtime_api.h:365
@ hipMemAdviseSetReadMostly
Definition: hip_runtime_api.h:675
hipFuncAttribute
Definition: hip_runtime_api.h:750
@ hipStreamCaptureStatusNone
Stream is not capturing.
Definition: hip_runtime_api.h:4554
hipError_t hipGraphMemsetNodeGetParams(hipGraphNode_t node, hipMemsetParams *pNodeParams)
Gets a memset node's parameters.
@ hipSharedMemBankSizeEightByte
Definition: hip_runtime_api.h:771
hipDeviceAttribute_t
Definition: hip_runtime_api.h:326
hipError_t hipGraphChildGraphNodeGetGraph(hipGraphNode_t node, hipGraph_t *pGraph)
Gets a handle to the embedded graph of a child graph node.
hipError_t hipFuncSetSharedMemConfig(const void *func, hipSharedMemConfig config)
Set shared memory configuation for a specific function.
hipError_t hipExtGetLinkTypeAndHopCount(int device1, int device2, uint32_t *linktype, uint32_t *hopcount)
Returns the link type and hop count between two devices.
@ hipMemAdviseSetAccessedBy
Definition: hip_runtime_api.h:681
@ hipGraphicsRegisterFlagsWriteDiscard
HIP will only write and will not read from this registered resource.
Definition: hip_runtime_api.h:900
Definition: driver_types.h:258
hipErrorMemoryAllocation
Memory allocation error.
Definition: hip_runtime_api.h:209
hipErrorLaunchFailure
An exception occurred on the device while executing a kernel.
Definition: hip_runtime_api.h:277
@ hipGraphicsRegisterFlagsSurfaceLoadStore
HIP will bind this resource to a surface.
Definition: hip_runtime_api.h:902
hipError_t hipDeviceSynchronize(void)
Waits on all active streams on current device.
hipError_t hipCtxGetCacheConfig(hipFuncCache_t *cacheConfig)
Set Cache configuration for a specific function.
size_t sharedMem
Shared memory.
Definition: hip_runtime_api.h:791
@ hipDeviceAttributeMaxBlocksPerMultiProcessor
Cuda only. Max block size per multiprocessor.
Definition: hip_runtime_api.h:357
hipError_t hipCtxDestroy(hipCtx_t ctx)
Destroy a HIP context.
@ hipDeviceAttributeGcnArch
Device gcn architecture.
Definition: hip_runtime_api.h:430
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...
@ hipDeviceAttributeConcurrentKernels
Device can possibly execute multiple kernels concurrently.
Definition: hip_runtime_api.h:338
hipError_t hipGraphKernelNodeGetParams(hipGraphNode_t node, hipKernelNodeParams *pNodeParams)
Gets kernel node's parameters.
@ hipDeviceAttributeMaxTexture2DLayered
Cuda only. Maximum dimensions of 2D layered texture.
Definition: hip_runtime_api.h:379
int cooperativeMultiDeviceUnmatchedSharedMem
Definition: hip_runtime_api.h:140
hipError_t hipProfilerStart()
Start recording of profiling information When using this API, start the profiler with profiling disab...
hipError_t hipDeviceGetSharedMemConfig(hipSharedMemConfig *pConfig)
Returns bank width of shared memory for current device.
hipErrorNotSupported
Produced when the hip API is not supported/implemented.
Definition: hip_runtime_api.h:283
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.
hipErrorLaunchOutOfResources
Out of resources error.
Definition: hip_runtime_api.h:264
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 hipStreamDestroy(hipStream_t stream)
Destroys the specified stream.
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 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.
@ hipStreamCaptureStatusActive
Stream is actively capturing.
Definition: hip_runtime_api.h:4555
hipError_t hipModuleLoad(hipModule_t *module, const char *fname)
Loads code object from file into a hipModule_t.
hipError_t hipProfilerStop()
Stop recording of profiling information. When using this API, start the profiler with profiling disab...
hipError_t hipGetTextureObjectResourceDesc(hipResourceDesc *pResDesc, hipTextureObject_t textureObject)
Gets resource descriptor for the texture object.
hipErrorInvalidResourceHandle
Resource handle (hipEvent_t or hipStream_t) invalid.
Definition: hip_runtime_api.h:256
hipError_t hipEventCreate(hipEvent_t *event)
@ hipDeviceAttributePageableMemoryAccess
Definition: hip_runtime_api.h:398
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 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.
Definition: driver_types.h:328
hipMemoryType
Definition: hip_runtime_api.h:156
@ hipDeviceAttributeSharedMemPerBlockOptin
Cuda only. Maximum shared memory per block usable by special opt in.
Definition: hip_runtime_api.h:410
@ hipGraphicsRegisterFlagsTextureGather
HIP will perform texture gather operations on this registered resource.
Definition: hip_runtime_api.h:903
@ hipGLDeviceListNextFrame
frame.
Definition: hip_runtime_api.h:888
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 hipMallocHost(void **ptr, size_t size)
Allocate pinned host memory [Deprecated].
hipError_t hipDrvPointerGetAttributes(unsigned int numAttributes, hipPointer_attribute *attributes, void **data, hipDeviceptr_t ptr)
Returns information about the specified pointer.[BETA].
@ hipDeviceAttributeMaxTexture2DWidth
Maximum dimension width of 2D texture.
Definition: hip_runtime_api.h:376
int pciDomainID
PCI Domain ID.
Definition: hip_runtime_api.h:111
@ hipDeviceAttributeCooperativeMultiDeviceUnmatchedFunc
Definition: hip_runtime_api.h:434
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 hipModuleLoadData(hipModule_t *module, const void *image)
builds module from code object which resides in host memory. Image is pointer to that location.
@ hipFuncCachePreferL1
prefer larger L1 cache and smaller shared memory
Definition: hip_runtime_api.h:761
#define hipMemAttachSingle
Definition: hip_runtime_api.h:606
@ hipGraphNodeTypeMemcpyToSymbol
MemcpyToSymbol node.
Definition: hip_runtime_api.h:4502
hipError_t hipMemcpyParam2D(const hip_Memcpy2D *pCopy)
Copies memory for 2D arrays.
@ hipDeviceAttributeMaxTexture2DLinear
Cuda only. Maximum dimensions (width, height, pitch) of 2D textures bound to pitched memory.
Definition: hip_runtime_api.h:380
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 hipMemPrefetchAsync(const void *dev_ptr, size_t count, int device, hipStream_t stream __dparm(0))
Prefetches memory to the specified destination device using HIP.
int managedMemory
Device supports allocating managed memory on this system.
Definition: hip_runtime_api.h:144
hipGraphExecUpdateResult
Definition: hip_runtime_api.h:4533
@ hipDeviceAttributeHdpMemFlushCntl
Address of the HDP_MEM_COHERENCY_FLUSH_CNTL register.
Definition: hip_runtime_api.h:432
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 hipDeviceDisablePeerAccess(int peerDeviceId)
Disable direct access from current device's virtual address space to memory allocations physically lo...
hipError_t hipModuleOccupancyMaxActiveBlocksPerMultiprocessor(int *numBlocks, hipFunction_t f, int blockSize, size_t dynSharedMemPerBlk)
Returns occupancy for a device function.
hipError_t hipFuncSetAttribute(const void *func, hipFuncAttribute attr, int value)
Set attribute for a specific function.
@ hipDeviceAttributeEccEnabled
Whether ECC support is enabled.
Definition: hip_runtime_api.h:329
hipError_t __hipPopCallConfiguration(dim3 *gridDim, dim3 *blockDim, size_t *sharedMem, hipStream_t *stream)
Pop configuration of a kernel launch.
struct hipDeviceProp_t hipDeviceProp_t
@ hipDeviceAttributeMaxSharedMemoryPerMultiprocessor
Maximum Shared Memory PerMultiprocessor.
Definition: hip_runtime_api.h:429
hipError_t hipDevicePrimaryCtxReset(hipDevice_t dev)
Resets the primary context on the GPU.
hipError_t hipModuleUnload(hipModule_t module)
Frees the module.
hipMemRangeAttribute
Definition: hip_runtime_api.h:711
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.
hipErrorPeerAccessAlreadyEnabled
Peer access was already enabled from the current device.
Definition: hip_runtime_api.h:266
hipError_t hipPointerGetAttribute(void *data, hipPointer_attribute attribute, hipDeviceptr_t ptr)
Returns information about the specified pointer.[BETA].
hipError_t hipImportExternalMemory(hipExternalMemory_t *extMem_out, const hipExternalMemoryHandleDesc *memHandleDesc)
Imports an external memory object.
hipStreamUpdateCaptureDependenciesFlags
Definition: hip_runtime_api.h:4560