tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-29 13:43:38 -04:00
Code Issues Packages Projects Releases Wiki Activity

HIP : Add HIP 7.0+ compatibility for hipBLAS compute types (#14634)

Browse Source
This commit is contained in:
Slobodan Josic
2025-07-11 18:55:00 +02:00
committed by GitHub
parent aaa088d87f
commit 756aa1020a
1 changed files with 14 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
ggml/src/ggml-cuda/vendors/hip.h vendored
View File

@@ -10,9 +10,6 @@
#include "rocblas/rocblas.h" #include "rocblas/rocblas.h"
#endif // __HIP_PLATFORM_AMD__ #endif // __HIP_PLATFORM_AMD__
#define CUBLAS_COMPUTE_16F HIPBLAS_R_16F
#define CUBLAS_COMPUTE_32F HIPBLAS_R_32F
#define CUBLAS_COMPUTE_32F_FAST_16F HIPBLAS_R_32F
#define CUBLAS_GEMM_DEFAULT HIPBLAS_GEMM_DEFAULT #define CUBLAS_GEMM_DEFAULT HIPBLAS_GEMM_DEFAULT
#define CUBLAS_GEMM_DEFAULT_TENSOR_OP HIPBLAS_GEMM_DEFAULT #define CUBLAS_GEMM_DEFAULT_TENSOR_OP HIPBLAS_GEMM_DEFAULT
#define CUBLAS_OP_N HIPBLAS_OP_N #define CUBLAS_OP_N HIPBLAS_OP_N
@@ -30,7 +27,6 @@
#define CU_CHECK(fn) {hipError_t err = fn; if(err != hipSuccess) { GGML_ABORT("HipVMM Failure: %s\n", hipGetErrorString(err)); }} #define CU_CHECK(fn) {hipError_t err = fn; if(err != hipSuccess) { GGML_ABORT("HipVMM Failure: %s\n", hipGetErrorString(err)); }}
#define __shfl_sync(mask, var, laneMask, width) __shfl(var, laneMask, width) #define __shfl_sync(mask, var, laneMask, width) __shfl(var, laneMask, width)
#define __shfl_xor_sync(mask, var, laneMask, width) __shfl_xor(var, laneMask, width) #define __shfl_xor_sync(mask, var, laneMask, width) __shfl_xor(var, laneMask, width)
#define cublasComputeType_t hipblasDatatype_t //deprecated, new hipblasComputeType_t not in 5.6
#define cublasCreate hipblasCreate #define cublasCreate hipblasCreate
#define cublasDestroy hipblasDestroy #define cublasDestroy hipblasDestroy
#define cublasGemmEx hipblasGemmEx #define cublasGemmEx hipblasGemmEx
@@ -42,7 +38,6 @@
#define cublasSgemm hipblasSgemm #define cublasSgemm hipblasSgemm
#define cublasStatus_t hipblasStatus_t #define cublasStatus_t hipblasStatus_t
#define cublasOperation_t hipblasOperation_t #define cublasOperation_t hipblasOperation_t
#define cudaDataType_t hipblasDatatype_t //deprecated, new hipblasDatatype not in 5.6
#define cudaDeviceCanAccessPeer hipDeviceCanAccessPeer #define cudaDeviceCanAccessPeer hipDeviceCanAccessPeer
#define cudaDeviceDisablePeerAccess hipDeviceDisablePeerAccess #define cudaDeviceDisablePeerAccess hipDeviceDisablePeerAccess
#define cudaDeviceEnablePeerAccess hipDeviceEnablePeerAccess #define cudaDeviceEnablePeerAccess hipDeviceEnablePeerAccess
@@ -144,6 +139,20 @@
#define CUBLAS_STATUS_INTERNAL_ERROR HIPBLAS_STATUS_INTERNAL_ERROR #define CUBLAS_STATUS_INTERNAL_ERROR HIPBLAS_STATUS_INTERNAL_ERROR
#define CUBLAS_STATUS_NOT_SUPPORTED HIPBLAS_STATUS_NOT_SUPPORTED #define CUBLAS_STATUS_NOT_SUPPORTED HIPBLAS_STATUS_NOT_SUPPORTED
#if defined(__HIP_PLATFORM_AMD__) && HIP_VERSION >= 70000000
#define CUBLAS_COMPUTE_16F HIPBLAS_COMPUTE_16F
#define CUBLAS_COMPUTE_32F HIPBLAS_COMPUTE_32F
#define CUBLAS_COMPUTE_32F_FAST_16F HIPBLAS_COMPUTE_32F_FAST_16F
#define cublasComputeType_t hipblasComputeType_t
#define cudaDataType_t hipDataType
#else
#define CUBLAS_COMPUTE_16F HIPBLAS_R_16F
#define CUBLAS_COMPUTE_32F HIPBLAS_R_32F
#define CUBLAS_COMPUTE_32F_FAST_16F HIPBLAS_R_32F
#define cublasComputeType_t hipblasDatatype_t
#define cudaDataType_t hipblasDatatype_t
#endif
#define __CUDA_ARCH__ 1300 #define __CUDA_ARCH__ 1300
#if defined(__gfx803__) || defined(__gfx900__) || defined(__gfx906__) #if defined(__gfx803__) || defined(__gfx900__) || defined(__gfx906__)