haraldwolff/llama.cpp
1
0
Fork 0
Code Issues Pull requests Packages Projects Releases Wiki Activity

CUDA: fixed tensor cores not being used on RDNA3 (#4697)

Browse source
This commit is contained in:
Johannes Gäßler 2023-12-30 13:52:01 +01:00 committed by GitHub
parent 24a447e20a
commit 39d8bc71ed
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 24 additions and 23 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

47
ggml-cuda.cu
View file

@ -119,10 +119,29 @@
#define MIN_CC_DP4A 610 // minimum compute capability for __dp4a, an intrinsic for byte-wise dot products #define MIN_CC_DP4A 610 // minimum compute capability for __dp4a, an intrinsic for byte-wise dot products
#define CC_VOLTA 700 #define CC_VOLTA 700
#define CC_OFFSET_AMD 1000000 #define CC_OFFSET_AMD 1000000
#define CC_RDNA1 (CC_OFFSET_AMD + 1010)
#define CC_RDNA2 (CC_OFFSET_AMD + 1030) #define CC_RDNA2 (CC_OFFSET_AMD + 1030)
#define CC_RDNA3 (CC_OFFSET_AMD + 1100)
#define GGML_CUDA_MAX_NODES 8192 #define GGML_CUDA_MAX_NODES 8192
// define this if you want to always fallback to MMQ kernels and not use cuBLAS for matrix multiplication
// on modern hardware, using cuBLAS is recommended as it utilizes F16 tensor cores which are very performant
// for large computational tasks. the drawback is that this requires some extra amount of VRAM:
// - 7B quantum model: +100-200 MB
// - 13B quantum model: +200-400 MB
//
//#define GGML_CUDA_FORCE_MMQ
// TODO: improve this to be correct for more hardware
// for example, currently fails for GeForce GTX 1660 which is TURING arch (> VOLTA) but does not have tensor cores
#if !defined(GGML_CUDA_FORCE_MMQ)
#define CUDA_USE_TENSOR_CORES
#endif
// max batch size to use MMQ kernels when tensor cores are available
#define MMQ_MAX_BATCH_SIZE 32
#if defined(GGML_USE_HIPBLAS) #if defined(GGML_USE_HIPBLAS)
#define __CUDA_ARCH__ 1300 #define __CUDA_ARCH__ 1300
@ -189,23 +208,6 @@ static __device__ __forceinline__ int __dp4a(const int a, const int b, int c) {
} }
#endif // defined(GGML_USE_HIPBLAS) #endif // defined(GGML_USE_HIPBLAS)
// define this if you want to always fallback to MMQ kernels and not use cuBLAS for matrix multiplication
// on modern hardware, using cuBLAS is recommended as it utilizes F16 tensor cores which are very performant
// for large computational tasks. the drawback is that this requires some extra amount of VRAM:
// - 7B quantum model: +100-200 MB
// - 13B quantum model: +200-400 MB
//
//#define GGML_CUDA_FORCE_MMQ
// TODO: improve this to be correct for more hardware
// for example, currently fails for GeForce GTX 1660 which is TURING arch (> VOLTA) but does not have tensor cores
#if !defined(GGML_CUDA_FORCE_MMQ) && (!defined(GGML_USE_HIPBLAS) || defined(RDNA3))
#define CUDA_USE_TENSOR_CORES
#endif
// max batch size to use MMQ kernels when tensor cores are available
#define MMQ_MAX_BATCH_SIZE 32
#if defined(_MSC_VER) #if defined(_MSC_VER)
#pragma warning(disable: 4244 4267) // possible loss of data #pragma warning(disable: 4244 4267) // possible loss of data
#endif #endif
@ -8661,13 +8663,12 @@ static void ggml_cuda_mul_mat(const ggml_tensor * src0, const ggml_tensor * src1
} }
#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) #if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)
const bool fp16_performance_good = true;
#ifdef RDNA3 const bool fp16_performance_good = min_compute_capability >= CC_RDNA1;
const bool use_mul_mat_q = false; bool use_mul_mat_q = ggml_is_quantized(src0->type);
#else #ifdef CUDA_USE_TENSOR_CORES
const bool use_mul_mat_q = true; use_mul_mat_q = use_mul_mat_q && min_compute_capability < CC_RDNA3;
#endif // RDNA3 #endif // CUDA_USE_TENSOR_CORES
#else #else