leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

CUTLASS 2.7 (#318)

Browse Source 
CUTLASS 2.7

Mainloop fusion for GEMM: summation over A or B
Strided DGRAD (optimized iterators)
Half-precision GELU_taylor activation functions
Use these when accumulation and epilogue compute types are all cutlass::half_t
Tuning and bug fixes to fused GEMM + GEMM example
Support for smaller than 128b aligned Convolutions: see examples
Caching of results to accelerate Convolution unit tests
Can be enabled or disabled by running cmake .. -DCUTLASS_TEST_ENABLE_CACHED_RESULTS=OFF
Corrections and bug fixes reported by the CUTLASS community
Thank you for filing these issues!

authored-by: Haicheng Wu haichengw@nvidia.com, Manish Gupta manigupta@nvidia.com, Dustyn Blasig dblasig@nvidia.com, Andrew Kerr akerr@nvidia.com
This commit is contained in:
Manish Gupta
2021-09-20 11:02:22 -07:00
committed by GitHub
parent 9ac255863f
commit 2e07c4cc2f
62 changed files with 5611 additions and 186 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
include/cutlass/arch/memory.h
View File

@@ -225,6 +225,34 @@ struct global_store;
//
/////////////////////////////////////////////////////////////////////////////////////////////////
template <typename AccessType>
struct global_store<AccessType, 64> {
CUTLASS_DEVICE
global_store(AccessType const &D, void *ptr, bool pred_guard) {
uint4 const *data = reinterpret_cast<uint4 const *>(&D);
asm volatile(
"{\n"
" .reg .pred p;\n"
" setp.ne.b32 p, %5, 0;\n"
" @p st.global.v4.u32 [%0], {%1, %2, %3, %4};\n"
" @p st.global.v4.u32 [%6], {%7, %8, %9, %10};\n"
" @p st.global.v4.u32 [%11], {%12, %13, %14, %15};\n"
" @p st.global.v4.u32 [%16], {%17, %18, %19, %20};\n"
"}\n"
:
: "l"(ptr), "r"(data[0].x), "r"(data[0].y), "r"(data[0].z),
"r"(data[0].w), "r"((int)pred_guard), "l"(((uint8_t *)ptr) + 16),
"r"(data[1].x), "r"(data[1].y), "r"(data[1].z), "r"(data[1].w),
"l"(((uint8_t *)ptr) + 32),
"r"(data[2].x), "r"(data[2].y), "r"(data[2].z), "r"(data[2].w),
"l"(((uint8_t *)ptr) + 48),
"r"(data[3].x), "r"(data[3].y), "r"(data[3].z), "r"(data[2].w));
}
};
template <typename AccessType>
struct global_store<AccessType, 32> {
CUTLASS_DEVICE