CUTLASS 2.1 (#83)

CUTLASS 2.1 contributes:
- BLAS-style host-side API added to CUTLASS Library
- Planar Complex GEMM kernels targeting Volta and Turing Tensor Cores
- Minor enhancements and bug fixes
This commit is contained in:
Andrew Kerr
2020-04-07 13:51:25 -07:00
committed by GitHub
parent 7c0cd26d13
commit 96dab34ad9
196 changed files with 20653 additions and 1995 deletions

View File

@@ -29,7 +29,11 @@
*/
#pragma once
#if defined(__CUDACC_RTC__)
#include <cuda/std/cstdint>
#else
#include <cstdint>
#endif
#include "cutlass/platform/platform.h"
@@ -48,7 +52,7 @@ struct integer_subbyte {
static bool const kSigned = Signed;
/// External type
using T = typename std::conditional<kSigned, int, unsigned>::type;
using T = typename platform::conditional<kSigned, int, unsigned>::type;
/// Storage type
using Storage = uint8_t;