leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 25 Packages Projects Releases Wiki Activity

v4.2 release. (#2587)

Browse Source 
* Fix default cluster callback values to 1 to avoid profiler failure when these values are not set in command line.

* v4.2 release.
This commit is contained in:
Junkai-Wu
2025-08-22 18:11:24 -04:00
committed by GitHub
parent 11cad1f67b
commit a49a78ffef
351 changed files with 28182 additions and 2032 deletions
Download Patch File Download Diff File Expand all files Collapse all files
include/cute/numeric/int.hpp
+6 -2
View File
@@ -29,9 +29,9 @@
*
**************************************************************************************************/
#pragma once
#include "cutlass/cutlass.h"
#if defined(__CUDACC_RTC__)
#include <cuda/std/cstdint>
#include CUDA_STD_HEADER(cstdint)
#else
#include <cstdint>
#endif
@@ -85,6 +85,8 @@ using CUTE_STL_NAMESPACE::uint16_t;
using CUTE_STL_NAMESPACE::uint32_t;
using CUTE_STL_NAMESPACE::uint64_t;
using cutlass::uint128_t;
using cutlass::uint256_t;
template <int N> struct uint_bit;
template <> struct uint_bit< 1> { using type = uint1_t; };
template <> struct uint_bit< 2> { using type = uint2_t; };
@@ -95,6 +97,8 @@ template <> struct uint_bit< 16> { using type = uint16_t; };
template <> struct uint_bit< 32> { using type = uint32_t; };
template <> struct uint_bit< 64> { using type = uint64_t; };
template <> struct uint_bit<128> { using type = cutlass::uint128_t; };
template <> struct uint_bit<256> { using type = cutlass::uint256_t; };
template <int N>
using uint_bit_t = typename uint_bit<N>::type;