leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 29 Packages Projects Releases Wiki Activity

CUTLASS 3.6.0 (#1850)

Browse Source 
* v3.6

* update changelog

* update readme

* fix typo

* fixing typos

* hopper gemm with weight prefetch

---------

Co-authored-by: yuzhai <yuzhai@nvidia.com>
Co-authored-by: Haicheng Wu <haichengw@nvidia.com>
This commit is contained in:
Yujia Zhai
2024-10-09 15:33:27 -04:00
committed by GitHub
co-authored by yuzhai Haicheng Wu
parent 0837a2a00a
commit cc3c29a81a
354 changed files with 105937 additions and 8197 deletions
Download Patch File Download Diff File Expand all files Collapse all files
include/cute/util/print.hpp
+15 -21
View File
@@ -30,9 +30,8 @@
**************************************************************************************************/
#pragma once
#include <cute/config.hpp>
#include <cute/util/type_traits.hpp>
#include <cute/config.hpp> // CUTE_HOST_DEVICE
#include <cute/util/type_traits.hpp> // cute::is_valid
//
// CUDA compatible print and printf
@@ -156,50 +155,45 @@ print(char const* format) {
// pretty printing
//
template <class T>
CUTE_HOST_DEVICE void
pretty_print(T const& v) {
printf(" "); print(v);
}
CUTE_HOST_DEVICE void
pretty_print(bool const& v) {
pretty_print(bool v) {
printf("%*d", 3, int(v));
}
CUTE_HOST_DEVICE void
pretty_print(int32_t const& v) {
pretty_print(int32_t v) {
printf("%*d", 5, v);
}
CUTE_HOST_DEVICE void
pretty_print(uint32_t const& v) {
pretty_print(uint32_t v) {
printf("%*d", 5, v);
}
CUTE_HOST_DEVICE void
pretty_print(int64_t const& v) {
pretty_print(int64_t v) {
printf("%*lld", 5, static_cast<long long>(v));
}
CUTE_HOST_DEVICE void
pretty_print(uint64_t const& v) {
pretty_print(uint64_t v) {
printf("%*llu", 5, static_cast<unsigned long long>(v));
}
CUTE_HOST_DEVICE void
pretty_print(half_t const& v) {
printf("%*.2f", 8, float(v));
}
CUTE_HOST_DEVICE void
pretty_print(float const& v) {
pretty_print(float v) {
printf("%*.2e", 10, v);
}
CUTE_HOST_DEVICE void
pretty_print(double const& v) {
pretty_print(double v) {
printf("%*.3e", 11, v);
}
template <class T>
CUTE_HOST_DEVICE void
pretty_print(T t) {
printf(" "); print(t);
}
} // end namespace cute
include/cute/util/type_traits.hpp
+6 -5
View File
@@ -44,7 +44,7 @@
#include <limits> // numeric_limits
#endif
#include <cute/config.hpp>
#include <cute/config.hpp> // CUTE_STL_NAMESPACE
namespace cute
{
@@ -79,6 +79,7 @@ using CUTE_STL_NAMESPACE::is_const_v;
using CUTE_STL_NAMESPACE::is_volatile;
using CUTE_STL_NAMESPACE::is_volatile_v;
// Defined in cute/numeric/integral_constant.hpp
// using CUTE_STL_NAMESPACE::true_type;
// using CUTE_STL_NAMESPACE::false_type;
@@ -278,14 +279,14 @@ struct conditional_template<false, True, False> {
// is_any_of
//
/// Member `value` is true if and only if T is same as (is_same_v) at least one of the types in Us
template <typename T, typename... Us>
// Member `value` is true if and only if T is same as (is_same_v) at least one of the types in Us
template <class T, class... Us>
struct is_any_of {
constexpr static bool value = (... || CUTE_STL_NAMESPACE::is_same_v<T, Us>);
};
/// Is true if and only if T is same as (is_same_v) at least one of the types in Us
template <typename T, typename... Us>
// Is true if and only if T is same as (is_same_v) at least one of the types in Us
template <class T, class... Us>
inline constexpr bool is_any_of_v = is_any_of<T, Us...>::value;
} // end namespace cute