ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-20 14:59:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

Gemm+layernorm instance, ckProfiler, client example (#568)

Browse Source 
* Add gemm + layernorm instance

* Add ckProfiler

* Add test

* Add client example

* Detect if user forger to set the workrspace

* Use literal in the example

* [What] use builtin function for sqrt
[Why] compiler will not use v_sqrt_f64_e64 if we use ::sqrt()

* check gemm vaildity in IsSupportedArgument

* Add more testcases

* Merge duplicated folder in client example

* Print more infomation

* Use better kernel parameter for MS problem size

* clang format

* Add constexpr for if condition and remove redundant include

* Remove cstdlib and add constexpr
This commit is contained in:
rocking5566
2023-02-10 05:02:55 +08:00
committed by GitHub
parent 76d144fa7c
commit f7d28f3e4b
29 changed files with 1635 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
include/ck/utility/math_v2.hpp
View File

@@ -158,9 +158,9 @@ static inline __device__ bool isnan(half_t x)
return (xx & 0x7FFF) > 0x7C00;
};
static inline __device__ float sqrt(float x) { return ::sqrtf(x); };
static inline __device__ float sqrt(float x) { return __builtin_amdgcn_sqrtf(x); };
static inline __device__ double sqrt(double x) { return ::sqrt(x); };
static inline __device__ double sqrt(double x) { return __builtin_amdgcn_sqrt(x); };
} // namespace math
} // namespace ck