youmuliu/stgsharp
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

mat kernel turns to be generic

Browse Source
This commit is contained in:
Youmu Liu
2025-09-18 21:55:26 +08:00
parent 6e27c6219d
commit 8007eee846
13 changed files with 109 additions and 188 deletions
Download Patch File Download Diff File Expand all files Collapse all files
StgSharp.Native/sn-src/load_intrinsic.c
+6 -6
View File
@@ -66,10 +66,10 @@ void load_intrinsic_sse(sn_intrinsic *context)
context->city_hash_simplify = city_hash_simplify_sse;
context->factorial_simd = factorial_simd_sse;
context->index_pair = index_pair_sse;
context->kernel_add = (MATARITHMETIC)kernel_add_sse;
context->kernel_add = (MATARITHMETIC)f32_add_sse;
context->kernel_fma = (void(*)(void*, void*, void*))kernel_fma_sse;
context->kernel_scalar_mul = kernel_scalar_mul_sse;
context->kernel_sub = (MATARITHMETIC)kernel_sub_sse;
context->kernel_sub = (MATARITHMETIC)f32_sub_sse;
context->kernel_transpose = (MATTRANSPOSEPROC)kernel_transpose_sse;
context->normalize_3 = normalize;
}
@@ -79,10 +79,10 @@ void load_intrinsic_avx(sn_intrinsic *context)
context->city_hash_simplify = city_hash_simplify_sse;
context->factorial_simd = factorial_simd_sse;
context->index_pair = index_pair_sse;
context->kernel_add = (MATARITHMETIC)kernel_add_avx;
context->kernel_add = (MATARITHMETIC)f32_add_avx;
context->kernel_fma = (void(*)(void*, void*, void*))kernel_fma_sse;
context->kernel_scalar_mul = kernel_scalar_mul_avx;
context->kernel_sub = (MATARITHMETIC)kernel_sub_avx;
context->kernel_sub = (MATARITHMETIC)f32_sub_avx;
context->kernel_transpose = (MATTRANSPOSEPROC)kernel_transpose_avx;
context->normalize_3 = normalize;
}
@@ -92,10 +92,10 @@ void load_intrinsic_512(sn_intrinsic *context)
context->city_hash_simplify = city_hash_simplify_sse;
context->factorial_simd = factorial_simd_sse;
context->index_pair = index_pair_sse;
context->kernel_add = (MATARITHMETIC)kernel_add_512;
context->kernel_add = (MATARITHMETIC)f32_add_512;
context->kernel_fma = (void(*)(void*, void*, void*))kernel_fma_sse;
context->kernel_scalar_mul = kernel_scalar_mul_512;
context->kernel_sub = (MATARITHMETIC)kernel_sub_512;
context->kernel_sub = (MATARITHMETIC)f32_sub_512;
context->kernel_transpose = (MATTRANSPOSEPROC)kernel_transpose_512;
context->normalize_3 = normalize;
}