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

Switch matrix parallel to task stealing, type and wrap are refacored, thread execution core has not yet

Browse Source
This commit is contained in:
Youmu Liu
2026-02-24 15:12:23 +08:00
parent d99c1e3aa6
commit d62e59cb93
21 changed files with 1000 additions and 965 deletions
Download Patch File Download Diff File Expand all files Collapse all files
StgSharp.Native/sn-src/load_intrinsic.x86.c
+13 -4
View File
@@ -14,14 +14,21 @@
#endif // _MSC_VER
PRIVATE void load_intrinsic_sse(sn_intrinsic *context, SIMDID id);
PRIVATE void load_intrinsic_avx_dual(sn_intrinsic *context, SIMDID id);
PRIVATE void load_intrinsic_avx(sn_intrinsic *context, SIMDID id);
PRIVATE void load_intrinsic_512_dual(sn_intrinsic *context, SIMDID id);
PRIVATE void load_intrinsic_512(sn_intrinsic *context, SIMDID id);
SN_API SIMDID load_intrinsic_function(void *intrinsic_context)
SN_API SIMDID load_intrinsic_function(void *intrinsic_context, uint64_t id)
{
sn_null_assert(intrinsic_context);
SIMDID simdid = sn_get_simd_level();
SIMDID simdid;
if (id) {
simdid.mask = id;
} else {
simdid = sn_get_simd_level();
}
sn_intrinsic *context = (sn_intrinsic *)intrinsic_context;
@@ -33,6 +40,8 @@ SN_API SIMDID load_intrinsic_function(void *intrinsic_context)
load_intrinsic_avx(context, simdid);
break;
case SIMDID_MAIN_LVL_AVX512:
case SIMDID_MAIN_LVL_AMX:
case SIMDID_MAIN_LVL_AVX10:
load_intrinsic_512(context, simdid);
break;
}
@@ -75,8 +84,8 @@ void load_intrinsic_sse(sn_intrinsic *context, SIMDID id)
void load_intrinsic_avx(sn_intrinsic *context, SIMDID id)
{
char avx_seg = id.make_byte[5];
char is_fma = (avx_seg & 1 << 3) != 0;
char is_f16c = (avx_seg & 1 << 4) != 0;
char is_fma = (avx_seg & SIMDID_AVX_FMA) != 0;
char is_f16c = (avx_seg & SIMDID_AVX_F16C) != 0;
context->city_hash_simplify = city_hash_simplify_sse;
context->factorial_simd = factorial_simd_sse;