kvcache-ai/ktransformers
1
0
Fork 0
mirror of https://github.com/kvcache-ai/ktransformers.git synced 2026-04-20 14:29:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

iq1s files

Browse Source
This commit is contained in:
moonshadow-25
2025-03-01 22:44:06 +08:00
parent c513ae59c3
commit d24d369332
2 changed files with 4 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
third_party/llamafile/sgemm.h vendored
View File

@@ -15,6 +15,8 @@ struct ggml_compute_params;
/*moonll old
add more params typeb...
*/
bool iqk_mul_mat(long, long, long,int, const void*, long, int, const void*, long,float*, long, int, int);
bool iqk_mul_mat_zen4(long, long, long,int, const void*, long, int, const void*, long,float*, long, int, int);
bool iqk_mul_mat_arm82(long, long, long,int, const void*, long, int, const void*, long,float*, long, int, int);

11
third_party/llamafile/tinyblas_cpu_sgemm.inc vendored
View File

@@ -323,24 +323,17 @@ bool llamafile_sgemm(long m, long n, long k, const void* A, long lda, const void
#if QK_K == 256
#if defined(__x86_64__) || defined(_M_X64)
#if defined(__AVX2__) && (defined(__FMA__) || (defined(_MSC_VER) && (defined(__AVX2__) || defined(__AVX512F__))))
/*
/*
moonll
more Btype accept
}*/
if (Ctype == GGML_TYPE_F32){
if (Ctype == GGML_TYPE_F32){
if (iqk_mul_mat(m, n, k * ggml_blck_size(ggml_type(Atype)), Atype, A,lda,Btype, B,ldb, (float*)C, ldc, ith, nth)) {
return true;
}
}
// }
#endif
#elif defined __aarch64__ && defined __ARM_FEATURE_DOTPROD && !defined _MSC_VER
if (Btype == GGML_TYPE_Q8_K && Ctype == GGML_TYPE_F32) {