ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-04-30 19:31:48 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix iq2_ks

Browse Source
This commit is contained in:
Iwan Kawrakow
2025-05-14 19:03:54 +03:00
parent f069a57817
commit 217905c8b3
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
ggml/src/ggml-cuda/mmq.cuh
View File

@@ -2404,7 +2404,7 @@ template <int mmq_y, int nwarps, bool need_check> static __device__ __forceinlin
const block_iq2_ks * bxi = (const block_iq2_ks *)(x + i*stride + sizeof(half)) + kbx0; const block_iq2_ks * bxi = (const block_iq2_ks *)(x + i*stride + sizeof(half)) + kbx0;
uint16_t extra = bxi->extra >> 8*(kqsx/8); uint16_t extra = bxi->extra >> 4*(kqsx/8);
int q2 = get_int_b2(bxi->qs, kqsx); int q2 = get_int_b2(bxi->qs, kqsx);
aux32[0] = ((q2 >> 0) & 0x03030303) | (((extra << 2) & 4) * 0x01010101); aux32[0] = ((q2 >> 0) & 0x03030303) | (((extra << 2) & 4) * 0x01010101);