vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Fix CI

Browse source
This commit is contained in:
Iwan Kawrakow 2024-02-27 10:59:29 +02:00
parent f162fcaf81
commit 801f998bd9
1 changed files with 6 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

12
ggml-quants.c
View file

@ -9698,8 +9698,8 @@ void ggml_vec_dot_iq2_s_q8_K(int n, float * restrict s, size_t bs, const void *
qs += 8; qs += 8;
vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[0] | (signs[1] << 16))); vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[0] | (signs[1] << 16)));
vs.val[1] = vandq_u8(vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2);
vs.val[0] = vandq_u8(vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2);
vs.val[0] = vceqq_u8(vs.val[0], mask2); vs.val[0] = vceqq_u8(vs.val[0], mask2);
vs.val[1] = vceqq_u8(vs.val[1], mask2); vs.val[1] = vceqq_u8(vs.val[1], mask2);
@ -9707,8 +9707,8 @@ void ggml_vec_dot_iq2_s_q8_K(int n, float * restrict s, size_t bs, const void *
q2s.val[1] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[1], m1)), q2s.val[1]); q2s.val[1] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[1], m1)), q2s.val[1]);
vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[2] | (signs[3] << 16))); vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[2] | (signs[3] << 16)));
vs.val[1] = vandq_u8(vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2);
vs.val[0] = vandq_u8(vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2);
vs.val[0] = vceqq_u8(vs.val[0], mask2); vs.val[0] = vceqq_u8(vs.val[0], mask2);
vs.val[1] = vceqq_u8(vs.val[1], mask2); vs.val[1] = vceqq_u8(vs.val[1], mask2);
@ -12332,13 +12332,13 @@ size_t quantize_iq4_xs(const float * src, void * dst, int nrow, int n_per_row, i
} }
void quantize_row_iq4_xs(const float * restrict x, void * restrict vy, int k) { void quantize_row_iq4_xs(const float * restrict x, void * restrict vy, int k) {
assert(k % QK4_XS == 0); assert(k % QK_K == 0);
block_iq4_xs * restrict y = vy; block_iq4_xs * restrict y = vy;
quantize_row_iq4_xs_reference(x, y, k); quantize_row_iq4_xs_reference(x, y, k);
} }
void quantize_row_iq4_xs_reference(const float * restrict x, block_iq4_xs * restrict y, int k) { void quantize_row_iq4_xs_reference(const float * restrict x, block_iq4_xs * restrict y, int k) {
assert(k % QK4_XS == 0); assert(k % QK_K == 0);
quantize_iq4_xs(x, y, 1, k, NULL, NULL); quantize_iq4_xs(x, y, 1, k, NULL, NULL);
} }