ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

vulkan: fix assertion when qy_needs_dequant (#12068)

Browse source 
Looks like a copy/paste bug from qx_needs_dequant.
This commit is contained in:
Jeff Bolz 2025-02-25 09:30:21 -06:00 • committed by GitHub
parent 401af80b54
commit a82c9e7c23
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 2 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

4
ggml/src/ggml-vulkan/ggml-vulkan.cpp
View file

@ -4192,7 +4192,7 @@ static void ggml_vk_mul_mat_q_f16(ggml_backend_vk_context * ctx, vk_context& sub
} }
if (qy_needs_dequant) { if (qy_needs_dequant) {
d_Y = ctx->prealloc_y; d_Y = ctx->prealloc_y;
GGML_ASSERT(d_Y->size >= y_sz * ne02 * ne03); GGML_ASSERT(d_Y->size >= y_sz * ne12 * ne13);
} else { } else {
d_Y = d_Qy; d_Y = d_Qy;
y_buf_offset = qy_buf_offset; y_buf_offset = qy_buf_offset;
@ -4769,7 +4769,7 @@ static void ggml_vk_mul_mat_id_q_f16(ggml_backend_vk_context * ctx, vk_context&
} }
if (qy_needs_dequant) { if (qy_needs_dequant) {
d_Y = ctx->prealloc_y; d_Y = ctx->prealloc_y;
GGML_ASSERT(d_Y->size >= y_sz * ne02 * ne03); GGML_ASSERT(d_Y->size >= y_sz * ne12 * ne13);
} else { } else {
d_Y = d_Qy; d_Y = d_Qy;
y_buf_offset = qy_buf_offset; y_buf_offset = qy_buf_offset;