Public/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-08-25 22:21:03 +02:00
Code Issues Packages Projects Releases Wiki Activity

remove elem row/col fast path, invalid for RDNA4

Browse Source
This commit is contained in:
Ruben Ortlam
2026-08-24 17:23:15 +02:00
parent 218c4e05fd
commit 6a08f25c40
1 changed files with 9 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_cm1.comp
+9 -16
View File
@@ -168,24 +168,17 @@ void main() {
uint elem_row[CM_ELEMS];
uint elem_col[CM_ELEMS];
if (WARP == 32) {
[[unroll]] for (uint e = 0; e < CM_ELEMS; ++e) {
elem_row[e] = gl_SubgroupInvocationID / TN + 2 * e;
elem_col[e] = gl_SubgroupInvocationID % TN;
}
} else {
for (uint i = gl_LocalInvocationID.x; i < TM * TN; i += BLOCK_SIZE) {
cm_layout_probe[i] = int32_t(i);
}
barrier();
for (uint i = gl_LocalInvocationID.x; i < TM * TN; i += BLOCK_SIZE) {
cm_layout_probe[i] = int32_t(i);
}
barrier();
coopmat<int32_t, gl_ScopeSubgroup, TM, TN, gl_MatrixUseAccumulator> probe;
coopMatLoad(probe, cm_layout_probe, 0, TN, gl_CooperativeMatrixLayoutRowMajor);
coopmat<int32_t, gl_ScopeSubgroup, TM, TN, gl_MatrixUseAccumulator> probe;
coopMatLoad(probe, cm_layout_probe, 0, TN, gl_CooperativeMatrixLayoutRowMajor);
[[unroll]] for (uint e = 0; e < CM_ELEMS; ++e) {
elem_row[e] = uint(probe[e]) / TN;
elem_col[e] = uint(probe[e]) % TN;
}
[[unroll]] for (uint e = 0; e < CM_ELEMS; ++e) {
elem_row[e] = uint(probe[e]) / TN;
elem_col[e] = uint(probe[e]) % TN;
}
const uint loadr_a = gl_LocalInvocationID.x % (BK / LOAD_VEC_A);