vulkan: copy iq4_nl LUT into shared memory (#10409)

2024-11-20 01:40:18 -06:00
parent 1bacb9f625
commit 8fd4b7fa29
6 changed files with 29 additions and 4 deletions
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp
@@ -10,6 +10,8 @@ layout (binding = 1) writeonly buffer D {D_TYPE data_b[];};
 void main() {
    const uint i = gl_WorkGroupID.x * 4 + gl_LocalInvocationID.x / 64;

+    init_iq4nl_shmem();
+
    const uint tid = gl_LocalInvocationID.x % 64;
    const uint il  = tid/32;
    const uint ir  = tid%32;