opencl: fix crash when warming up MoE on Adreno (#22876)

diff --git a/ggml/src/ggml-opencl/ggml-opencl.cpp b/ggml/src/ggml-opencl/ggml-opencl.cpp
index 61bdc62cd108c2377718cd2091c75d26565c4fe7..248124c2896ff4a680350ef0b37e57c6c4e94dcd 100644 (file)
--- a/ggml/src/ggml-opencl/ggml-opencl.cpp
+++ b/ggml/src/ggml-opencl/ggml-opencl.cpp
@@ -13132,7 +13132,7 @@ static void moe_router_reoerder(ggml_backend_t backend, const ggml_tensor * src,
     CL_CHECK(clSetKernelArg(kernel, 4, sizeof(int), &ne02));
 
     size_t histogram_global_size[] = {(size_t)(((ne21 + 63) / 64) * 64), static_cast<size_t>(ne20), 1};
-    size_t histogram_local_size[] = {64, static_cast<size_t>(ne20), 1};
+    size_t histogram_local_size[] = {64, 1, 1};
     backend_ctx->enqueue_ndrange_kernel(kernel, 3, histogram_global_size, histogram_local_size, src);
 
     // Scan