From fec7911f8febb22c27ce1acb12441800da24cc06 Mon Sep 17 00:00:00 2001 From: Ruben Ortlam Date: Mon, 1 Sep 2025 20:58:35 +0200 Subject: [PATCH] vulkan: disable large mmv subgroups on older Nvidia GPUs (#15717) --- ggml/src/ggml-vulkan/ggml-vulkan.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/ggml/src/ggml-vulkan/ggml-vulkan.cpp b/ggml/src/ggml-vulkan/ggml-vulkan.cpp index 4057ce8556..3ac459bc4f 100644 --- a/ggml/src/ggml-vulkan/ggml-vulkan.cpp +++ b/ggml/src/ggml-vulkan/ggml-vulkan.cpp @@ -4662,7 +4662,7 @@ static vk_pipeline ggml_vk_get_dequantize_mul_mat_vec(ggml_backend_vk_context * // heuristic to choose workgroup size uint32_t dmmv_wg = DMMV_WG_SIZE_SUBGROUP; - if (ctx->device->vendor_id == VK_VENDOR_ID_NVIDIA || ctx->device->vendor_id == VK_VENDOR_ID_INTEL) { + if ((ctx->device->vendor_id == VK_VENDOR_ID_NVIDIA && ctx->device->architecture != vk_device_architecture::NVIDIA_PRE_TURING) || ctx->device->vendor_id == VK_VENDOR_ID_INTEL) { // Prefer larger workgroups when M is small, to spread the work out more // and keep more SMs busy. // q6_k seems to prefer small workgroup size even for "medium" values of M.