We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent b489b4e commit 4e6d54fCopy full SHA for 4e6d54f
ggml/src/ggml-cuda/mmq.cuh
@@ -250,13 +250,15 @@ static constexpr __device__ int mmq_get_granularity_device(const int /*mmq_x*/)
250
}
251
#endif // AMD_MFMA_AVAILABLE
252
253
-static int mmq_get_nwarps_host(const int cc) {
254
#if defined(GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)
+static int mmq_get_nwarps_host(const int cc) {
255
return amd_mfma_available(cc) ? 8 : 4;
256
+}
257
#else
258
+static int mmq_get_nwarps_host(const int /*cc*/) {
259
return 8;
-#endif // (GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)
260
261
+#endif // (GGML_USE_HIP) && defined(__HIP_PLATFORM_AMD__)
262
263
static constexpr __device__ int mmq_get_nwarps_device() {
264
0 commit comments