From 3465b9d471f1b9f6896d608b4f64d40c4bf97f14 Mon Sep 17 00:00:00 2001 From: biondizzle Date: Sat, 16 May 2026 20:42:49 +0000 Subject: [PATCH] remove torch.cuda.synchronize() from run_nvfp4_grouped_gemm (cudagraph-safe) --- cutedsl/bridge.py | 1 - 1 file changed, 1 deletion(-) diff --git a/cutedsl/bridge.py b/cutedsl/bridge.py index dfbdd521..4d9ba7a0 100644 --- a/cutedsl/bridge.py +++ b/cutedsl/bridge.py @@ -344,6 +344,5 @@ def run_nvfp4_grouped_gemm( stream, global_scale_a=gsa_c, global_scale_b=gsb_c, ) - torch.cuda.synchronize() return out