Update flash_attention.yml (#61)

FA repo moved to adopting nvidia-cutlass-dsl version 4.1.0
Nvidia changed B200 host label to linux.dgx.b200.8

Author: jduprat

.github/workflows/flash_attention.yml

@@ -15,7 +15,7 @@ on:
 jobs:
   benchmark-flash-attn:
     name: Flash Attention CuTe DSL Benchmark
-    runs-on: B200
+    runs-on: linux.dgx.b200.8
     container:
       # https://catalog.ngc.nvidia.com/orgs/nvidia/containers/pytorch/
       image: nvcr.io/nvidia/pytorch:25.06-py3
@@ -29,7 +29,7 @@ jobs:
         run: |
           set -x
           echo "Installing nvidia-cutlass-dsl"
-          pip install nvidia-cutlass-dsl==4.1.0.dev0
+          pip install nvidia-cutlass-dsl==4.1.0
       - name: Buid and Run FlashAttention CuTe DSL
         run: |
           set -x
@@ -41,4 +41,4 @@ jobs:
           export PYTHONPATH=$(pwd)
           python benchmarks/benchmark_attn.py >> $GITHUB_STEP_SUMMARY
 
-          popd
+          popd