fixed the XLA flag typo

NVIDIA · Sep 4, 2024 · bef0376 · bef0376
1 parent 6accf4a
commit bef0376
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/rosetta/docs/PGLE.md b/rosetta/docs/PGLE.md
@@ -61,7 +61,7 @@ PGLE found latency for async op custom-call-start.1 and (assumed)custom-call-don
 In order to get the best performance with PGLE, here is a list of all recommended XLA flags:
 ```
 export XLA_FLAGS="--xla_gpu_enable_latency_hiding_scheduler=true
---xla_gpu_enable_triton_gemm-false
+--xla_gpu_enable_triton_gemm=false
 --xla_gpu_graph_level=0
 --xla_gpu_all_reduce_combine_threshold_bytes=1073741824
 --xla_gpu_all_gather_combine_threshold_bytes=1073741824

diff --git a/rosetta/rosetta/projects/maxtext/scripts/example_slurm.sub b/rosetta/rosetta/projects/maxtext/scripts/example_slurm.sub
@@ -53,7 +53,7 @@ export NCCL_IB_SL=1
 
 # Set XLA Flags
 export XLA_FLAGS="--xla_gpu_enable_latency_hiding_scheduler=true
-                --xla_gpu_enable_triton_gemm-false
+                --xla_gpu_enable_triton_gemm=false
                 --xla_gpu_graph_level=0
                 --xla_gpu_all_reduce_combine_threshold_bytes=1073741824
                 --xla_gpu_all_gather_combine_threshold_bytes=1073741824