cp: perf: [Perf recipe] Change TP 16->32 for deepseek GB200 sync benchmark (1715) into r0.5.0 (#1716)

Signed-off-by: Guyue Huang <guyueh@nvidia.com>
Signed-off-by: NeMo Bot <nemo-bot@nvidia.com>
Co-authored-by: Guyue Huang <140554423+guyueh1@users.noreply.github.com>

Author: chtruong814

examples/configs/recipes/llm/performance/grpo-deepseek-v3-32n4g.yaml

@@ -11,7 +11,7 @@ policy:
     num_layers_in_last_pipeline_stage: 6
   generation:
     vllm_cfg:
-      tensor_parallel_size: 16
+      tensor_parallel_size: 32
 logger:
   log_dir: logs/grpo-deepseek-v3-32n4g
   wandb: