File size: 1,976 Bytes
dfbcd52 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 | # Profiling Examples
End-to-end GRPO runs that enable one of verl's profilers so you can capture a performance/memory trace without authoring a bespoke launcher. All scripts use the current `verl.trainer.main_ppo` entry point and the current Hydra API.
## Canonical Scripts
| Script | Profiler | Model | Infer | Train | Platform |
|-----------------------------------------|---------------|--------------------|--------|-------|----------|
| `run_qwen3_8b_npu_profile_e2e.sh` | NPU (E2E) | Qwen3-8B | vLLM | FSDP | NPU |
| `run_qwen3_8b_npu_profile_discrete.sh` | NPU (discrete)| Qwen3-8B | vLLM | FSDP | NPU |
| `run_qwen2_5_vl_7b_torch_memory.sh` | torch_memory | Qwen2.5-VL-7B | SGLang | FSDP | NVIDIA |
### NPU profiling
- `*_profile_e2e.sh` — one end-to-end timeline for all ranks.
- `*_profile_discrete.sh` — per-stage (rollout/ref/actor) discrete traces.
Controlled via `global_profiler.tool=npu`, `global_profiler.steps=[...]`, `global_profiler.save_path=...`, plus per-role `actor_rollout_ref.*.profiler.*` overrides. Override any of `PROFILE_STEPS`, `PROFILE_SAVE_PATH`, `PROFILE_LEVEL`, `PROFILE_CONTENTS`, `PROFILE_DISCRETE`, `PROFILE_RANKS_ALL` to adjust behavior.
### Torch memory profiling
- `run_qwen2_5_vl_7b_torch_memory.sh` dumps `torch.cuda._record_memory_history` snapshots to `global_profiler.save_path` (default `./mem_snapshots`). Load the `.pickle` in PyTorch's memory viz UI. Override `TRACE_ALLOC_MAX_ENTRIES`, `STACK_DEPTH`, `PROFILE_SAVE_PATH` as needed.
## Conventions
- `VAR=${VAR:-default}` for `MODEL_PATH`, batch sizes, learning rate, rollout TP, profile options, etc.
- Dynamic batch size and `trainer.balance_batch=True` are enabled by default.
- No deprecated config (`ppo_megatron_trainer.yaml`, `ppo_micro_batch_size`, `data.val_batch_size`, top-level `reward_model.*`, `actor.ulysses_sequence_parallel_size`).
|