Megatron FSDP Skill

For stable background and recommendation level, see:

• @docs/training/megatron-fsdp.md
• @skills/nemo-mbridge-perf-megatron-fsdp/card.yaml

Enablement

Minimal Megatron FSDP override in Bridge:

cfg.dist.use_megatron_fsdp = True
cfg.ddp.use_megatron_fsdp = True
cfg.ddp.data_parallel_sharding_strategy = "optim_grads_params"
cfg.ddp.average_in_collective = False
cfg.checkpoint.ckpt_format = "fsdp_dtensor"

Example recipe fixup:

cfg = llama3_8b_pretrain_config()
cfg.dist.use_megatron_fsdp = True
cfg.ddp.use_megatron_fsdp = True
cfg.ddp.data_parallel_sharding_strategy = "optim_grads_params"
cfg.ddp.average_in_collective = False
cfg.checkpoint.ckpt_format = "fsdp_dtensor"
cfg.checkpoint.save = "/tmp/fsdp_ckpts"
cfg.checkpoint.load = None

Performance harness note:

python scripts/performance/launch.py --use_megatron_fsdp true

Code Anchors

Bridge config definition:

```148:154:src/megatron/bridge/training/config.py usemegatronfsdp: bool = False """Use Megatron's Fully Sharded Data Parallel. Cannot be used together with usetorchfsdp2."""

usetorchfsdp2: bool = False """Use the torch FSDP2 implementation. FSDP2 is not currently working with Pipeline Parallel. It is still not in a stable release stage, and may therefore contain bugs or other potential issues."""

Bridge validation:

```1533:1578:src/megatron/bridge/training/config.py
if self.dist.use_megatron_fsdp and self.dist.use_torch_fsdp2:
    raise ValueError(...)
...
assert not self.dist.use_tp_pp_dp_mapping, "use_tp_pp_dp_mapping is not supported with Megatron FSDP"
...
assert self.checkpoint.ckpt_format == "fsdp_dtensor", (
    "Megatron FSDP only supports fsdp_dtensor checkpoint format"
)

Runtime wrapper selection:

```217:243:src/megatron/bridge/models/common/unimodal.py if usemegatronfsdp: DP = FullyShardedDataParallel elif usetorchfsdp2: DP = TorchFullyShardedDataParallel else: DP = DistributedDataParallel ... DP( config=getmodelconfig(modelchunk), ddpconfig=ddpconfig, module=modelchunk, ... pgcollection=pgcollection, )

Perf harness overrides:

```74:98:scripts/performance/utils/overrides.py
recipe.ddp.use_megatron_fsdp = True
recipe.ddp.data_parallel_sharding_strategy = "optim_grads_params"
recipe.ddp.keep_fp8_transpose_cache = False
recipe.ddp.average_in_collective = False
...
recipe.checkpoint.load = None

Pitfalls

1. Public recipes often expose usemegatronfsdp but still default to ckptformat="torchdist". If save/load is enabled, switch to fsdp_dtensor.
2. usetorchfsdp2 exists, but on the validated branch Bridge still fails before training because ddpwrap passes pg_collection.
3. CPU offloading is only valid when pipelinemodelparallel_size == 1 and activation recomputation is disabled.
4. Upstream warns that FSDP and TP/CP can want different CUDADEVICEMAX_CONNECTIONS settings on Hopper and earlier.
5. Megatron FSDP and FSDP2 are mutually exclusive.

Verification

Use the existing 2-GPU functional smoke test:

CUDA_VISIBLE_DEVICES=0,1 uv run python -m torch.distributed.run --nproc_per_node=2 \
  -m pytest tests/functional_tests/training/test_megatron_fsdp.py::TestMegatronFSDP::test_fsdp_pretrain_basic -v -s

Success criteria:

• Pytest reports 1 passed
• The log shows finite loss at the last iteration
• The run finishes without a checkpoint format assertion