fix: lower num_workers to 4 (#4535)

For multi-task training in pytorch, each data source will have their own
dataloader. If the number of workers of dataloaders is large, there will
be many (number of tasks * num_workers) worker processes stressing CPU.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

- **Performance Optimization**
  - Adjusted default maximum worker configuration from 8 to 4 CPUs
  - Reduced potential parallel processing resources for the environment
- **Documentation**
- Updated documentation to reflect the change in default value for
`NUM_WORKERS` from 8 to 4
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Signed-off-by: Chun Cai <[email protected]>

deepmd/pt/utils/env.py

@@ -21,7 +21,7 @@
     ncpus = len(os.sched_getaffinity(0))
 except AttributeError:
     ncpus = os.cpu_count()
-NUM_WORKERS = int(os.environ.get("NUM_WORKERS", min(8, ncpus)))
+NUM_WORKERS = int(os.environ.get("NUM_WORKERS", min(4, ncpus)))
 # Make sure DDP uses correct device if applicable
 LOCAL_RANK = os.environ.get("LOCAL_RANK")
 LOCAL_RANK = int(0 if LOCAL_RANK is None else LOCAL_RANK)

doc/env.md

@@ -72,7 +72,7 @@ Default backend.
 
 :::{envvar} NUM_WORKERS
 
-**Default**: 8 or the number of cores (whichever is smaller)
+**Default**: 4 or the number of cores (whichever is smaller)
 
 {{ pytorch_icon }} Number of subprocesses to use for data loading in the PyTorch backend.
 See [PyTorch documentation](https://pytorch.org/docs/stable/data.html) for details.