This PR is greatly simplified now that it stacked on top of a PR that builds with distributed always. We only need to stub functions that may not be defined due to a backend not being enabled.
Signed-off-by: Edward Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159889
Approved by: https://github.com/wconstab
ghstack dependencies: #160449
This PR is greatly simplified now that it stacked on top of a PR that builds with distributed always. We only need to stub functions that may not be defined due to a backend not being enabled.
Signed-off-by: Edward Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159889
Approved by: https://github.com/wconstab
ghstack dependencies: #160449
This PR is greatly simplified now that it stacked on top of a PR that builds with distributed always. We only need to stub functions that may not be defined due to a backend not being enabled.
Signed-off-by: Edward Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159889
Approved by: https://github.com/wconstab
ghstack dependencies: #160449
# Context
In #161183, we added NUMA-binding support for `Callable` entrypoints to `elastic_launch`.
However, we would raise an exception if the subprocesses would be spawned in parallel via `ThreadPoolExecutor`, which is an option configurable via the `TORCH_MP_PARALLEL_START` environment variable (see diff).
The logic here was that `os.sched_setaffinity`, which we used to set CPU affinities, is [per process](https://docs.python.org/3/library/os.html#os.sched_setaffinity), so there could be a race condition during a parallel start:
> Restrict the process with PID pid (or the current process if zero) to a set of CPUs. mask is an iterable of integers representing the set of CPUs to which the process should be restricted.
But on further reading, the Linux docs say [`sched_setaffinity` is per *thread*.](https://man7.org/linux/man-pages/man2/sched_setaffinity.2.html) As it turns out, the Python doc is a misnomer.
I [verified that `sched_setaffinity` only affects the calling thread, not the entire calling process.](https://gist.github.com/pdesupinski/7e2de3cbe5bb48d489f257b83ccddf07)
The upshot is that we actually *can* safely use the inheritance trick from #161183 even with parallel start, since the setting will be inherited from the calling thread, and `os.sched_setaffinity` only affects the calling thread.
# This PR
Remove restrictions against parallel start for NUMA binding.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/161576
Approved by: https://github.com/d4l3k
# Context
In #160163, we added support for NUMA binding for `Callable` entrypoints to `elastic_launch`. This requires special consideration, because they go through a different path to spawn subprocesses compared to `str` entrypoints, a path which does not provide a straightforward way to utilize `numactl` CLI. See #160006 for a full description of the challenges.
Although #160163 worked in initial local experiments, we ran into some linker errors in other environments when we tried to call `numactl`. This appeared to be due to interactions with how the `LD_PRELOAD` environment variable was being set.
# This PR
On further thought, the most straightforward, foolproof solution here is to use [the trick that @d4l3k suggested.](https://github.com/pytorch/pytorch/issues/160006#issuecomment-3162018836)
Specifically, for each local rank `i`:
1. The parent process sets its own CPU affinity to what local rank `i`'s should be.
2. Then, the parent spawns the subprocess for local rank `i`.
3. Finally, the parent resets its own CPU affinity to what it was originally.
There were other solutions that would work just for `Callable` entrypoints, but I believe this is the simplest one that can work for *both* `str` and `Callable`, and it's pretty simple.
This required a bit of refactoring:
1. Turn all the `_get_.*_numactl_options` into functions which return a set of logical CPUs to bind to, rather than options like `--cpunodebind=0`.
2. Instead of wrapping commands with `numactl`, use `os.sched_setaffinity` to bind to the CPUs from (1.).
3. Put this all inside a context manager which encapsulates applying and restoring the bindings in the parent process.
4. Use the context manager for both `str` and `Callable` paths
# Test Plan
## Automated
`$ pytest test/test_numa_binding.py`
## Manual
See [doc.](https://docs.google.com/document/d/1vxD-OKYBTT27jbBwtW9iz9g0tNM0u-i0tiTJg_ieQA8/edit?tab=t.0) Meta only, but TLDR tried out every combination of `str`, `Callable`, binding disabled, and binding enabled on the same model and saw 2x SM utilization for binding enabled.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/161183
Approved by: https://github.com/d4l3k
# Context
Another fix to enable broad rollout of #149334.
The implementation assumes that the trainer process with local rank `n` only uses device `cuda:n`. However, there are sometimes jobs with more than one GPU per process, in which case our assumption could be incorrect and actually lead to worse memory locality.
# This PR
As titled.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160848
Approved by: https://github.com/kiukchung
# Context
Broader context in #160163.
In order for the _utils_internal version of signpost_event to do proper logging, its parameters argument needs to be json serializable.
# This PR
Convert `NumaOptions` to serializable form before inputting to `signpost_event`.
# Test Plan
## Automated
Added tests `$ pytest test/test_numa_binding.py`.
## Manual
See [D80317206](https://www.internalfb.com/diff/D80317206).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160710
Approved by: https://github.com/kiukchung
# Context
This is an extension of #149334.
# This PR
Add support for NUMA bindings with Callable entrypoints, such as `do_train` instead of `/usr/local/bin/python`.
Most notably, we utilize a hack in order to force `Process.start()` to use custom NUMA bindings for each subprocess. Please search for `HACK:` in the code to see a description of the implementation we chose, and #160006 for discussion of alternatives and why this is necessary.
Other changes:
* Remove unnecessary `--preferred` option from all binding strategies. By default, Linux already allocates memory to the NUMA node local to the CPU which triggered the allocation. (See [MPOL_LOCAL](https://man7.org/linux/man-pages/man2/set_mempolicy.2.html).)
* Refactor so that the main API is `maybe_wrap_command_with_numa_bindings`, which computes bindings for a single rank at a time, rather than `maybe_wrap_with_numa_bindings` which computed bindings for all ranks at once. This allowed for more code sharing between `Callable` and `str` entrypoints.
# Test Plan
## Automated
`$ pytest test/test_numa_binding.py`
## Manual
Using [this benchmark,](https://gist.github.com/pdesupinski/bbe01ade455d86e989794f2c612e2d91), ran
```
$ PYTHONUNBUFFERED=1 LOGLEVEL=INFO perf stat -e ls_dmnd_fills_from_sys.dram_io_far,ls_dmnd_fills_from_sys.dram_io_near -- python -m torch.distributed.run --standalone --nproc-per-node=8 --numa-binding=node --run-path mlp_train.py 2>&1 | tee node_callable.txt && PYTHONUNBUFFERED=1 LOGLEVEL=INFO perf stat -e ls_dmnd_fills_from_sys.dram_io_far,ls_dmnd_fills_from_sys.dram_io_near -- python -u -m torch.distributed.run --standalone --nproc-per-node=8 --run-path mlp_train.py 2>&1 | tee none_callable.txt
```
and observed
* 6.6% remote memory accesses with 'node' bindings
* 11.6% remote without bindings
I also ran similar with `str` entrypoints as before just to be sure it's still working.
NOTE: [--run-path triggers the code to be run inside a `Callable`.](017259f9c6/torch/distributed/run.py (L870))
Pull Request resolved: https://github.com/pytorch/pytorch/pull/160163
Approved by: https://github.com/d4l3k
