Fix "unable to open shared memory" while using MPFuture #517
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
borzunov
force-pushed
the
fix-mpfuture-shmem
branch
from
d2c7433
to
ea6e3fc
Compare
borzunov
commented
Nov 1, 2022
| @@ -303,7 +303,7 @@ def __del__(self): | |||
| def __getstate__(self): | |||
| return dict( | |||
| _sender_pipe=self._sender_pipe, | |||
| _shared_state_code=self._shared_state_code, | |||
| _shared_state_code=ForkingPickler.dumps(self._shared_state_code).tobytes(), | |||
There was a problem hiding this comment.
Despite the name, it works correctly with all multiprocessing start methods (fork, spawn, and forkserver). This may be tested using this short script:
import multiprocessing as mp
import torch
import time
from multiprocessing.reduction import ForkingPickler
torch.multiprocessing.set_sharing_strategy("file_system")
def foo(q):
t = torch.zeros(3).share_memory_()
q.put(ForkingPickler.dumps(t).tobytes())
t[1] = 777
if __name__ == '__main__':
mp.set_start_method('spawn')
q = mp.Queue()
p = mp.Process(target=foo, args=(q,))
p.start()
time.sleep(1)
print(ForkingPickler.loads(q.get()))
p.join()
Codecov Report
@@ Coverage Diff @@
## master #517 +/- ##
==========================================
- Coverage 75.97% 75.91% -0.07%
==========================================
Files 81 81
Lines 7943 7946 +3
==========================================
- Hits 6035 6032 -3
- Misses 1908 1914 +6
|
justheuristic
approved these changes
Nov 1, 2022
mryab
pushed a commit
that referenced
this pull request
Nov 29, 2022
Currently, one may sometimes get the "unable to open shared memory" error (see the screenshot) while using `hivemind.MPFuture`. Interestingly, the smaller `HIVEMIND_SHM_BUFFER_SIZE` is, the more often the error occurs (e.g., in Petals, it occurs right after starting the server if `HIVEMIND_SHM_BUFFER_SIZE=2`). Turns out, it happens when the origin process garbage-collects all instances of MPFuture using the same shmem buffer, then the underlying buffer is freed, and target processes can't reconnect to it anymore when unpickling its instances of MPFuture. This PR fixes this important issue. (cherry picked from commit 94c985d)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Currently, one may sometimes get the "unable to open shared memory" error (see the screenshot) while using
hivemind.MPFuture. Interestingly, the smallerHIVEMIND_SHM_BUFFER_SIZEis, the more often the error occurs (e.g., in Petals, it occurs right after starting the server ifHIVEMIND_SHM_BUFFER_SIZE=2).Turns out, it happens when the origin process garbage-collects all instances of MPFuture using the same shmem buffer, then the underlying buffer is freed, and target processes can't reconnect to it anymore when unpickling its instances of MPFuture.
This PR fixes this important issue.