Leverage TensorPipe's automatic SHM address selection

[lw]

Summary: TensorPipe until now required PyTorch to come up and provide a unique identifier to use as address for the UNIX domain socket used in the SHM transport. However the Linux kernel can automatically assign an available address (like it does with IP ports), and TensorPipe now supports it, so we can remove that useless PyTorch logic.

Test Plan: CI

Differential Revision: D30220352

cc @pietern @mrshenli @pritamdamania87 @zhaojuanmao @satgera @rohan-varma @gqchen @aazzolini @osalpekar @jiayisuse @agolynski @SciPioneer @H-Huang @mrzzd @cbalioglu @gcramer23

[facebook-github-bot]

🔗 Helpful links

• 🧪 See artifacts and rendered test results at hud.pytorch.org/pr/63028
• 📄 Preview docs built from this PR

💊 CI failures summary and remediations

As of commit 45a6f46 (more details on the Dr. CI page):

---

• 1/1 failures introduced in this PR

---

🕵️ 1 new failure recognized by patterns

The following CI failures do not appear to be due to upstream breakages:

win-vs2019-cuda10.1-py3 / test (default, 1, 1, windows.8xlarge.nvidia.gpu) (1/1)

Step: "Run test scripts" (full log | diagnosis details | 🔁 rerun)

2021-08-26T14:26:02.2852374Z AssertionError: True is not false

2021-08-26T14:26:02.2840986Z ======================================================================
2021-08-26T14:26:02.2842400Z FAIL [48.434s]: test_fd_sharing (__main__.TestMultiprocessing)
2021-08-26T14:26:02.2843723Z ----------------------------------------------------------------------
2021-08-26T14:26:02.2844680Z Traceback (most recent call last):
2021-08-26T14:26:02.2847676Z   File "test_multiprocessing.py", line 327, in test_fd_sharing
2021-08-26T14:26:02.2848993Z     self._test_sharing(repeat=TEST_REPEATS)
2021-08-26T14:26:02.2849897Z   File "test_multiprocessing.py", line 283, in _test_sharing
2021-08-26T14:26:02.2850502Z     test_receive()
2021-08-26T14:26:02.2851141Z   File "test_multiprocessing.py", line 278, in test_receive
2021-08-26T14:26:02.2851798Z     self.assertFalse(p.is_alive())
2021-08-26T14:26:02.2852374Z AssertionError: True is not false
2021-08-26T14:26:02.2852725Z 
2021-08-26T14:26:02.2951818Z ----------------------------------------------------------------------
2021-08-26T14:26:02.2952503Z Ran 2 tests in 165.253s
2021-08-26T14:26:02.2952784Z 
2021-08-26T14:26:02.2953133Z FAILED (failures=1)
2021-08-26T14:26:02.2953411Z 
2021-08-26T14:26:02.2953808Z Generating XML reports...
2021-08-26T14:26:02.2955017Z Generated XML report: test-reports\python-unittest\test_multiprocessing\TEST-TestMultiprocessing-20210826142317.xml
2021-08-26T14:26:02.6355658Z Traceback (most recent call last):
2021-08-26T14:26:02.6356775Z   File "run_test.py", line 1011, in <module>

---

This comment was automatically generated by Dr. CI (expand for details).

Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions to the (internal) Dr. CI Users group.

Click here to manually regenerate this comment.

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D30220352

[mrshenli]

oops, test failures are real:

RuntimeError: In addrImplFromLoop at tensorpipe/transport/shm/listener_impl.cc:97 "errorgetsockname: Bad file descriptor (this error originated at tensorpipe/common/socket.cc:129)"

[lw]

Yep, I'm debugging that, I should have marked this as WIP sorry

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D30220352

[codecov]

Codecov Report

Merging #63028 (45a6f46) into master (61d88cd) will increase coverage by 7.52%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master   #63028      +/-   ##
==========================================
+ Coverage   59.23%   66.75%   +7.52%     
==========================================
  Files         695      695              
  Lines       90739    90739              
==========================================
+ Hits        53751    60577    +6826     
+ Misses      36988    30162    -6826     

[mrshenli]

LGTM!

[facebook-github-bot]

This pull request has been merged in 48c57b9.