pytorch - ✅(Solved) Fix ImportError: /usr/local/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so: undefined symbol: ncclAlltoAll [1 pull requests, 5 comments, 5 participants]

pytorch2026-03-08 15:36:23

ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

GitHub issue URL

Helpful · Quick feedback

GitHub stats

pytorch/pytorch#176829•Fetched 2026-04-08 00:24:16

View on GitHub

Comments

Participants

Timeline

Reactions

Author

Participants

Timeline (top)

mentioned ×21subscribed ×21labeled ×7commented ×5

Error Message

ImportError: /usr/local/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so: undefined symbol: ncclAlltoAll

Fix Action

Fixed

Fixed by PR: [BE] Add in-place collect_env execution (https://github.com/pytorch/pytorch/pull/176904)

PR fix notes

PR #176904: [BE] Add in-place collect_env execution

Repository: pytorch/pytorch
Author: malfet
State: closed | merged: False
Link: https://github.com/pytorch/pytorch/pull/176904

Description (problem / solution / changelog)

Stack from ghstack (oldest at bottom):

-> #176904

First of all, fewer systems now how wget installed by default, but almost all Linux/MacOS comes with curl

If script with such name already exists, wget will download it with .${NUM} suffixed alias, which results in reporter posting results from something else, for example see https://github.com/pytorch/pytorch/issues/176829

wget https://raw.githubusercontent.com/pytorch/pytorch/main/torch/utils/collect_env.py
python collect_env.py
--2026-03-08 23:35:37--  https://raw.githubusercontent.com/pytorch/pytorch/main/torch/utils/collect_env.py
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 2606:50c0:8001::154, 2606:50c0:8003::154, 2606:50c0:8000::154, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|2606:50c0:8001::154|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 31107 (30K) [text/plain]
Saving to: ‘collect_env.py.1’

collect_env.py.1            100%[==========================================>]  30.38K  --.-KB/s    in 0.04s

2026-03-08 23:35:38 (707 KB/s) - ‘collect_env.py.1’ saved [31107/31107]

Traceback (most recent call last):
  File "/mnt/d/my/work/study/ai/kaggle_code/aimo2/collect_env.py", line 15, in <module>

Changed files

.github/ISSUE_TEMPLATE/bug-report.yml (modified, +5/-1)
.github/ISSUE_TEMPLATE/pt2-bug-report.yml (modified, +4/-0)

Code Example

pip3 install torch torchvision --index-url https://download.pytorch.org/whl/cu130

---

ImportError: /usr/local/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so: undefined symbol: ncclAlltoAll

---

wget https://raw.githubusercontent.com/pytorch/pytorch/main/torch/utils/collect_env.py
# For security purposes, please check the contents of collect_env.py before running it.
python collect_env.py
--2026-03-08 23:35:37--  https://raw.githubusercontent.com/pytorch/pytorch/main/torch/utils/collect_env.py
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 2606:50c0:8001::154, 2606:50c0:8003::154, 2606:50c0:8000::154, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|2606:50c0:8001::154|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 31107 (30K) [text/plain]
Saving to: ‘collect_env.py.1’

collect_env.py.1            100%[==========================================>]  30.38K  --.-KB/s    in 0.04s   

2026-03-08 23:35:38 (707 KB/s) - ‘collect_env.py.1’ saved [31107/31107]

Traceback (most recent call last):
  File "/mnt/d/my/work/study/ai/kaggle_code/aimo2/collect_env.py", line 15, in <module>
    from vllm.envs import environment_variables
  File "/usr/local/lib/python3.12/site-packages/vllm/__init__.py", line 14, in <module>
    import vllm.env_override  # noqa: F401
    ^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.12/site-packages/vllm/env_override.py", line 87, in <module>
    import torch
  File "/usr/local/lib/python3.12/site-packages/torch/__init__.py", line 431, in <module>
    from torch._C import *  # noqa: F403
    ^^^^^^^^^^^^^^^^^^^^^^
ImportError: /usr/local/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so: undefined symbol: ncclAlltoAll

RAW_BUFFERClick to expand / collapse

🐛 Describe the bug

pip3 install torch torchvision --index-url https://download.pytorch.org/whl/cu130

Official source. The Alibaba source does not have this problem. then get error

ImportError: /usr/local/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so: undefined symbol: ncclAlltoAll

The same command didn't have this problem earlier, but there were the following incompatibility issues. However, the latest same link encountered this error. https://github.com/vllm-project/vllm/issues/36302

Versions

wget https://raw.githubusercontent.com/pytorch/pytorch/main/torch/utils/collect_env.py
# For security purposes, please check the contents of collect_env.py before running it.
python collect_env.py
--2026-03-08 23:35:37--  https://raw.githubusercontent.com/pytorch/pytorch/main/torch/utils/collect_env.py
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 2606:50c0:8001::154, 2606:50c0:8003::154, 2606:50c0:8000::154, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|2606:50c0:8001::154|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 31107 (30K) [text/plain]
Saving to: ‘collect_env.py.1’

collect_env.py.1            100%[==========================================>]  30.38K  --.-KB/s    in 0.04s   

2026-03-08 23:35:38 (707 KB/s) - ‘collect_env.py.1’ saved [31107/31107]

Traceback (most recent call last):
  File "/mnt/d/my/work/study/ai/kaggle_code/aimo2/collect_env.py", line 15, in <module>
    from vllm.envs import environment_variables
  File "/usr/local/lib/python3.12/site-packages/vllm/__init__.py", line 14, in <module>
    import vllm.env_override  # noqa: F401
    ^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.12/site-packages/vllm/env_override.py", line 87, in <module>
    import torch
  File "/usr/local/lib/python3.12/site-packages/torch/__init__.py", line 431, in <module>
    from torch._C import *  # noqa: F403
    ^^^^^^^^^^^^^^^^^^^^^^
ImportError: /usr/local/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so: undefined symbol: ncclAlltoAll

cc @ezyang @gchanan @kadeng @msaroufim @seemethere @malfet @atalman @tinglvv @nWEIdia @awgu @wanchaol @fegin @fduwjj @wz337 @wconstab @d4l3k @pragupta @dcci @aditvenk @xmfan

extent analysis

Fix Plan

1. Update PyTorch and Torchvision

Update PyTorch and torchvision to the latest version using pip. This might resolve the issue.

pip3 install --upgrade torch torchvision --index-url https://download.pytorch.org/whl/cu130

2. Check for Conflicting Packages

Run the collect_env.py script to check for any conflicting packages. If the script reports any issues, try to resolve them before proceeding.

wget https://raw.githubusercontent.com/pytorch/pytorch/main/torch/utils/collect_env.py
python collect_env.py

3. Remove Conflicting Packages

If the collect_env.py script reports any conflicting packages, try to remove them. For example, if it reports a conflict with vllm, try to uninstall it.

pip3 uninstall vllm

4. Reinstall PyTorch and Torchvision

After removing any conflicting packages, try reinstalling PyTorch and torchvision.

pip3 install torch torchvision --index-url https://download.pytorch.org/whl/cu130

Verification

To verify that the fix worked, try running the original command again.

pip3 install torch torchvision --index-url https://download.pytorch.org/whl/cu130

If the installation completes successfully, the issue should be resolved.

Vote matrix · Quick signals

Works

Did the solution work? Tap to confirm.

Easy Fix

Was it a quick fix?

Time Saver

Did it save you time?

Blocking

Was it severely blocking?

Common Issue

Are others likely hitting this too?

Flaky / Intermittent

Is it intermittent?

Verified / Reproducible

Can you reproduce it reliably?

#api #ssr #installation #tensor shape #autograd error #inference speed #output truncation #response parsing #generation error #database connection

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Data

Security

Network

Code

UI/UX

Text

System

Multimedia

Protocol

API

Engineering

pytorch - ✅(Solved) Fix ImportError: /usr/local/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so: undefined symbol: ncclAlltoAll [1 pull requests, 5 comments, 5 participants]

Recommended Tools

GitHub issue graph ai analysis

Error Message

Fix Action

Fixed

PR fix notes

PR #176904: [BE] Add in-place collect_env execution

Description (problem / solution / changelog)

Changed files

Code Example

🐛 Describe the bug

Versions

extent analysis

Fix Plan

1. Update PyTorch and Torchvision

2. Check for Conflicting Packages

3. Remove Conflicting Packages

4. Reinstall PyTorch and Torchvision

Verification

Still need to ship something?

TRENDING

pytorch - ✅(Solved) Fix ImportError: /usr/local/lib/python3.12/site-packages/torch/lib/libtorch_cuda.so: undefined symbol: ncclAlltoAll [1 pull requests, 5 comments, 5 participants]

Recommended Tools

GitHub issue graph ai analysis

Error Message

Fix Action

Fixed

PR fix notes

PR #176904: [BE] Add in-place collect_env execution

Description (problem / solution / changelog)

Changed files

Code Example

🐛 Describe the bug

Versions

extent analysis

Fix Plan

1. Update PyTorch and Torchvision

2. Check for Conflicting Packages

3. Remove Conflicting Packages

4. Reinstall PyTorch and Torchvision

Verification

Still need to ship something?

RELATED_DISCOVERY

TRENDING