Error when trying to use cuda in PyTorch

The error below happens when using any command related to cuda on PyTorch. I made sure my glibc is correctly at 2.14 and my conda environment contains only the PyTorch installation and corresponding dependencies. Anyone else seen this?

import torch
torch.cuda.current_device()
*** glibc detected *** python: double free or corruption (!prev): 0x00007ffff80d1180 ***
======= Backtrace: =========
/lib64/libc.so.6(+0x33f4675dee)[0x7ffff734adee]
/lib64/libc.so.6(+0x33f4678c80)[0x7ffff734dc80]
/usr/lib64/libcuda.so.1(+0x30b736)[0x7fffec01e736]
/usr/lib64/libcuda.so.1(+0x30b7a4)[0x7fffec01e7a4]
/usr/lib64/libcuda.so.1(+0x1f03d0)[0x7fffebf033d0]
/usr/lib64/libcuda.so.1(+0x1c877b)[0x7fffebedb77b]
/usr/lib64/libcuda.so.1(cuInit+0x4d)[0x7fffebf2e7cd]
/home/kumarv/willa099/miniconda3/envs/pytorch4/lib/python3.6/site-packages/torch/…/…/…/libcudart.so.9.0(+0x1c8aa)[0x7fffe433b8aa]
/home/kumarv/willa099/miniconda3/envs/pytorch4/lib/python3.6/site-packages/torch/…/…/…/libcudart.so.9.0(+0x1c901)[0x7fffe433b901]
/lib64/libpthread.so.0(pthread_once+0x53)[0x7ffff7675e03]
/home/kumarv/willa099/miniconda3/envs/pytorch4/lib/python3.6/site-packages/torch/…/…/…/libcudart.so.9.0(+0x54869)[0x7fffe4373869]
/home/kumarv/willa099/miniconda3/envs/pytorch4/lib/python3.6/site-packages/torch/…/…/…/libcudart.so.9.0(+0x18b6a)[0x7fffe4337b6a]
/home/kumarv/willa099/miniconda3/envs/pytorch4/lib/python3.6/site-packages/torch/…/…/…/libcudart.so.9.0(+0x1dd8b)[0x7fffe433cd8b]
/home/kumarv/willa099/miniconda3/envs/pytorch4/lib/python3.6/site-packages/torch/…/…/…/libcudart.so.9.0(cudaGetDeviceCount+0x4a)[0x7fffe435425a]
/home/kumarv/willa099/miniconda3/envs/pytorch4/lib/python3.6/site-packages/torch/_C.cpython-36m-x86_64-linux-gnu.so(_Z29THCPModule_isDriverSufficientP7_object+0x1e)[0x7fffeb33970e]
python(_PyCFunction_FastCallDict+0x19a)[0x7ffff7bb8a2a]
python(+0x19cdfc)[0x7ffff7c45dfc]
python(_PyEval_EvalFrameDefault+0x2fa)[0x7ffff7c6a94a]
python(+0x196f8b)[0x7ffff7c3ff8b]
python(+0x19ced5)[0x7ffff7c45ed5]
python(_PyEval_EvalFrameDefault+0x2fa)[0x7ffff7c6a94a]
python(+0x196f8b)[0x7ffff7c3ff8b]
python(+0x19ced5)[0x7ffff7c45ed5]
python(_PyEval_EvalFrameDefault+0x2fa)[0x7ffff7c6a94a]
python(+0x196f8b)[0x7ffff7c3ff8b]
python(+0x19ced5)[0x7ffff7c45ed5]
python(_PyEval_EvalFrameDefault+0x2fa)[0x7ffff7c6a94a]
python(PyEval_EvalCodeEx+0x329)[0x7ffff7c40cb9]
python(PyEval_EvalCode+0x1c)[0x7ffff7c41a4c]
python(+0x214c44)[0x7ffff7cbdc44]
python(+0xdb84e)[0x7ffff7b8484e]
python(PyRun_InteractiveLoopFlags+0xf3)[0x7ffff7b84a04]
python(+0xdbaa4)[0x7ffff7b84aa4]
python(+0xdd994)[0x7ffff7b86994]
python(main+0xee)[0x7ffff7b8975e]
/lib64/libc.so.6(__libc_start_main+0xfd)[0x7ffff72f3d1d]
python(+0x1c847b)[0x7ffff7c7147b]
======= Memory map: ========
7fff60000000-7fff60021000 rw-p 00000000 00:00 0
7fff60021000-7fff64000000 —p 00000000 00:00 0
7fff66bf8000-7fff7a777000 r-xp 00000000 00:1b 1406612009 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcudnn.so.7.1.2
7fff7a777000-7fff7a977000 —p 13b7f000 00:1b 1406612009 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcudnn.so.7.1.2
7fff7a977000-7fff7a9d2000 rw-p 13b7f000 00:1b 1406612009 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcudnn.so.7.1.2
7fff7a9d2000-7fff7aa24000 rw-p 00000000 00:00 0
7fff7aa24000-7fff7aa27000 rw-p 13bda000 00:1b 1406612009 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcudnn.so.7.1.2
7fff7aa27000-7fff82855000 r-xp 00000000 00:1b 3168206853 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcufft.so.9.0.176
7fff82855000-7fff82a55000 —p 07e2e000 00:1b 3168206853 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcufft.so.9.0.176
7fff82a55000-7fff82a64000 rw-p 07e2e000 00:1b 3168206853 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcufft.so.9.0.176
7fff82a64000-7fff82ac8000 rw-p 00000000 00:00 0
7fff82ac8000-7fff82aca000 rw-p 07e3e000 00:1b 3168206853 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcufft.so.9.0.176
7fff82aca000-7fff85cbb000 r-xp 00000000 00:1b 1140328941 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcublas.so.9.0.176
7fff85cbb000-7fff85eba000 —p 031f1000 00:1b 1140328941 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcublas.so.9.0.176
7fff85eba000-7fff85ef1000 rw-p 031f0000 00:1b 1140328941 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcublas.so.9.0.176
7fff85ef1000-7fff85f00000 rw-p 00000000 00:00 0
7fff85f00000-7fff85f03000 rw-p 03228000 00:1b 1140328941 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcublas.so.9.0.176
7fff85f03000-7fff8943f000 r-xp 00000000 00:1b 2771446159 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcusparse.so.9.0.176
7fff8943f000-7fff8963f000 —p 0353c000 00:1b 2771446159 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcusparse.so.9.0.176
7fff8963f000-7fff89659000 rw-p 0353c000 00:1b 2771446159 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcusparse.so.9.0.176
7fff89659000-7fff89669000 rw-p 00000000 00:00 0
7fff89669000-7fff8966e000 rw-p 03556000 00:1b 2771446159 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libcusparse.so.9.0.176
7fff8966e000-7fff8d459000 r-xp 00000000 00:1b 2793288561 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_core.so
7fff8d459000-7fff8d658000 —p 03deb000 00:1b 2793288561 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_core.so
7fff8d658000-7fff8d65f000 r–p 03dea000 00:1b 2793288561 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_core.so
7fff8d65f000-7fff8d68c000 rw-p 03df1000 00:1b 2793288561 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_core.so
7fff8d68c000-7fff8d6a0000 rw-p 00000000 00:00 0
7fff8d6a0000-7fff8ebdd000 r-xp 00000000 00:1b 2881525789 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_gnu_thread.so
7fff8ebdd000-7fff8eddd000 —p 0153d000 00:1b 2881525789 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_gnu_thread.so
7fff8eddd000-7fff8ede0000 r–p 0153d000 00:1b 2881525789 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_gnu_thread.so
7fff8ede0000-7fff8edf7000 rw-p 01540000 00:1b 2881525789 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_gnu_thread.so
7fff8edf7000-7fff8f6e9000 r-xp 00000000 00:1b 1976849771 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_gf_lp64.so
7fff8f6e9000-7fff8f8e8000 —p 008f2000 00:1b 1976849771 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_gf_lp64.so
7fff8f8e8000-7fff8f8ea000 r–p 008f1000 00:1b 1976849771 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_gf_lp64.so
7fff8f8ea000-7fff8f8fc000 rw-p 008f3000 00:1b 1976849771 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libmkl_gf_lp64.so
7fff8f8fc000-7fff8f902000 rw-p 00000000 00:00 0
7fff8f902000-7fff91a64000 r-xp 00000000 00:1b 432129565 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libnccl.so.1.3.5
7fff91a64000-7fff91c63000 —p 02162000 00:1b 432129565 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libnccl.so.1.3.5
7fff91c63000-7fff91c64000 r–p 02161000 00:1b 432129565 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libnccl.so.1.3.5
7fff91c64000-7fff91c65000 rw-p 02162000 00:1b 432129565 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/libnccl.so.1.3.5
7fff91c65000-7fffab2f7000 r-xp 00000000 00:1b 1480490381 /panfs/roc/groups/6/kumarv/willa099/miniconda3/envs/pytorch4/lib/python3.6/site-packages/torch/lib/libATen.soAborted

Hi,

Can you run the samples provided in the CUDA install properly?

I solved the problem by switching to a different cluster. I think it’s possible the GPU was bad. Thanks for the response, though.