[theano-users] Error - cudnn conflict

[Javier Coronel]

Hi guys,


I've been working wih theano for a while with no problem in a shared
cluster in my university, however, a few days ago an error showed up when
running my usual training session. When I run my code the next warning is
showed and then the following error appears:


*Warning *Using gpu device 0: Tesla K20c (CNMeM is disabled, cuDNN 6021)
/lustre/apps/anaconda/2/lib/python2.7/site-packages/theano/sandbox/cuda/__init__.py:631:

UserWarning: Your cuDNN version is more recent than the one Theano
officially supports. If you see any problems, try updating Theano or
downgrading cuDNN to version 5.1.


*Error *... File
"/lustre/apps/anaconda/2/lib/python2.7/site-packages/theano/sandbox/cuda/nvcc_compiler.py",
line 405, in compile_str 'for cmd', ' '.join(cmd)) Exception: ('The
following error happened while compiling the node',
GpuDnnConv{algo='small', inplace=True}(GpuContiguous.0, GpuContiguous.0,
GpuAllocEmpty.0, GpuDnnConvDesc{border_mode='valid',
.... I/hpcfs/home/ingenieria/biomedica/postgrado/jd.coronel30/.theano/compiledir_Linux-2.6-el6.x86_64-x86_64-with-centos-6.5-Final-x86_64-2.7.13-64/cuda_ndarray
-I/usr/local/cuda-7.5/include
-I/lustre/apps/anaconda/2/lib/python2.7/site-packages/theano/sandbox/cuda
... /hpcfs/home/ingenieria/biomedica/postgrado/jd.coronel30/.theano/compiledir_Linux-2.6-el6.x86_64-x86_64-with-centos-6.5-Final-x86_64-2.7.13-64/tmp02jhaS/ea4e203b6529466794536f8a1bfa77ae.so
mod.cu -lcudart -lcublas -lcuda_ndarray -lcudnn -lpython2.7',
"[GpuDnnConv{algo='small', inplace=True}(, , , , Constant{1.0},
Constant{0.0})]")



The theano version installed is the last one and I don't know what is the
problem, anyone could help me with this?



Thanks in advance!



Javier Coronel
------------------------------

[Pascal Lamblin]

You seem to be using the old GPU back-end, which does not support your
version of cuDNN.

In fact, you seem to have an older version of Theano as well, as the
latest release, 0.9, should display a warning prompting you to swith to
the new GPU back-end,
https://github.com/Theano/Theano/wiki/Converting-to-the-new-gpu-back-end%28gpuarray%29

Post by Javier Coronel
Hi guys,
I've been working wih theano for a while with no problem in a shared
cluster in my university, however, a few days ago an error showed up
when running my usual training session. When I run my code the next
*Warning *Using gpu device 0: Tesla K20c (CNMeM is disabled, cuDNN 6021)
UserWarning: Your cuDNN version is more recent than the one Theano
officially supports. If you see any problems, try updating Theano or
downgrading cuDNN to version 5.1.
*Error *... File
"/lustre/apps/anaconda/2/lib/python2.7/site-packages/theano/sandbox/cuda/nvcc_compiler.py",
line 405, in compile_str 'for cmd', ' '.join(cmd)) Exception: ('The
following error happened while compiling the node',
GpuDnnConv{algo='small', inplace=True}(GpuContiguous.0, GpuContiguous.0,
GpuAllocEmpty.0, GpuDnnConvDesc{border_mode='valid',
.... I/hpcfs/home/ingenieria/biomedica/postgrado/jd.coronel30/.theano/compiledir_Linux-2.6-el6.x86_64-x86_64-with-centos-6.5-Final-x86_64-2.7.13-64/cuda_ndarray
-I/usr/local/cuda-7.5/include
-I/lustre/apps/anaconda/2/lib/python2.7/site-packages/theano/sandbox/cuda ... /hpcfs/home/ingenieria/biomedica/postgrado/jd.coronel30/.theano/compiledir_Linux-2.6-el6.x86_64-x86_64-with-centos-6.5-Final-x86_64-2.7.13-64/tmp02jhaS/ea4e203b6529466794536f8a1bfa77ae.so
mod.cu -lcudart -lcublas -lcuda_ndarray -lcudnn -lpython2.7',
"[GpuDnnConv{algo='small', inplace=True}(, , , , Constant{1.0},
Constant{0.0})]")
The theano version installed is the last one and I don't know what is
the problem, anyone could help me with this?
Thanks in advance!
Javier Coronel
------------------------------------------------------------------------
--
---
You received this message because you are subscribed to the Google
Groups "theano-users" group.
To unsubscribe from this group and stop receiving emails from it, send
For more options, visit https://groups.google.com/d/optout.