Differences

This shows you the differences between two versions of the page.

--- gpu [2018/04/23 12:53]
vodrazka [Set-up CUDA and CUDNN]
+++ gpu [2018/05/30 09:54]
vodrazka [Set-up CUDA and CUDNN]
@@ Line 12: / Line 12: @@
 | dll5 |  GeForce GTX 1080 Ti |  384.69 |  6.1 |  10 |  11 |  250 |
 | dll6 |  GeForce GTX 1080 Ti |  384.69 |  6.1 |  9 |  11 |  122 |
-| gpu |  GeForce GTX TITAN Z |  381.22 |  3.5 |  2 |  6 |  31 |
+| titan-gpu |  GeForce GTX TITAN Z |  381.22 |  3.5 |  2 |  6 |  31 |
 | iridium |  Quadro K2000 |  367.48 |  3.0 |  1 |  2 |  504 |
 | kronos |  GeForce GTX 1080 Ti |  384.81 |  6.1 |  1 |  11 |  125 |
@@ Line 64: / Line 64: @@
 You also need to use ''qsub -q gpu.q@dll[256]'' because only those machines have drivers which support CUDA 9.
-**Testing configuration (so far on twister1 & 2 only)**
+**THE NEW CLUSTER (SGE 8.1.9)**
-Multiple versions of ''cuda'' can be accessed in ''/opt''.
+Multiple versions of ''cuda'' can be accessed in ''/opt/cuda''. **Compared to the old cluster there is a difference in setting the CUDA_DIR_OPT variable!!**
-System default version for cuda library is configured in ''/etc/ld.so.conf.d/cuda.conf'' as:
-  /opt/cuda/lib64
+You need to set library path from your ''~/.bashrc'':
-  /opt/cuda/extras/CUPTI/lib64
-Actual version used depends on the link in ''/opt''. For example:
+  CUDNN_version=7.0
+  CUDA_version=9.0
-  ls -l /opt
+  CUDA_DIR_OPT=/opt/cuda/$CUDA_version
-  ...
+  if [ -d "$CUDA_DIR_OPT" ] ; then
-  lrwxrwxrwx 1 root root  8 dub  9 12:30 cuda -> cuda-9.0
+    CUDA_DIR=$CUDA_DIR_OPT
-  ...
+    export CUDA_HOME=$CUDA_DIR
+    export THEANO_FLAGS="cuda.root=$CUDA_HOME,device=gpu,floatX=float32"
-This means that the system is using ''cuda 9.0'' by default.
+    export PATH=$PATH:$CUDA_DIR/bin
+    export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$CUDA_DIR/cudnn/$CUDNN_version/lib64:$CUDA_DIR/lib64
-If system default version does not work for you, you can set library path from your ''~/.bashrc''. Your LD_LIBRARY_PATH should contain:
+    export CPATH=$CUDA_DIR/cudnn/$CUDNN_version/include:$CPATH
+  fi
-  /opt/cuda-$CUDA_version/lib64
-In order to link specific version of ''cudnn'' your LD_LIBRARY_PATH should contain:
-  /opt/cuda-$CUDA_version/cudnn/$CUDNN_version/lib64
+  * When not using Theano, just Tensorflow this can be simplified to ''export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/cuda/9.0/cudnn/7.0/lib64:/opt/cuda/8.0/lib64''.
-Please check if required CUDNN_version is available for given CUDA_version.
+  * There is no default and you always need to set ''LD_LIBRARY_PATH'' to suit explicitly.
+  * Note that ''cudnn'' library is compiled for specific version of ''cuda''. If you need specific version of ''cudnn'', you can look in ''/opt/cuda/$CUDA_version/cudnn/'' whether it is available for given ''$CUDA_version''.

[ Back to the navigation ] [ Back to the content ]

Institute of Formal and Applied Linguistics Wiki

Differences