[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki

[ Back to the navigation ]

This is an old revision of the document!

Table of Contents


This page summarizes which UFAL servers have some GPU card, and suggests basic diagnostic commands, paths to installed tools, etc., simply everything necessary at the very beginning of using GPUs for experiments.

Servers with GPU units

machine GPU; Capability [cc] cores GPU RAM Cuda Theano TensorFlow Comment
titan-gpu GeForce GTX Titan Z; cc3.5 2 6 GB each core V7.5.17 works works
twister1 …has burnt
twister2 GeForce GTX 570; cc2.0 1 1 GB V7.5.17 0.7.0 no, needs cc3.0+
kronos-dev Tesla K40c; cc3.5 1 12 GB V7.5.17 0.6 0.8 works
kronos-dev Tesla K40c; cc? 1 - - - - missing power cable
now unused Quadro K2000; cc3.0 1 2 GB worked with V6.5.12 worked with 0.6.0 had ImportError: libcudart.so.7.5 (should work though)
victoria GeForce GT 630; cc3.0 1 2 GB V7.5 0.7.0 0.8 works Ondrej Bojar's desktop machine
arc GeForce GT 630; cc3.0 1 2 GB V7.5 0.7.0 0.8 works Ales/Tom's desktop machine

Summary of future plans:

Performance tests:

Installed toolkits

This should mention where each interesting toolkit lives (on a particular machine).


This script installs TensorFlow 0.7.1 (and all other dependencies we need for Multimodal Translation) into `tf' and `tf-gpu' virtual environments. The GPU environment can be loaded by calling

source tf-gpu/bin/activate-gpu

OP: I created script which install Tensorflow 0.8 and test it if it uses GPU. TF is installed into `user` or `global` installation either for `python3.4` or `python2.7`

Select GPU device

Use variable CUDA_VISIBLE_DEVICES to constrain tensorflow to compute only on the selected one. For the use of first GPU use:


Basic commands

  # is any such hardware there?
  # more details, incl. running processes on the GPU
  # nvidia-* are typically located in /usr/bin
watch nvidia-smi
  # For monitoring GPU activity in a separate terminal (thanks to Jindrich Libovicky for this!)
nvcc --version
  # this should tell CUDA version
  # nvcc is typically installed in /usr/local/cuda/bin/
  # dela to vubec neco uzitecneho? :-)
  # theano-* are typically located in /usr/local/bin/
  # shows CUDA capability etc.

GPU specs for those GPUs we have:

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