ccu, Administrators
684
edits
Changes
Jump to navigation
Jump to search
m
Note: the GPU cluster recently moved to a new hardware infrastructure, which means that a lot of the documentation below is outdated. So far, the only page which is up to date is [https://ccu.uni-konstanz.de/wiki/CCU:GPU_Cluster_Quick_Start the new cluster quick start documentation]. Please only refer to this doc to get started, and ignore the rest for now.
== How The following information is partially outdated, and refers to get started ==older system architectures (Ubuntu 18.04). In particular, the install scripts probably do not work anymore. Instead, refer to current online documentation on how to install e.g. nvidia drivers and nvidia docker.
** Step 3: [[Tutorials:Link to container registry on our server|Link to container registry on our server]]
** Step 4: [[Tutorials:Persistent volumes on the Kubernetes cluster|Persistent volumes on the GPU cluster]]
** Step 5: [[Tutorials:Monitoring with Tensorboard on the GPU cluster|Monitoring with Tensorboard on the GPU cluster]]
== Tips and Tricks ==
** [[Tips:How to ensure your pod ends up on a specific node|How to ensure your pod ends up on a specific node]]
== Reference documents ==
* [[Cluster:Namespaces|Which namespaces am I allowed to use?]]
* [[Cluster:Nodes|Which compute nodes are available?]]
* [[Cluster:Nodes|Which namespace has access to which compute nodes?]]
→How to get started
This means for you that you have to be able to take the necessary steps to wrap your own code into a container. All this is covered in an easy, introductory way in the short tutorials below, which should be sufficient to get you started. At some point, you might want to learn about docker in a more in-depth manner, for this, I refer you to the excellent tutorials available elsewhere, some of which are linked [[CCU:tutorials|here]].
== Compute nodes ==
See [[Cluster:Compute_nodes|this page]] for a current list of compute nodes, their hardware, and which groups they serve.
== Changelog ==
See [[Cluster:Changelog|this page]] for a history of cluster updates and necessary changes on your side.
== What you need ==
* An account for the CCU.
* Ideally, a desktop PC with an nVidia GPU to test your code before pushing it to the cluster. However, you can develop for and control the cluster on any machine, it's not mandatory that you can actually run the code locally. Note, however, that it makes debugging harder if you cannot do this (you have to do everything on the console).
* Your PC ideally runs a flavor of Linux, all example scripts were tested against Ubuntu 1820.04 (should also work on derivatives, such as Linux Mint 19, based on that Ubuntu edition). If you use Windows, you are on your own.
* Admin access to your own PC to install lots of stuff (or a friendly administrator).
* More specific needs will be covered in the in-depth tutorials.
== How to get started ==
Most up to date information for the current cluster:
* [[CCU:GPU_Cluster_Quick_Start | Quick start tutorial]]
* [[CCU:Perstistent storage on the Kubernetes cluster | How to use persistent storage]]
* [[Tutorials:Link_to_container_registry_on_our_server | How to use the CCU image repository]]
* [[Tutorials:Mount_cifs_storage_in_a_pod | How to mount cifs storage]]
* [[Cluster:Compute nodes | How to target different compute nodes]]
* Preparing your system
** Step 1: [[Tutorials:Install nVidia CUDA and GPU drivers|Install nVidia CUDA and GPU drivers]]
** Step 2: [[Tutorials:Install the nVidia docker system|Install the nVidia docker system]]
** For the impatient: [[Tutorials:Complete install script for a fresh Ubuntu 18.04|Complete install script for a fresh Ubuntu 18.04]]
These tutorials should still work: * Learning the basics of Docker(requires docker or nvidia-docker for the GPU containers)** [[Tutorials:Link to container registry on our server|Link to container registry on our server]]
** An in-depth look at a [[Example:container which trains MNIST using Tensorflow|container which trains MNIST using Tensorflow]], with the following steps:
*** Step 1: create a local python tensorflow application.
** Step 2: [[Tutorials:Set up your Kubernetes user account|Set up your Kubernetes user account]]
** Step 3: [[Tutorials:Run the example container on the cluster|Run the example container on the cluster]] and make sure that it works correctly.
