techstaff:aicluster
Differences
This shows you the differences between two versions of the page.
Both sides previous revisionPrevious revisionNext revision | Previous revisionLast revisionBoth sides next revision | ||
techstaff:aicluster [2020/11/11 11:31] – kauffman | techstaff:aicluster [2020/12/10 12:57] – [Batch] kauffman | ||
---|---|---|---|
Line 5: | Line 5: | ||
Feedback is requested: | Feedback is requested: | ||
- | # | + | [[https:// |
Knowledge of how to use Slurm already is preferred at this stage of testing. | Knowledge of how to use Slurm already is preferred at this stage of testing. | ||
- | ===== TODO ===== | + | The information from the older cluster mostly applies and I suggest you read that documentation: |
- | Since I'm still working on it, I don't guarantee any uptime yet. Mainly I need to make sure TRES tracking is working like we want. This will involve restarting slurmd and slurmctld which will kill running jobs. | ||
+ | ====== Infrastructure ====== | ||
+ | Summary of nodes installed on the cluster | ||
- | * < | + | ===== Computer/GPU Nodes ===== |
- | * < | + | * 6x nodes |
- | * < | + | * 2x Xeon Gold 6130 CPU @ 2.10GHz (64 threads) |
- | * < | + | * 192G RAM |
- | * < | + | * 4x Nvidia GeForce RTX2080Ti |
- | * < | + | |
- | * < | + | |
- | * < | + | |
- | * < | + | |
- | * < | + | |
- | * < | + | |
- | * < | + | |
- | * < | + | |
- | * figure why summary view is no longer a thing. | + | |
- | * < | + | |
- | * < | + | |
- | * < | + | |
- | * home directory | + | |
- | * setup backups for home dirs | + | |
- | * default quota | + | |
- | * home directory usage report | + | |
+ | * 2x nodes | ||
+ | * 2x Xeon Gold 6130 CPU @ 2.10GHz (64 threads) | ||
+ | * 384G RAM | ||
+ | * 4x Nvidia Quadro RTX 8000 | ||
+ | |||
+ | * all: | ||
+ | * zfs mirror mounted at /local | ||
+ | * compression to lz4: Usually this has a performance gain as less data is read and written to disk with a small overhead in CPU usage. | ||
+ | * As of right now there is no mechanism to clean up /local. At some point I'll probably put a find command in cron that deletes files older than 90 days or so. | ||
+ | |||
+ | ===== Storage ===== | ||
+ | |||
+ | * ai-storage1: | ||
+ | * 41T total storage | ||
+ | * uplink to cluster network: 2x 25G | ||
+ | * / | ||
+ | * We intend to set user quotas, however, there are no quotas right now. | ||
+ | * / | ||
+ | * Lives on the home directory server. | ||
+ | * Idea would be to create a dataset with a quota for people to use. | ||
+ | * Normal LDAP groups that you are used to and available everywhere else would control access to these directories. e.g. jonaslab, sandlab | ||
+ | |||
+ | * ai-storage2: | ||
+ | * 41T total storage | ||
+ | * uplink to cluster network: 2x 25G | ||
+ | * / | ||
+ | * Eventually data will be auto deleted after X amount of time. Maybe 90 days or whatever we determine makes sense. | ||
+ | |||
+ | * ai-storage3: | ||
+ | * zfs mirror with previous snapshots of ' | ||
+ | * NOT a backup. | ||
+ | * Not enabled yet. | ||
+ | |||
+ | |||
+ | ====== Login ====== | ||
+ | There are a set of front end nodes that give you access to the Slurm cluster. You will connect through these nodes and need to be on these nodes to submit jobs to the cluster. | ||
+ | |||
+ | ssh cnetid@fe.ai.cs.uchicago.edu | ||
+ | |||
+ | * Requires a CS account. | ||
+ | |||
+ | ==== File Transfer ==== | ||
+ | You will use the FE nodes to transfer your files onto the cluster storage infrastructure. The network connections on those nodes are 2x 10G each. | ||
+ | |||
+ | === Quota === | ||
+ | * By default users are given a quota of 20G. | ||
====== Demo ====== | ====== Demo ====== | ||
- | kauffman3 is my test CS account. | + | kauffman3 is my CS test account. |
< | < | ||
Line 92: | Line 123: | ||
</ | </ | ||
+ | ==== Notes on CUDA_VISIBLE_DEVICES ==== | ||
+ | CUDA_VISIBLE_DEVICES: | ||
- | ===== Fairshare ===== | + | * This variable should NOT be modified. Ever. |
+ | * Relative means that if you requested one gpu it will show up as 0. Even if all other gpus on the server are being used by others. | ||
- | # Check out the fairshare values | ||
- | < | ||
- | kauffman3@fe01: | ||
- | | ||
- | -------------------- ---------- ---------- ----------- ----------- ----------- ------------- ---------- ---------- ------------------------------ ------------------------------ | ||
- | kauffman3 | ||
- | | ||
- | kauffman4 | ||
- | | ||
- | </ | ||
+ | ===== Fairshare/ | ||
+ | By default all usage is tracked and charged to a users default account. A fairshare value is computed and used in prioritizing a job on submission. | ||
- | We are using the FairTree (fairshare algorithm). This is the default in Slurm these days and from what I can tell probably better suits our needs. It is no big deal to change to classic fairshare. | + | Details |
- | + | You will need to charge an account | |
- | As the system | + | |
- | + | ||
- | < | + | |
- | | + | |
- | | + | |
- | User: kauffman | + | |
- | </ | + | |
- | We will probably assign fairshare points | + | |
- | + | ||
- | + | ||
- | + | ||
- | ====== Storage ====== | + | |
- | + | ||
- | / | + | |
- | | + | |
- | + | ||
- | / | + | |
- | Lives on the home directory server. | + | |
- | Idea would be to create a dataset with a quota for people | + | |
- | Normal LDAP groups | + | |
- | e.g. jonaslab, sandlab | + | |
- | + | ||
- | + | ||
- | Currently there is no quota on home directories. This is set per user per dataset. | + | |
- | + | ||
- | + | ||
- | I was able to get homes and scratch each connected via 2x 25G. Both are SSD only so the storage should be FAST. | + | |
- | + | ||
- | + | ||
- | Each compute node (nodes with gpus) has a zfs mirror mounted at /local | + | |
- | I set compression to lz4 by default. Usually this has a performance gain as less data is read and written to disk with a small overhead in CPU usage. | + | |
- | As of right now there is no mechanism to clean up /local. At some point I'll probably put a find command in cron that deletes files older than 90 days or so. | + | |
Line 147: | Line 141: | ||
> Do we have a max job runtime? | > Do we have a max job runtime? | ||
- | Yes. 4 hours. This is done per partition. | + | Yes. 4 hours. This is done per partition. |
< | < | ||
PartitionName=geforce Nodes=a[001-006] Default=YES DefMemPerCPU=2900 MaxTime=04: | PartitionName=geforce Nodes=a[001-006] Default=YES DefMemPerCPU=2900 MaxTime=04: | ||
Line 156: | Line 151: | ||
- | You can take a look at all the values we set here: | + | ===== Jupyter Notebook Tips ===== |
+ | ==== Batch ==== | ||
+ | The process for a batch job is very similar. | ||
+ | |||
+ | jupyter-notebook.sbatch | ||
+ | < | ||
+ | # | ||
+ | unset XDG_RUNTIME_DIR | ||
+ | NODEIP=$(hostname -i) | ||
+ | NODEPORT=$(( $RANDOM + 1024)) | ||
+ | echo "ssh command: ssh -N -L 8888: | ||
+ | . ~/ | ||
+ | jupyter-notebook --ip=$NODEIP --port=$NODEPORT --no-browser | ||
+ | </ | ||
+ | |||
+ | Check the output of your job to find the ssh command to use when accessing your notebook. | ||
+ | |||
+ | Make a new ssh connection to tunnel your traffic. The format will be something like: | ||
+ | |||
+ | '' | ||
- | fe0[1,2]$ cat /etc/slurm-llnl/slurm.conf | + | This command will appear to hang since we are using the -N option which tells ssh not to run any commands including a shell on the remote machine. |
- | The man page: https://slurm.schedmd.com/slurm.conf.html | + | Open your local browser and visit: '' |
+ | ==== Interactive ==== | ||
+ | - '' | ||
+ | - '' | ||
+ | - '' | ||
+ | - '' | ||
+ | - '' | ||
+ | - '' | ||
+ | - Make a new ssh connection with a tunnel to access your notebook | ||
+ | - '' | ||
+ | - This will make an ssh tunnel on your local machine that fowards traffic sent to '' | ||
+ | - Open your local browser and visit: '' | ||
/var/lib/dokuwiki/data/pages/techstaff/aicluster.txt · Last modified: 2021/01/06 16:11 by kauffman