Compute Servers (For EECs Staff/PhDs Only)

The Faculty of EECS has a large number of Linux-based compute servers (or grunt-servers) available for any type of computational tasks required by today’s Research, and they are accompanied by a large variety of up-to-date Software, IDEs, frameworks, libraries and toolkits.

 

Servers List

The list shows the available Compute Servers owned by the School and by Research Groups.




Name OS CPU GPU GPU RAM HTT? RAM Cores Owned Access
albans CentOS 7 2 * Intel Xeon 5318Y @ 2 * A100 80 GB Yes 512 GB 96 Hao, Yang Antennas School
ant-head1 N/A N/A N/A N/A Yes 512 GB 84 Hao, Yang Antennas School
ant-head2 N/A N/A N/A N/A Yes 512 GB 84 Hao, Yang Antennas School
agnus CentOS 7 2 * Xeon(R) Gold 5318Y CPU @ 2.10GHz N/A N/A Yes 256 GB 96 Benetos, Emmanouil C4DM
arius CentOS 7 2 * Xeon(R) Silver 4316 CPU @ 2.30GHz 1 * A40 48 GB Yes 256 GB 78 Castro, Ignacio Networks
bath Centos 7 2 * Xeon 5122 @ 3.60GHz 2 * P100 16 GB Yes 192 GB 16 EECS Systems School
bibury CentOS 7 2 * Xeon Gold 5218R @ 2.10GHz 4 * RTX 2080 Ti 44 GB Yex 257 GB 80 Dixon, Simon C4DM (AIM)
bluck CentOS 7 2 * Xeon Gold 5318Y @ 2.10GHz 8 * A5000 24 Yes 256 GB 80 Dixon, Simon C4DM (AIM)
bowers CentOS 7 2 * Xeon Gold 5318Y @ 2.10GHz 8 * A5000 24 Yes 256 GB 80 Dixon, Simon C4DM (AIM)
braintree CentOS 7 2 * E5-2637 @3.5 GHz 4 * RTX 2080 Ti 45 GB Yes 128 GB 16 Dixon, Simon C4DM (Frontier)
buxton CentOS 7 2 * E5-2637 @3.5 GHz 4 * RTX 2080 Ti 45 GB Yes 128 GB 16 Dixon, Simon C4DM (AIM)
wynne CentOS 7 2 * Xeon Gold 5218R @ 2.10GHz 8 * A5000 24 Yes 256 GB 80 Dixon, Simon C4DM (AIM)
carlisle CentOS 7 2 * Xeon(R) Silver 4112 @ 2.60GHz 4 * GTX 1080 48 GB Yes 64 GB 24 Mazzon Vision (Smart Cameras)
canterbury CentOS 7 2 * E5-2640 @ 2.50GHz 1 * K20m 4 GB Yes 128 GB 24 Dubrovka Antennas School
lincoln CentOS 7 2 * Intel Xeon 5318Y @ 2 * A100 80 GB Yes 512 GB 96 Hao, Yang Antennas School
carmathen CentOS 7 2 * Xeon-5115 CPU @ 2.40GHz 2 * Tesla V100 64 GB Yes 128 GB 40 Gong, Sean Vision
chichester CentOS 7 4 * RTX 2080 Ti Yes Gong, Sean Vision
clytemnestra CentOS 7 2 * Xeon(R) Gold 5218R @ 2.10GHz 3 * RTX 6000 69 GB Yes 289 GB 80 Patras MMV
doncaster CentOS 7 2 * Xeon-5122 @3.60 GHz 4 * RTX 2070 32 GB Yes 128 GB 16 Gong, Xiang Vision
dover CentOS 7 2 * Xeon-5122 @3.60 GHz 4 * RTX 2080 Ti 45 GB Yes 128 GB 16 EECS Systems  School
dorset CentOS 7 2 * Xeon 5218 CPU@ 2.30GHz 4 * RTX 2080 Ti 45 GB Yes 256 GB 16 Lucas, Simon Game AI
durham CentOS 7 2 * Intel Xeon 5320 @ 2.20GHz 2 * A100 80 GB Yes 512 GB 104 Alomainy, Akram Antennas School
earl-01 CentOS 7 2 * Xeon-4114 @ 2.20GHz Yes 320 GB 40 Uhlig Networks
earl-02 Ubuntu 2 * Xeon-4114 @ 2.20GHz Yes 64 GB 40 Uhlig Networks
earl-03 Ubuntu 2 * Xeon-4114 @ 2.20GHz Yes 64 GB 40 Uhlig Networks
earl-04 Ubuntu 2 * Xeon-4114 @ 2.20GHz Yes 64 GB 40 Uhlig Networks
epstein Centos 7 2 * E5-2620 @2.1 GHz 2 * K80 24 GB Yes 64 GB 32 Sandler  C4DM
exeter CentOS 7 2 * E5-2680 @ 2.70GHz 1 * K40 + 1 * K20 18 GB Yes 128 GB 32  EECS Systems  School
colchester Centos 7 2 * E5-2620 @ 2.40GHz 4 * GTX 1080 Ti 48 GB Yes 32 GB 12 Gong, Xiang  Vision
dgx1-1 Ubuntu 4 * E5-2698 @2.20 GHz 8 * P100 128 GB Yes 512 GB 80 Gong Vision
dgx1-2 Ubuntu 4 * E5-2698 @2.20 GHz 8 * P100 128 GB Yes 512 GB 80 Gong Vision (Deep Insight)
doncaster CentOS 7  2 * Xeon(R) 5122@3.60GHz  4 * RTX 2070  32 GB  Yes  128 GB  24 Gong Vision
dorchester Centos 7 2 * Xeon 5122 @ 3.60GHz  2 * P100  16 GB  Yes 192 GB 16 EECS Systems  School
dover CentOS 7  2 * Xeon(R) 5122@3.60GHz  4 * RTX 2080  48 GB  Yes  128 GB  24  EECS Systems  School
dunstable Centos 7 2 * E5-2660 @ 2.6GHz 1 * M5000 8 GB Yes 64GB 40 Cavallero  CIS
harlow Centos 7 2 * E5-2640 @ 2.4GHz Yes 148GB 32 Sadrzadeh Theory
ilkley Centos 7 2 * Xeon E5-2620 v4 @ 2.10GHz 4 * GTX 1080 Ti 48 GB Yes 128 GB 32 Gong, Sean
Vision
jackal CentOS 6 2 * X5570 @ 2.93GHz Yes 48 GB 16  EECS Systems School
kinloch Centos 7 2 * Xeon Gold 6330 @ 2GHz 4 * NVIDIA A40 48 GB Yes 512 GB 112 Poesio,Massimo Cogsci
leeds CentOS 7 2 * E5-2690 @ 2.6 GHz 2 * GTX Titan X 24 GB Yes 384 48 Vision  Vision
lichfield CentOS 7 2 * Xeon Gold 6442Y  2 * Nvidia A40  48GB Yes 256 GB 96 Liakata, Maria NLP
lime-01 Ubuntu 2 * Xeon-4210R @ 2.4G 1 * RTX 2080 Ti 12 GB Yes 768 GB 10 Antichi, Gianni Networks
medusa Centos 7 2 * Xeon 5122 @ 3.60GHz 2 * P100 16 GB yes 192 GB 16 I. Patras MMV
monmouth Centos 7 2 * Intel Xeon Gold 5318S @ 2.10GHz 2 * Nvidia A40 48 GB yes 256 GB 96 Liu, Yuanwei CSR
nausicaa CentOS 7 2 * E5-2630L v4 @ 1.80GHz 4 * GTX 1080  32 GB 128 GB  40 I. Patras MMV
naples CentOS 7 2 * E5-2630L v4 @ 1.80GHz 3 * RTX 6000  66 GB 386 GB  80 Purver, Matt CogSci
pandora CentOS 7 2 * Xeon Gold 5218R CPU @ 2.10GHz 8 * RTX 3090  192 GB 258 GB  80 I. Patras MMV
penelope CentOS 7 2 * E5-2630L v4 @ 1.80GHz 4 * GTX 1080  32 GB 128 GB  40 I. Patras MMV
pisa CentOS 7  2 * E5-2643 v3 @ 3.40GHz2  2 * GTX 1080Ti  24 GB 192 GB  24 Purver,Matthew CogSci
sanctus CentOS 7 2 * Xeon Gold 5218  3 * Nvidia A30  24GB Yes 384 GB 40 Liakata, Maria NLP
towcester CentOS 7 2 * E5-26602 @ 2.2GHz 2 * K40 CUDA GPU 24 GB Yes 128 GB 40 Patras  MMV
turin CentOS 7 2 * E5-2687W v4 @ 3.00GHz 3 * GTX 1080Ti 36 GB 384 GB 48 Poesio,Massimo CogSci (DALI)
venice CentOS 7 2 * E5-2687W v4 @ 3.00GHz 3 * GTX 1080Ti 36 GB 384 GB 48 Poesio,Massimo CogSci (DALI)
whitby CentOS 7 2 * Xeon Gold 5218R @ 2.10GHz 3 * RTX 6000 69 GB Yes 289 GB 80 Dixon, Simon C4DM
wroxeter CentOS 7 2 * Xeon(R) Silver 4110 @  2.10GHz 2 * GTX 1080 Ti 22 GB Yes 128GB 32 Farkhatdinov, Ildar  Robotics

Access

Users can access the servers based on the “Access” field in the above list.

  • School: Open access to all Research members
  • <Research Group Name>: Access to members of that Research Group

Users can access the servers via SSH. Access to servers owned by Research Groups may be granted by an official request to the Group’s supervisors.

Some servers my have restricted access only to members of a Research Group working on a Research Project and will be displayed on the above table.

Shares

The following Shares are available to all Compute Servers (and Managed Desktops):

Home folder

Your EECS home directory is available to all Compute Servers and the ‘login servers’. You can write data into your home directory from Server A and they will also be available on Server B.

 

Group Shares

Group shares are available on all Compute Servers and ‘login servers’ via NFS. Typically, they can be accessible via the /import/$GROUP_SHARE_NAME, by the members of a Research Group.

For example, to access the Group Share named ‘demo-groupshare-1‘, you can access it from any Compute Server or Linux Managed Desktop via /import/demo-group-share-1 . In this example, you will be able to access that group share, only if you are a member of the ‘demogroup‘, as seen from the Group Permissions:

harry@exeter ~ $ ls -lha /import/demo-groupshare-1/
total 1.5M
drwxrws---.   4 harry demogroup 4.0K May  1 15:39 .
drwxr-xr-x. 119 root  root       0 May 10 11:48 ..
drwxr-xr-x.   2 root  demogroup 4.0K May  1 11:37 DATASETS

For the creation of a new Group Share, please raise a HelpDesk Ticket, detailing the reason and the requirements for that Share.

Software Share

Software, IDEs, packages, libraries etc are available on the compute servers either locally (based on past Requests) or centrally in the /import/linux/ Software Share. This software share is available to all Compute Servers and Managed Desktops in EECS and it is available via NFS.

Applications installed inside that Software Share may also be available as custom binaries on all Servers and Desktops, which, when users execute them, prepares the user’s environment accordingly and launch the intended application.

Most of the libraries and toolkits available in the ‘/import/linux’ Software Share are also available via the environment modules, for ease of use.

 

harry@exeter ~ $ ls /import/linux
android-studio                  erlang              opencv
atk                             feko                python
batik                           tensorflow          matlab
cuda                            ...                 ...
...
...

 

Usage

NOTE: You should not run compute-intensive processes on the ‘login-servers’ like ‘frank.eecs’ or ‘login.strudent’. Such jobs that exhaust the resources of those machines will be killed without any warning!

Use nice to make your long-running jobs give way to lower-latency jobs (like text editors..). The nice value you use should be proportional to the amount of time you expect your job to run. Given in practice the longest-running job will always finish last anyway, it’s polite not to make text editors unnecessarily slow.

Useful applications are screen and tmux, which allow you to detach from a long-running terminal, leave a task running at the background and re-attach to it later. This way, if you suddenly lose your network connection to the server unexpectedly, your processes won’t quit and your Session will still be alive.