System Overview

You can monitor the current state of the cluster using:

Note

Everything listed below is connected together over a super-fast 25-gigabit fibre optic network.

Compute cluster

The tables below list the various batches of hardware the make up the current cluster, with each set of nodes named something ‘fun’ on a (geeky) thematic basis.

  • Year: 2023

  • Theme: Harry Potter

  • Hardware: Dell PowerEdge R6525 (1x) and XE8545 (2x)

  • CPUs/GPUs: 3rd-Gen AMD EPYC (Milan) / Nvidia Ampere

  • Capacity: 96 CPU cores (192 threads), 13,824 Cuda cores, and 864 Tensor cores

Name

CPUs

GPUs

Memory

Disk

voldemort

2x 7543 @2.8GHz 32C/64T

4.0 TB

1.9 TB (SSD)

pigwidgeon

2x 7543 @2.8GHz 32C/64T

4x A100 80 GB

512 GB

960 GB (SSD)

nagini

2x 7543 @2.8GHz 32C/64T

4x A100 80 GB

512 GB

960 GB (SSD)

delacour

2x 7543 @2.8Ghz 32C/64T

512 GB

960 GB (SSD)

snape

2x 7543 @2.8Ghz 32C/64T

512 GB

960 GB (SSD)

tonks

2x 7543 @2.8Ghz 32C/64T

512 GB

960 GB (SSD)

  • Year: 2021

  • Theme: Buck Rogers

  • Hardware: Dell PowerEdge R7525 (2x)

  • CPUs/GPUs: 2nd-Gen AMD EPYC (Rome) / Nvidia Quadro

  • Capacity: 80 CPU cores (160 threads), 9,216 Cuda cores, and 1,152 Tensor cores

Name

CPUs

GPUs

Memory

Disk

buck

2x 7502 @2.5GHz 32C/64T

2.0 TB

890 GB (SSD)

twiki

2x 7252 @3.1GHz 8C/16T

2x RTX 8000 48GB

256 GB

440 GB (SSD)

  • Year: 2019 (October)

  • Theme: Marvel Cinematic Universe

  • Hardware: Dell PowerEdge C6420 (48x) and R940xa (1x)

  • CPUs/GPUs: 2nd-Gen Intel Xeon Scalable (Cascade Lake) / Nvidia Tesla

  • Capacity: 1,600 CPU cores (3,200 threads), 10,240 Cuda cores, and 1,280 Tensor cores

Name

CPUs

GPUs

Memory

Disk

abomination

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

antman

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

blackpanther

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

blackwidow

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

captainamerica

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

captainmarvel

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

crossbones

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

doctorstrange

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

drax

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

ego

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

falcon

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

gamora

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

ghost

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

groot

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

hawkeye

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

heimdall

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

hela

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

hulk

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

ironman

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

kaeciliu

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

killmonger

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

loki

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

malekith

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

mandarin

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

mantis

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

mysterio

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

nebula

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

odin

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

quicksilver

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

redskull

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

rocket

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

ronan

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

scarletwitch

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

spiderman

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

starlord

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

taserface

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

thanos

4x Gold 6242 @2.8Ghz 16C/32T

2x V100 32GB

3.0 TB

1.3 TB (SSD)

thor

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

ultron

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

valkyrie

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

vision

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

vulture

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

warmachine

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

wasp

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

whiplash

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

wintersolider

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

wong

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

yellowjacket

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

yondu

2x Silver 4216 @2.1Ghz 16C/32T

192 GB

1.3 TB (SSD)

  • Year: 2019 (March)

  • Theme: Bond Villians

  • Hardware: Dell PowerEdge R640 (2x) and R740 (1x)

  • CPUs/GPUs: 1st-Gen Intel Xeon Scalable (Skylake) / Nvidia Tesla

  • Capacity: 80 CPU cores (160 threads), 5,120 Cuda cores, and 640 Tensor cores

Name

CPUs

GPUs

Memory

Disk

jaws

2x Silver 4116 @2.1GHz 12C/24T

1x V100 16GB

192 GB

900 GB (SSD)

oddjob

2x Gold 6132 @2.6GHz 14C/28T

384 GB

1.3 TB (SSD)

nicknack

2x Gold 6132 @2.6GHz 14C/28T

384 GB

1.3 TB (SSD)

  • Year: 2017

  • Theme: Battlestar Galactica

  • Hardware: Dell PowerEdge R630 (3x)

  • CPUs: Intel E5 v4 (Broadwell)

  • Capacity: 84 CPU cores (168 threads)

Name

CPUs

GPUs

Memory

Disk

apollo

2x E5-2680 v4 @2.4GHz 14C/24T

1.5 TB

370 GB (SSD)

boomer

2x E5-2680 v4 @2.4GHz 14C/24T

256 GB

370 GB (SSD)

starbuck

2x E5-2680 v4 @2.4GHz 14C/24T

256 GB

370 GB (SSD)

Storage and backup

The BeeGFS storage array runs alongside the cluster, however the backup system is located in another location for redundancy.

  • Year: 2019 (October)

  • Theme: Back to the Future / Discworld

  • Hardware: Dell PowerEdge R740xa (4x), MD1400 (4x), R640 (1x), and ME484 (1x)

  • CPUs: 2nd-Gen Intel Xeon Scalable (Cascade Lake)

  • Capacity: 1.5 PB (storage) and 1.0 PB (backup)

Name

CPUs

Memory

Disk

delorean

2x Silver 4208 @2.1G 8C/16T

32 GB

240 GB (SSD)

delorean (exp)

1.0 PB (HDD)

librarian

2x Silver 4208 @2.1G 8C/16T

192 GB

3.8 TB (SSD)

rincewind

2x Silver 4210 @2.2G 10C/20T

96 GB

288 TB (HDD)

twoflower

2x Silver 4210 @2.2G 10C/20T

96 GB

288 TB (HDD)

death

2x Silver 4210 @2.2G 10C/20T

96 GB

288 TB (HDD)

luggage

2x Silver 4210 @2.2G 10C/20T

96 GB

288 TB (HDD)

rincewind (exp)

144 TB (HDD)

twoflower (exp)

144 TB (HDD)

death (exp)

144 TB (HDD)

luggage (exp)

144 TB (HDD)

  • Year: 2023

  • Theme: Back to the Future

  • Hardware: Dell PowerEdge R7525 (1x) and ME5084 (1x)

  • CPUs: 3rd-Gen AMD EPYC (Milan)

  • Capacity: 1.3 PB (backup)

Name

CPUs

Memory

Disk

docbrown

2x 7313 @3.0GHz 16C/32T

256 GB

480 GB (SSD)

docbrown (exp)

1.3 PB (HDD)

Virtual infrastructure

The rest of the infrastructure runs on VMWare ESXi hosts, with various virtual machines (VMs) for handling web servers and applications, databases, domain control, monitoring, remote access, etc. Although they have a small amount of local disk, the VMs primarily use a separate Storage Area Network (SAN) for data.

  • Year: 2019 (October)

  • Hardware: Dell PowerEdge R640 (2x)

  • CPUs: 2nd-Gen Intel Xeon Scalable (Cascade Lake)

Name

CPUs

Memory

Disk

hpcvmh01

2x Gold 6242 CPU @2.8G 16C/32T

384 GB

240 GB (SSD)

hpcvmh02

2x Gold 6242 CPU @2.8G 16C/32T

384 GB

240 GB (SSD)