High Performance Computing Systems

System Name Pluto Orion AI Cloud
Brand / Model Fujitsu Fujitsu/SMC
CPU Platform & GHz Compute & GPU: Intel 8268 
48-Cores @ 2.9 GHz
SMP: Intel Xeon 8276M 112-Cores @ 2.2GHz
Training: 128 vCPUs
Inference: 64 vCPUs
Pre/Post: 256 vCPUs
Number of nodes Compute: 112
GPU: 5 (8 x V100/node)
SMP: 2
Training: 23 (8 x H200/node)
Inference: 2 (8 x L40s/node)
Pre/Post: 4
Memory per node (GB) Compute: 192GB (4G/core)
GPU: 192GB (4G/core)
SMP: 6TB
Training: 2TB
Inference: 2TB
Pre/Post: 4TB
Storage GPFS 1PB HPS 300TB
Obj 100TB
Interconnect HDR IB 200Gbps   SuperPOD

 



High-throughput Computing Systems (To be decommission on 15 June 26)

System name Private Cloud Computing
Management Portal
https://openstack.acrc.a-star.edu.sg
System Architecture Private Cloud Platform
Operating System Linux
OS Distribution RHEL
Total Number of nodes 24
Total CPU
4272
Interconnect 100GE
Uplink 2 X 10G
Platform Red Hat Open Stack

Nodes Specifications  HPE (Non GPU)  HPE (GPU)   Dell (Non GPU)   Dell ( GPU)
Number of Compute 13  3  4  4
GPU Card ( Tesla V100) / Node
0  8  0  2
CPU Platform & GHz AMD Rome 7702 128-Cores @ 2 GHz Intel Xeon 6230 40-Cores @ 2.1 GHz Nvidia V-100 Intel(R) Xeon(R) Gold 6152 44-Cores @ 2.10 GHz Intel(R) Xeon(R) Gold 6152 44-Cores @ 2.10 GHz
Cores per node 128  40  44  44
Memory per node (GB) 512GB (4G/core)
192GB (4.8G/Core)
768 x 2 nodes
1536 x 2 nodes
768 x 3 nodes
1536 x 1 nodes