HPC Resources – Hardware

hpc hardware

Technion HPC Resources – Detailed Overview

The Technion High Performance Computing (HPC) infrastructure is a robust and versatile resource designed to support the diverse computational needs of researchers and research groups.

The Technion HPC clusters consist of 200 nodes with a total of 12,036 Intel and AMD CPUs, and 86 Nvidia GPUs. All components are interconnected through a high-speed network based on 200 Gbps InfiniBand connectivity, ensuring rapid data transfer and efficient communication between nodes. Central storage based on NetApp

Researchers and research groups can procure compute nodes according to the cluster’s standard and expansion policy. These nodes will be integrated into the Zeus cluster, gaining access to high-speed communication services and storage solutions provided by the cluster. Comprehensive support is available from the cluster support team to ensure optimal performance and address any technical issues.

Below is a detailed inventory of our HPC clusters, highlighting the various configurations and capabilities available:

Cluster nameUnitCore#GPU#QtyRAM
GB
ProcessorGPUTypeCPU Utilization
Zeus Public SD530CIS960
0249216Intel Xeon Gold 6230CPU Click for details
Zeus Public C4140CIS8082768Intel Xeon Gold 6230
NVIDIA V100GPU Click for details
Zeus Public XL220CIS768
01212288Intel Xeon Platinum 8358CPU Click for details
Zeus-MK DL560CIS320004030200Intel Xeon Gold 6230
CPU Click for details
Zeus-YD SD530CIS9600249216Intel Xeon Gold 6230CPU Click for details
Zeus-MT SD530CIS400
0103840Intel Xeon Gold 6230CPU Click for details
Zeus-MT DL365CIS384023072AMD EPYC 9654CPU Click for details
Zeus SB SD530CIS5202384Intel Xeon Gold 6230CPU Click for details
Zeus MAFAT XL225CIS230401818432AMD EPYC 7713CPU Click for details
Zeus MAFAT XL170CIS6400164096Intel Xeon Gold 6148CPU Click for details
Zeus MAFAT DL385CIS19200101536AMD EPYC 9654CPU Click for details
Zeus MAFAT DL380CIS4021760Intel Xeon Gold 6230NNVIDIA V100CPU Click for details
Zeus MAFAT SX40CIS4041192Intel Xeon Gold 5115NVIDIA V100GPU Click for details
Zeus MAFAT ZFSCIS1601256Intel Xeon Silver 4110Storage
Zeus DM SD530CIS4800124608Intel Xeon Gold 6230CPU Click for details
Zeus DH DL365CIS384041536AMD EPYC 9654CPU Click for details
Zeus AGD DL365CIS576034608AMD EPYC 9654CPU Click for details
DGXCIS115272912288AMD EPYC 77428x NVIDIA A100GPU Click for details
CIS TOTAL1203686188114198

Compute Capabilities

  • CPU Nodes: Our clusters feature a variety of high-performance CPUs from Intel and AMD, including Intel Xeon Gold and Platinum series, as well as AMD EPYC processors.
  • GPU Nodes: For GPU-intensive tasks, the clusters include Nvidia V100 and A100 GPUs, providing exceptional parallel processing power.
  • Storage Nodes: Dedicated storage nodes ensure efficient data management and accessibility.

High-Speed Network

The entire HPC infrastructure is supported by a high-speed 200 Gbps InfiniBand network. This network ensures that data is transferred quickly and reliably between nodes, minimizing latency and maximizing throughput for high-performance applications.

Expansion and Support

The Technion HPC cluster is designed for scalability. Researchers can expand the cluster by adding new nodes, which will seamlessly integrate with the existing infrastructure. The support team provides comprehensive assistance, ensuring that all nodes and services operate at peak efficiency.