hpcwiki

Diese Seite ist nicht editierbar. Sie können den Quelltext sehen, jedoch nicht verändern. Kontaktieren Sie den Administrator, wenn Sie glauben, dass hier ein Fehler vorliegt.
===== Massiv Parallel Compute Cluster MaPaCC 2023 =====
{{:wiki:mapacc4_frontview.jpg?400|Massiv Parallel Compute Cluster 4th generation}}\\
Massiv Parallel Compute Cluster 4th generation

Hardwarecomponents:

==== 70 compute nodes ====
  * 2x CPU Intel [[https://ark.intel.com/content/www/de/de/ark/products/232383/intel-xeon-platinum-8462y-processor-60m-cache-2-80-ghz.html|Platinum 8462Y+]] 2,2 GHz base frequency
  * 256 GByte local meeory (RAM)
  * local SSD 960 Gbyte capacity
  * HDR100 infiniband interconnect

==== 5 highmem nodes ====
  * 2x CPU AMD [[https://www.amd.com/de/products/cpu/amd-epyc-7601|AMD EPYC 7601 32-Core]] 2,2 GHz base frequency
  * 2 TeraByte local memory (RAM)
  * local SSD 512 Gbyte capacity
  * HDR100 infiniband interconnect

==== 9 visualization nodes ====
  * 2x CPU Intel 
  * 512-786 GByte local memory (RAM)
  * local SSD 512 Gbyte capacity
  * HDR100 infiniband interconnect

==== 2 login nodes ====
  * 2x CPU Intel [[https://ark.intel.com/content/www/de/de/ark/products/215276/intel-xeon-gold-6342-processor-36m-cache-2-80-ghz.html|Intel Gold 6342]] 2,3 GHz base frequency
  * 512 GByte local memory (RAM)
  * local SSD
  * QDR/FDR infiniband interconnect

==== 3 transfer nodes ====
  * 4x AMD Intel 
  * 128 GByte local memory (RAM)
  * local SSD
  * QDR/FDR infiniband interconnect
  * 
==== GPU nodes ====
  * 2x [[https://www.supermicro.com/en/Aplus/system/4U/4124/AS-4124GS-TNR.cfm|Supermicro A+ Server 4124GS-TNR]]: 2x AMD EPYC 7F52 16-Core, 512 GByte RAM, 4x Nvidia A100 PCIe, local NVMe
  * 3x [[https://www.supermicro.com/en/Aplus/system/2U/2124/AS-2124GQ-NART.cfm|Supermicro A+ Server 2124GQ-NART]]: 2x AMD EPYC 7F52 16-Core, 512 GByte RAM, 4x Nvidia A100 SXM4, local NVMe
  * 2x [[https://www.gigabyte.com/Enterprise/GPU-Server/G492-ZD0-rev-100|Gigabyte G492-ZD0]]: 2x AMD EPYC 7742 64-Core, 1024 GByte RAM, 8x Nvidia A100 SXM4, local NVMe 

==== scratch storage ====
  * 500 Terabyte capacity
  * [[https://www.beegfs.io/|Beegfs Filesystem]]
  * 4 Objectserver & 1 Metadataserver
  * FDR infiniband interconnect
  * **no backup!**

==== nfs based storage ====
  * 300 Terabyte capacity
  * 6 NFS fileserver
  * 6 filesystems
  * FDR infiniband interconnect
  * **with backup!**