{"id":4493,"date":"2023-01-24T11:42:32","date_gmt":"2023-01-24T02:42:32","guid":{"rendered":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/?page_id=4493"},"modified":"2023-10-27T08:26:10","modified_gmt":"2023-10-26T23:26:10","slug":"newsystemc_hardware","status":"publish","type":"page","link":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/en\/system\/newsystemc\/newsystemc_hardware","title":{"rendered":"System C kugui_Hardware"},"content":{"rendered":"

Hardware Configuration<\/h2>\r\n

HPE Apollo 2000 Gen10 Plus\/HPE Apollo 6500 Gen10 Plus<\/h4>\r\n

The system consists of a standard configuration CPU node equipped with AMD Milan featuring 128 cores per node,<\/span>
\r\n and an ACC node equipped with 4 NVIDIA A100 GPUs per node. <\/span>
\r\nIt has a total theoretical computing performance of 0.973 PFLOPS and a total main memory capacity of 35.25 TB. <\/span><\/p>\r\n

 <\/p>\r\n

Hardware<\/h5>\r\n
\u00a0 – CPU node x 128 nodes<\/div>\r\n

\u00a0 \u00a0 – CPU: AMD EPYC 7763, 2.45GHz, (64core)\u00d72
\r\n\u00a0 \u00a0 – Main memory capacity: DDR4-3200, 16 GB\u00d716 = 256 GB\/node
\r\n\u00a0 \u00a0 – Theoretical performance per node
\r\n\u00a0 \u00a0 \u00a0 – Theoretical calculation performance: 5.0176 TFLOPS (2.45 GHz x 64 cores x 16 instructions x 2 sockets)
\r\n\u00a0 \u00a0 \u00a0 – Memory transfer performance: 409.6 GByte\/sec 3,200 MHz x 16 channels x 8 Bytes)
\r\n\u00a0 \u00a0 \u00a0 – Cores per node: 128 cores\/node
\r\n<\/span><\/p>\r\n

\u00a0 \u00a0 \u00a0 – Memory capacity per core: 2 GB\/core<\/div>\r\n
\u00a0 \u00a0 – Total theoretical performance: 642.25 TFLOPS
\r\n\u00a0 \u00a0 – Total capacity of main memory: 32 TB<\/div>\r\n

\u00a0<\/span><\/p>\r\n

\u00a0 – ACC node x 8 nodes<\/div>\r\n

\u00a0 \u00a0 – CPU: AMD EPYC 7763, 2.45GHz, (64core)\u00d71
\r\n<\/span><\/p>\r\n

\u00a0 \u00a0 – GPU: NVIDIA A100 40GB for HGX \u00d74<\/div>\r\n
\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 https:\/\/www.nvidia.com\/ja-jp\/data-center\/hgx\/<\/a><\/div>\r\n
\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 https:\/\/developer.nvidia.com\/blog\/nvidia-ampere-architecture-in-depth\/<\/a><\/div>\r\n

\u00a0 \u00a0 – Main memory capacity: DDR4-3200, 32GB\u00d732 = 256GB\/node
\r\n\u00a0 \u00a0 – Theoretical performance per node
\r\n\u00a0 \u00a0 \u00a0 – Theoretical computing performance
\r\n\u00a0 \u00a0 \u00a0 \u00a0 – CPU units: 2.5088 TFLOPS (2.45GHz x 64cores x 16 instructions x 1socket)
\r\n<\/span>\u00a0 \u00a0 \u00a0 \u00a0 – GPU unit: 38.8 TFLOPS (9.7 TFLOPS x 4)<\/span>
\r\n\u00a0 \u00a0 \u00a0 – Number of cores per node
\r\n\u00a0 \u00a0 \u00a0 \u00a0 – CPU unit: 64cores\/node
\r\n\u00a0 \u00a0 \u00a0 – Memory capacity per node
\r\n\u00a0 \u00a0 \u00a0 \u00a0 – CPU unit: 256GB\/node
\r\n<\/span><\/p>\r\n

\u00a0 \u00a0 \u00a0 \u00a0 – GPU unit: 160GB\/node<\/div>\r\n
\u00a0 \u00a0 – Total theoretical performance: 330.4704TFLOPS
\r\n\u00a0 \u00a0 – Total main memory capacity
\r\n\u00a0 \u00a0 \u00a0 \u00a0 – CPU unit: 2TB
\r\n\u00a0 \u00a0 \u00a0 \u00a0 – GPU unit: 1.25TB<\/div>\r\n

\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0
\r\n\u00a0 – Node-to-node communication
\r\n\u00a0 \u00a0 – Topology: FAT-tree
\r\n\u00a0 \u00a0 – Node interconnect: HDR200 InfiniBand
\r\n\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0
\r\n\u00a0 – File system
\r\n\u00a0 \u00a0 – Lustre file system
\r\n\u00a0 \u00a0 – \/home 200TB, \/work 950TB<\/span><\/p>","protected":false},"excerpt":{"rendered":"Hardware Configuration HPE Apollo 2000 Gen10 Plus\/HPE Apollo 6500 Gen10 Plus The system consists of a standard configuration CPU node equipped with AMD Milan featuring 128 cores per node, and an ACC node equipped with 4 NVIDIA A100 GPUs per node. It has a total theoretical computing performance of 0.973 PFLOPS and a total main […]","protected":false},"author":1,"featured_media":0,"parent":4491,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"page-system.php","meta":{"_acf_changed":false,"_locale":"en_US","_original_post":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/?page_id=2758","footnotes":""},"class_list":["post-4493","page","type-page","status-publish","hentry","en-US"],"acf":[],"_links":{"self":[{"href":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/wp-json\/wp\/v2\/pages\/4493","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/wp-json\/wp\/v2\/comments?post=4493"}],"version-history":[{"count":5,"href":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/wp-json\/wp\/v2\/pages\/4493\/revisions"}],"predecessor-version":[{"id":5646,"href":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/wp-json\/wp\/v2\/pages\/4493\/revisions\/5646"}],"up":[{"embeddable":true,"href":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/wp-json\/wp\/v2\/pages\/4491"}],"wp:attachment":[{"href":"https:\/\/mdcl.issp.u-tokyo.ac.jp\/scc\/wp-json\/wp\/v2\/media?parent=4493"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}