{"id":7238,"date":"2025-08-26T14:03:45","date_gmt":"2025-08-26T05:03:45","guid":{"rendered":"https:\/\/mdx.jp\/?post_type=mdx2_news&#038;p=7238"},"modified":"2026-02-16T13:27:53","modified_gmt":"2026-02-16T04:27:53","slug":"mdx-ii-gpu%e3%83%8e%e3%83%bc%e3%83%89%e6%8b%a1%e5%bc%b5%e4%ba%88%e5%ae%9a%e3%81%ab%e3%81%a4%e3%81%84%e3%81%a6%e3%81%ae%e3%81%8a%e7%9f%a5%e3%82%89%e3%81%9b-2026-4%ef%bd%9e","status":"publish","type":"mdx2_news","link":"https:\/\/mdx.jp\/en\/mdx2\/news\/7238","title":{"rendered":"Notice on Planned Expansion of MDX II GPU Nodes (from April 2026)"},"content":{"rendered":"<p>From April 2026, <strong>8 additional GPU nodes<\/strong> will be introduced into <strong>mdx II<\/strong>.<br \/>\n&nbsp;<br \/>\nWe are currently preparing for the start of this service.<br \/>\nDetails such as the application period will be announced on the official website and by email as soon as preparations are complete.<br \/>\n&nbsp;<\/p>\n<p><img data-recalc-dims=\"1\" loading=\"lazy\" decoding=\"async\" src=\"https:\/\/i0.wp.com\/mdx.jp\/wp-content\/uploads\/uploads\/mdx2\/2025\/01\/mdx2_GPU.png?resize=281%2C177&#038;ssl=1\" alt=\"\" width=\"281\" height=\"177\" class=\"alignnone size-full wp-image-4570\" \/><\/p>\n<p>With this addition, the existing 7 GPU nodes will be expanded by 8, bringing the total number of <strong>GPU nodes in mdx II to 15<\/strong>.<br \/>\nThe system configuration of the GPU nodes after expansion is shown below.<\/p>\n<h3>GPU Node Configuration from April 2026<\/h3>\n<div style=\"margin-left: 1em;\">\n<table border=\"1\" cellpadding=\"8\" cellspacing=\"0\">\n<tbody>\n<tr>\n<th>Total Computing Performance<\/th>\n<td colspan=\"3\"><span style=\"color:red; font-weight:bold;\">2,104.5 TFLOPS<\/span><\/td>\n<\/tr>\n<tr>\n<th rowspan=\"2\">Node Configuration<\/th>\n<td>GPU node cluster<br \/><span style=\"color:red; font-weight:bold;\">15 Nodes<\/span><\/td>\n<td colspan=\"2\">\n          &lt;per node&gt;<br \/>\n          \u30fbCPU: Intel Xeon Gold 6530 Processor(2.1 GHz, 32 cores) x 2<br \/>\n          \u30fbGPU: NVIDIA H200 SXM x 4<br \/>\n          \u30fbTheoretical Computing Performance: 140.300 TFLOPS<br \/>\n          \u30fbMain Memory Capacity: 1,024 GB<br \/>\n          \u30fbSecondary Storage Capacity: 960 GB\n        <\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>\u203bThe theoretical performance is calculated using double precision.<br \/>\n\u203bThe values represent the combined performance of both CPU and GPU. The breakdown is as follows:<\/p>\n<div style=\"margin-left: 1em;\">\n<table border=\"1\" style=\"border-collapse: collapse; text-align: center;\">\n<tr>\n<th><\/th>\n<th>Theoretical Computing Performance(per node)<\/th>\n<\/tr>\n<tr>\n<td><b>CPU<\/b><\/td>\n<td>4.3 TFLOPS<\/td>\n<\/tr>\n<tr>\n<td><b>GPU<\/b><\/td>\n<td>136.0 TFLOPS<\/td>\n<\/tr>\n<\/table>\n<\/div>\n","protected":false},"featured_media":0,"parent":0,"template":"","meta":{"_acf_changed":false,"advanced_seo_description":"","jetpack_seo_html_title":"","jetpack_seo_noindex":false},"mdx2_news-cat":[124860],"class_list":["post-7238","mdx2_news","type-mdx2_news","status-publish","hentry","mdx2_news-cat-news","en-US"],"acf":[],"jetpack_likes_enabled":true,"jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/mdx.jp\/wp-json\/wp\/v2\/mdx2_news\/7238","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/mdx.jp\/wp-json\/wp\/v2\/mdx2_news"}],"about":[{"href":"https:\/\/mdx.jp\/wp-json\/wp\/v2\/types\/mdx2_news"}],"version-history":[{"count":7,"href":"https:\/\/mdx.jp\/wp-json\/wp\/v2\/mdx2_news\/7238\/revisions"}],"predecessor-version":[{"id":9213,"href":"https:\/\/mdx.jp\/wp-json\/wp\/v2\/mdx2_news\/7238\/revisions\/9213"}],"wp:attachment":[{"href":"https:\/\/mdx.jp\/wp-json\/wp\/v2\/media?parent=7238"}],"wp:term":[{"taxonomy":"mdx2_news-cat","embeddable":true,"href":"https:\/\/mdx.jp\/wp-json\/wp\/v2\/mdx2_news-cat?post=7238"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}