{"id":10756,"date":"2026-02-26T07:02:22","date_gmt":"2026-02-26T07:02:22","guid":{"rendered":"https:\/\/mailitics.com\/index.php\/2026\/02\/26\/breaking-the-host-memory-bottleneck\/"},"modified":"2026-02-26T07:02:22","modified_gmt":"2026-02-26T07:02:22","slug":"breaking-the-host-memory-bottleneck","status":"publish","type":"post","link":"https:\/\/mailitics.com\/index.php\/2026\/02\/26\/breaking-the-host-memory-bottleneck\/","title":{"rendered":"Breaking the Host Memory Bottleneck: How Peer Direct Transformed Gaudi\u2019s Cloud Performance"},"content":{"rendered":"<p>    Breaking the Host Memory Bottleneck: How Peer Direct Transformed Gaudi\u2019s Cloud Performance<br \/>\n \t<BR><br \/>\n<BR><\/BR><br \/>\n    <!-- no image --><br \/>\n \t<BR><br \/>\n<BR><\/BR><\/p>\n<div>\n<p>Engineering RDMA-like performance over cloud host NICs using libfabric, DMA-BUF, and HCCL to restore distributed training scalability <\/p>\n<p>The post <a href=\"https:\/\/towardsdatascience.com\/breaking-the-host-memory-bottleneck\/\">Breaking the Host Memory Bottleneck: How Peer Direct Transformed Gaudi\u2019s Cloud Performance<\/a> appeared first on <a href=\"https:\/\/towardsdatascience.com\/\">Towards Data Science<\/a>.<\/p>\n<\/div>\n<p> \t<BR><br \/>\n <BR><\/BR><br \/>\n    Maria Piterberg<br \/>\n \t<BR><br \/>\n<BR><\/BR><br \/>\n<a href=\"https:\/\/towardsdatascience.com\/breaking-the-host-memory-bottleneck\/\">Go to original source<\/a><br \/>\n \t<BR><br \/>\n <BR><\/BR><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Breaking the Host Memory Bottleneck: How Peer Direct Transformed Gaudi\u2019s Cloud Performance Engineering RDMA-like performance over cloud host NICs using libfabric, DMA-BUF, and HCCL to restore distributed training scalability The post Breaking the Host Memory Bottleneck: How Peer Direct Transformed Gaudi\u2019s Cloud Performance appeared first on Towards Data Science. Maria Piterberg Go to original source<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2804,62,69,240,4836,4837,222],"tags":[1716,4771,1194],"class_list":["post-10756","post","type-post","status-publish","format-standard","hentry","category-ai-engineering","category-aimldsaimlds","category-artificial-intelligence","category-editors-pick","category-gaudi","category-hccl","category-mlops","tag-cloud","tag-host","tag-performance"],"_links":{"self":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/posts\/10756"}],"collection":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/comments?post=10756"}],"version-history":[{"count":0,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/posts\/10756\/revisions"}],"wp:attachment":[{"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/media?parent=10756"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/categories?post=10756"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/mailitics.com\/index.php\/wp-json\/wp\/v2\/tags?post=10756"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}