<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Memory Performance of AMD EPYC Rome and Intel Cascade Lake SP Server Processors · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2204.03290/</link><description>Collected HN comments and submissions linking this paper.</description><language>en</language><lastBuildDate>Mon, 02 Nov 2026 11:16:22 GMT</lastBuildDate><atom:link href="https://hackerpaper.kaipfister.de/p/2204.03290/feed.xml" rel="self" type="application/rss+xml" /><item><title>Memory Performance of AMD EPYC Rome and Intel Cascade Lake SP Server Processors — mentioned by magicalhippo</title><link>https://news.ycombinator.com/item?id=41888296</link><guid isPermaLink="false">urn:hackerpapers:mention:2204.03290:41888296</guid><pubDate>Sat, 19 Oct 2024 15:16:50 GMT</pubDate><description>&lt;p&gt;Curious about the latency of this. This[1] page, which is from 2022, seems to indicate that the CLX memory controllers adds about 200ns, and you got some retimers which add a few tens of ns. Then there&amp;#x27;s the cabling, assuming 2/3 the speed of light that&amp;#x27;s 5ns per meter. With the memory latency itself it seems this adds up about 350ns, while a local memory hit is about 100ns[2][3]. For random access workloads this is certainly better than not having stuff in memory, but I guess for streaming workloads a sled full of NVMe drives might be just as good. [1]…&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2204.03290/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.AR</category></item><item><title>Memory Performance of AMD EPYC Rome and Intel Cascade Lake SP Server Processors — HN submission</title><link>https://news.ycombinator.com/item?id=31328410</link><guid isPermaLink="false">urn:hackerpapers:submission:2204.03290:31328410</guid><pubDate>Tue, 10 May 2022 15:39:16 GMT</pubDate><description>&lt;p&gt;Memory Performance of AMD EPYC Rome and Intel Cascade Lake SP Server Processors&lt;/p&gt;&lt;p&gt;&lt;/p&gt;&lt;p&gt;3 points; 0 discussion replies on HN. &lt;a href="https://hackerpaper.kaipfister.de/p/2204.03290/"&gt;Paper and source history&lt;/a&gt;. A submission is not an endorsement.&lt;/p&gt;</description><category>cs.AR</category></item></channel></rss>
