<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Apple vs. Oranges: Evaluating the Apple Silicon M-Series SoCs for HPC Performance and Efficiency · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2502.05317/</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/2502.05317/feed.xml" rel="self" type="application/rss+xml" /><item><title>Apple vs. Oranges: Evaluating the Apple Silicon M-Series SoCs for HPC Performance and Efficiency — mentioned by fvwqcecvq</title><link>https://news.ycombinator.com/item?id=48902154</link><guid isPermaLink="false">urn:hackerpapers:mention:2502.05317:48902154</guid><pubDate>Tue, 14 Jul 2026 04:00:42 GMT</pubDate><description>&lt;p&gt;No. It&amp;#x27;s always been about power efficiency. Apple&amp;#x27;s GPU are very good in that area just like the CPU. https://arxiv.org/html/2502.05317v1 Apple vs. Oranges: Evaluating the Apple Silicon M-Series SoCs for HPC Performance and Efficiency &amp;quot;Apple&amp;#x27;s M-series GPUs offer massive performance-per-watt, scaling efficiently from roughly 5W in base chips to around 40-50W in top-tier Max chips. This efficiency generally sits between 200 and 250 GFLOPS per Watt.&amp;quot; Sure. nVidia GPUs have higher performance. But they do it with 450W+ a.k.a. 10x the power&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2502.05317/"&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>Apple vs. Oranges: Evaluating the Apple Silicon M-Series SoCs for HPC Performance and Efficiency — mentioned by woadwarrior01</title><link>https://news.ycombinator.com/item?id=47585721</link><guid isPermaLink="false">urn:hackerpapers:mention:2502.05317:47585721</guid><pubDate>Tue, 31 Mar 2026 11:19:00 GMT</pubDate><description>&lt;p&gt;&amp;gt; Sorry to shatter your bubble, but this is patently false, LLMs are far more efficient on hardware that simultaneously serves many requests at once. You might want to read this: https://arxiv.org/abs/2502.05317v2&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2502.05317/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.AR</category></item></channel></rss>
