LWN: Comments on "User-data replication on NUMA systems" https://lwn.net/Articles/45334/ This is a special feed containing comments posted to the individual LWN article titled "User-data replication on NUMA systems". en-us Wed, 01 Oct 2025 01:21:22 +0000 Wed, 01 Oct 2025 01:21:22 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net User data Page replication https://lwn.net/Articles/47328/ https://lwn.net/Articles/47328/ balbir-_singh I think only executable text pages are replicated. As far as <br>I can remember data is never replicated (I worked on such systems <br>about two years ago).<p>The title can be misleading if what I am saying is correct.<p>Balbir Tue, 02 Sep 2003 10:32:24 +0000 Numbers are off https://lwn.net/Articles/45661/ https://lwn.net/Articles/45661/ corbet I read the numbers right, I just wrote poorly. It was getting late in a long week... In any case, you're right; the new performance is as good as 143% of the old value, the improvement is 43%. Thu, 21 Aug 2003 15:52:09 +0000 Numbers are off https://lwn.net/Articles/45651/ https://lwn.net/Articles/45651/ pflugstad I think you're looking at the numbers wrong. The posting has:<p>SDET Average Throughput (NUMA-Q): <br> 2.6.0-test3 100.0%<br> 2.6.0-test3+urepl 143.1%<p>SDET Average Throughput (16-way P4): <br> 2.6.0-test3 100.0%<br> 2.6.0-test3+urepl 108.8%<p>The baseline is 100%, so that's an *improvement* of 43.1% or 8.8%, not 143.2/108.8. <p>While these are good numbers, one wonders how the &quot;production ready&quot; patch would change them. <p> Thu, 21 Aug 2003 15:19:31 +0000