Date: Thu, 09 Nov 2023 18:08:53 +0000 From: bugzilla-noreply@freebsd.org To: bugs@FreeBSD.org Subject: [Bug 274994] Regression of iperf3 network throughput tests with erms "rep movsb" copyto loops Message-ID: <bug-274994-227-JTB4aAKZhT@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-274994-227@https.bugs.freebsd.org/bugzilla/> References: <bug-274994-227@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D274994 --- Comment #1 from Richard Scheffenegger <rscheff@freebsd.org> --- (Just to make it clear, there were essentially loopback tests within the sa= me system, and the absolute values matter less than the relative comparison between the tests with the erms rep movsb vs. hand-rolled loops. Even on the one platform which is most recent, the hand-rolled loop was still appreciable faster over the rep movsb in this test... Since this is rather unexpected, an independent confirmation of these relat= ive results on a plain hardware platform running MAIN with and without the patch to effectively collapse copyin to the nosmap_std routing would be highly=20 appreciated. --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-274994-227-JTB4aAKZhT>