Date: Sun, 16 Mar 2003 13:41:24 -0800 From: Brooks Davis <brooks@one-eyed-alien.net> To: Andy Isaacson <adi@hexapodia.org> Cc: Brooks Davis <brooks@one-eyed-alien.net>, smp@freebsd.org Subject: Re: weird power usage on dual Xeon Message-ID: <20030316134124.B4374@Odin.AC.HMC.Edu> In-Reply-To: <20030315010321.A12293@hexapodia.org>; from adi@hexapodia.org on Sat, Mar 15, 2003 at 01:03:21AM -0600 References: <20030314174113.A16583@Odin.AC.HMC.Edu> <20030315010321.A12293@hexapodia.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--8P1HSweYDcXXzwPJ Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Sat, Mar 15, 2003 at 01:03:21AM -0600, Andy Isaacson wrote: >=20 > Do you get a different answer if you run only two copies of SETI? I > can't imagine that their FFT routine takes kindly to having to share L1 > cache, perhaps it's spending all its time waiting for the L2... and do > you actually process more work units per hour with 4 SETI processes > running on HT as opposed to 2 SETIs without HT? I kinda doubt it, but > would welcome real information. Given the way our schedular handles HTT (it doesn't) you're just as likely to end up with two SETI procs on one CPU as with them on different CPUs. In any case, SETI actually gets fairly significantly better results when you run four with HTT enabled then when you run two without. With two I'm seeing about 4hr per workunit or one every 2hrs, with four I'm seeing about 6hrs per workunit or one every 1.5hrs. -- Brooks --=20 Any statement of the form "X is the one, true Y" is FALSE. PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4 --8P1HSweYDcXXzwPJ Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQE+dO+DXY6L6fI4GtQRAsDsAJ94QfOL5w2oJIQPvAoKIMG+4ypeQACglU/3 AIoX+3v+oYSvlsB6w03QOXE= =nKy9 -----END PGP SIGNATURE----- --8P1HSweYDcXXzwPJ-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030316134124.B4374>