From owner-freebsd-current Tue Jul 2 15:17:55 2002 Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.FreeBSD.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 462B937B40C for ; Tue, 2 Jul 2002 15:17:26 -0700 (PDT) Received: from apollo.backplane.com (apollo.backplane.com [216.240.41.2]) by mx1.FreeBSD.org (Postfix) with ESMTP id 03F0D44001 for ; Tue, 2 Jul 2002 15:08:58 -0700 (PDT) (envelope-from dillon@apollo.backplane.com) Received: from apollo.backplane.com (localhost [127.0.0.1]) by apollo.backplane.com (8.12.4/8.12.3) with ESMTP id g62M54T4007849 for ; Tue, 2 Jul 2002 15:05:04 -0700 (PDT) (envelope-from dillon@apollo.backplane.com) Received: (from dillon@localhost) by apollo.backplane.com (8.12.4/8.12.3/Submit) id g62M54MF007848; Tue, 2 Jul 2002 15:05:04 -0700 (PDT) (envelope-from dillon) Date: Tue, 2 Jul 2002 15:05:04 -0700 (PDT) From: Matthew Dillon Message-Id: <200207022205.g62M54MF007848@apollo.backplane.com> To: FreeBSD current users Subject: Final test numbers for buildworld under stable, current / with and without softupdates (was Re: something funny with soft updates?) References: <200206290440.g5T4e6EU009321@apollo.backplane.com> Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG I have some new numbers. I finally was able to run the test on -current with an SMP build. All the results are below. It seems to confirm my hypothesis that the new cpu-hungry gcc is the main source of the timing differences. -Matt Matthew Dillon Things to note: * sys time is an approximate indicator of I/O and matches-up fairly well between -stable and -current, indicating that the builds on -current and -stable are doing approximately the same amount of I/O. * differences in sys times for SMP vs UP tests are a fairly good indicator of reduced efficiency under SMP, a metric we may be able to use in future tests. * user times are consistent comparing SMP & UP under current and a good indication that -current is timing the processes properly. * massive differences in user times between -current and -stable is a fair indication that the new GCC on -current is the cpu-cookie-monster culprit and also responsible for masking I/O improvements from softupdates (which remains relatively fixed between -stable and -current tests). There's nothing wrong with softupdates. #!/bin/tcsh -f # # (all tests run on a DELL2550 2xCPUx/1.1GHz P3s) cd /usr/src mv ~dillon/bwtest.out ~dillon/bwtest.out.bak ( /usr/bin/time -l make -j 5 buildworld ) >>& ~dillon/bwtest.out ( /usr/bin/time -l make -j 5 buildworld ) >>& ~dillon/bwtest.out ( /usr/bin/time -l make -j 5 buildworld ) >>& ~dillon/bwtest.out current, without softupdates (SMP BUILD): 2760.07 real 2514.37 user 777.96 sys 2705.11 real 2507.46 user 784.09 sys 2697.63 real 2513.73 user 778.65 sys current, with softupdates (SMP BUILD): 2358.11 real 2511.09 user 777.87 sys 2330.06 real 2512.97 user 781.82 sys 2335.10 real 2516.46 user 781.29 sys current, without softupdates (UP BUILD): 3361.70 real 2365.23 user 535.50 sys 3451.55 real 2368.22 user 537.26 sys 3454.85 real 2369.42 user 536.69 sys current, with softupdates (UP BUILD): 3122.30 real 2360.70 user 532.54 sys 3083.17 real 2361.14 user 529.53 sys 3085.05 real 2361.59 user 529.32 sys FOR REFERENCE (builds on -stable) : (original tests under -stable) :test1# cat x1 (SMP BUILD, STABLE, WITH SOFTUPDATES) : 1497.09 real 1397.98 user 612.06 sys : 1500.12 real 1399.33 user 609.79 sys : 1494.82 real 1398.30 user 612.46 sys :test1# cat x2 (SMP BUILD, STABLE, WITHOUT SOFTUPDATES) : 2449.14 real 1401.34 user 625.54 sys : 2389.75 real 1400.38 user 629.86 sys : 2358.82 real 1403.26 user 624.93 sys : To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message