From owner-freebsd-current Thu Jul 9 23:34:20 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id XAA28761 for freebsd-current-outgoing; Thu, 9 Jul 1998 23:34:20 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.15.68.22]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id XAA28756 for ; Thu, 9 Jul 1998 23:34:17 -0700 (PDT) (envelope-from bde@godzilla.zeta.org.au) Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.7/8.8.7) id QAA09859; Fri, 10 Jul 1998 16:34:13 +1000 Date: Fri, 10 Jul 1998 16:34:13 +1000 From: Bruce Evans Message-Id: <199807100634.QAA09859@godzilla.zeta.org.au> To: current@FreeBSD.ORG, jdp@polstra.com Subject: Re: Does "make -j4 buildworld" actually help anybody? Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG >I recently got what I thought would be a perfect box to take advantage >of "make -j4": a PII/400 with 128 MB of RAM. But "-j4" slows things >down considerably. Under identical conditions, the timings were: I think it only helps with multiple CPUs and enough disk parallelism to prevent too many extra seeks. >Without -j4: > real 47m14.643s > user 36m7.531s > sys 12m20.942s It is hard to improve on a negative amount of time waiting for the disks :-). I still haven't committed a fix for fork time getting counted twice. Here is a cryptic version: diff -c2 kern_resource.c~ kern_resource.c *** kern_resource.c~ Fri May 29 13:02:39 1998 --- kern_resource.c Fri Jul 10 16:26:57 1998 *************** *** 523,526 **** --- 523,527 ---- */ microuptime(&tv); + switchtime = tv; /* XXX */ totusec += (tv.tv_usec - p->p_switchtime.tv_usec) + (tv.tv_sec - p->p_switchtime.tv_sec) * (int64_t)1000000; Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message