port make index (was: Re: make -j$n buildworld : use of -j investigated)
On Thu, 25 Nov 2004 16:19:02 +0900, Rob [EMAIL PROTECTED] wrote: Rob wrote: Brian Szymanski wrote: Did you try any machines that used Hyperthreading? I'd be interested to see how those machines fare based on the number of logical and real CPUs. Although people suggest -j4 as optimal in general case, I have come to a very different conclusion: 1) single CPU with enough RAM (2 GHz, 512 MB) there's no significant speed up in the range -j1 to -j9. So -j1 is as good as -j9. If you went to all that trouble, you might as well post the numbers :-) Time unit is minutes. CPU: 2x800 MHz2000 MHz333 MHz RAM: 1024 MB 512 MB 64 MB -j 199 50276 258 49291 358 50367 457 50547 558 49 658 50 757 50 858 50 958 50 I have run another test on a 700 MHz, 128 MB PC, and the following equation seems to hold for all my tests. Calculate: time(minutes) * speed(MHz) * nproc / 1000 MHz and if this results in approximately 1, the system is optimized. For example, in the above case, column 1: -j1 : 99 * 800 * 2 / 1000 = 1.5 -j2 : 58 * 800 * 2 / 1000 = 0.928 column 2: -j1 : 50 * 2000 * 1 / 1000 = 1 column 3: -j1 : 276 * 333 * 1 / 1000 = 0.919 another PC: -j1 : 142 * 700 * 1 / 1000 = 0.994 -- All PCs have standard hardware. Off-the-shelf mainboard, IDE harddisks, nothing special really. All this is done on 5.3-Stable systems and the time listed (in minutes) is for the buildworld only: make -jn buildworld Rob. Would all this work for 'make index' for the ports also? Or is this more io bound? I can't test this myself, because my laptop is to slow for making these tests any fun. Ronald. -- Ronald Klop, Amsterdam, The Netherlands ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: port make index (was: Re: make -j$n buildworld : use of -j investigated)
On Thu, Nov 25, 2004 at 10:09:35AM +0100, Ronald Klop wrote: Would all this work for 'make index' for the ports also? Or is this more io bound? I can't test this myself, because my laptop is to slow for making these tests any fun. Based on my tests, 'make index' is only faster with -j on smp machines (the default is -j2, which only penalizes UP by a second or two, but nearly halves the build time on an smp machine). Kris pgpO01V3Pxxuq.pgp Description: PGP signature
Re: port make index (was: Re: make -j$n buildworld : use of -j investigated)
On Thu, 25 Nov 2004 01:28:55 -0800, Kris Kennaway [EMAIL PROTECTED] wrote: On Thu, Nov 25, 2004 at 10:09:35AM +0100, Ronald Klop wrote: Would all this work for 'make index' for the ports also? Or is this more io bound? I can't test this myself, because my laptop is to slow for making these tests any fun. Based on my tests, 'make index' is only faster with -j on smp machines (the default is -j2, which only penalizes UP by a second or two, but nearly halves the build time on an smp machine). But -j1 saves on mem usage on UP I presume, so it makes my workstation more useable while being 2 secs.faster. Is this set by an INDEX_JOBS make var in /etc/make.conf? Found it in /usr/ports/Makefile. Ronald. -- Ronald Klop, Amsterdam, The Netherlands ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: port make index (was: Re: make -j$n buildworld : use of -j investigated)
On Thu, 25 Nov 2004 16:19:02 +0900, Rob [EMAIL PROTECTED] wrote: time(minutes) * speed(MHz) * nproc / 1000 MHz Looking at your examples, it seems you divide by 1e5, not by 1000. In other words, buildworld is CPU bound and takes about 6e12 clock cycles. Use -jnproc. Nick B ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: port make index (was: Re: make -j$n buildworld : use of -j investigated)
On Thu, Nov 25, 2004 at 11:14:42AM +0100, Ronald Klop wrote: On Thu, 25 Nov 2004 01:28:55 -0800, Kris Kennaway [EMAIL PROTECTED] wrote: On Thu, Nov 25, 2004 at 10:09:35AM +0100, Ronald Klop wrote: Would all this work for 'make index' for the ports also? Or is this more io bound? I can't test this myself, because my laptop is to slow for making these tests any fun. Based on my tests, 'make index' is only faster with -j on smp machines (the default is -j2, which only penalizes UP by a second or two, but nearly halves the build time on an smp machine). But -j1 saves on mem usage on UP I presume, so it makes my workstation more useable while being 2 secs.faster. Is this set by an INDEX_JOBS make var in /etc/make.conf? Found it in /usr/ports/Makefile. Yes. Kris pgp0Y9FFa4krZ.pgp Description: PGP signature