Collecting SBUs now

Michael Brömer mib at telebel.de
Thu May 30 12:31:53 PDT 2002


Am Donnerstag, 30. Mai 2002 20:56 schrieben Sie:
> On Thu, 30 May 2002 19:59:18 +0200
>
> Michael Brömer <mib at telebel.de> wrote:
> > I use MAKE="make -j2" most of the time. -j3 works fine too, but
> > it is a few seconds slower if i compile an entire LFS system.
> > Overall there is not much to gain by starting more processes than
> > you got CPUs if you got enough RAM.
>
> Thats not quite true.
>
> I find -j4 is moticably quicker on my UNI-processor box.
True, the machines can profit from having something to do
while doing I/O. But if you have enough RAM there is not that
much real disk "I/O" going on during compilation.

I while ago i experimented a bit and compiled LFS
systems with several settings for MAKE and there was not a
great deal of difference. 
                        -j1      -j2      -j3  ...  -j
            bash-2.05a  1:13     1:06     1:06     1:06
       binutils-2.11.2  2:03     2:03     2:01     2:02
           bzip2-1.0.1  0:06     0:05     0:05     0:06
         diffutils-2.7  0:07     0:07     0:07     0:07
         fileutils-4.1  0:55     0:46     0:45     0:45
            gcc-2.95.3  5:06     3:02     3:03     3:04
            grep-2.4.2  0:10     0:08     0:08     0:09
           gzip-1.2.4a  0:02     0:02     0:02     0:02
          linux-2.4.16  1:18     0:54     0:55     0:57
           make-3.79.1  0:14     0:12     0:11     0:12
            mawk-1.3.3  0:04     0:04     0:04     0:04
           patch-2.5.4  0:07     0:07     0:07     0:06
              sed-3.02  0:06     0:06     0:06     0:06
          sh-utils-2.0  0:37     0:30     0:30     0:29
              tar-1.13  0:18     0:15     0:16     0:14
           texinfo-4.0  0:21     0:15     0:15     0:17
         textutils-2.0  0:33     0:26     0:24     0:26
              time-1.7  0:06     0:04     0:04     0:04
-------------------------------------------------------
Total for BASE static  13:35    10:21    10:18    10:26

           glibc-2.2.4 17:51    10:41    10:46    11:57
...
...
-------------------------------------------------------
LFS done in m:s        57:49    42:41    42:44    failed

Usually gcc-2.95.3 failed with -j ;-(

I could never beat the time that i got with -j2, so
i stick with it.

Michael

-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message



More information about the lfs-dev mailing list