irt: tools/squashfs4: enable parallel builds

Elliott Mitchell ehem+openwrt at m5p.com
Sun Apr 2 15:28:51 PDT 2023


Am I the only person who finds commit 3efd49a588
("tools/squashfs4: enable parallel builds") a bit odd?  Specifically from
the commit message:

> Run `time make tools/squashfs4/{clean,compile} -j$(nproc)`
> Before:
> real    0m8.803s
> user    0m12.415s
> sys     0m1.317s
> After:
> real    0m13.781s
> user    0m13.290s
> sys     0m1.528s

So a parallel build of squashfs4 took more processor time in user mode,
more processor time in supervisor mode, and *more* wall clock time?

More processor time is fine, that comes with coordinating parallel build
processes.  The goal of parallel builds is to spend a bit of processor
time to *reduce* wall clock time use.  Yet a parallel build of
`squashfs4` is stated to have resulted in a 50% increase?

Uhm.  Perhaps your computer's processor interconnect is very slow?
Does this computer have an extremely high processor count?  I do note
`squashfs4` doesn't have very many source files.

Generally I would support the use of parallel builds, but unless I'm
missing something this appears the case where it should be avoided.


-- 
(\___(\___(\______          --=> 8-) EHM <=--          ______/)___/)___/)
 \BS (    |         ehem+sigmsg at m5p.com  PGP 87145445         |    )   /
  \_CS\   |  _____  -O #include <stddisclaimer.h> O-   _____  |   /  _/
8A19\___\_|_/58D2 7E3D DDF4 7BA6 <-PGP-> 41D1 B375 37D0 8714\_|_/___/5445





More information about the openwrt-devel mailing list