Some thoughts about our tests and the build time

[Bjoern Michaelsen]

Hi,

On Tue, May 17, 2016 at 04:25:35PM +0200, Markus Mohrhard wrote:
> > Did you by chance have an opportunity to also ran "make" from scratch on
> > that machine? I assume/hope it wont take the 70 min + 11 min = 82 min. a
> > straight addition would make one assume, because make will use idle job slots
> > during a full build.
> real    83m37.618s
> user    493m30.052s
> sys    29m44.404s

Ah ok, so these add up pretty linear in both user and real.

> And then the following time make (therefore again only executing tests):
> 
> real    9m16.741s
> user    45m31.208s
> sys    0m37.516s

Thats interesting as it should be comparable with:

> > > real    6m37.479s
> > > user    45m4.740s

and while it is for CPU time it isnt for wall time by a factor of 50%. So seems
we have quite some variance here ...

> > P.S.: For reference, the output of "time make build-nocheck" would be
> > helpful too (aka a noop incremental build time overhead in make/dep parsing
> > etc.)
> real    0m38.787s
> user    0m16.300s
> sys    0m0.896s

Expected I guess -- and it seems heavily I/O bound, this is single threaded
and the CPU is waiting 2/3rds of the time.

Thanks for the additional numbers!

Best,

Bjoern