DragonFly BSD
DragonFly users List (threaded) for 2009-11
[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]

Re: Analysis on make parallelism for buildworld


From: Saifi Khan <saifi.khan@xxxxxxxxxxxxxxx>
Date: Thu, 5 Nov 2009 21:54:00 +0530 (IST)

On Thu, 5 Nov 2009, Simon 'corecode' Schubert wrote:

> Saifi Khan wrote:
> > Please review and let me know if there is any tweak/correction
> > that you deem necessary.
> 
> Hey!
> 
> that is interesting.  Could you run it with 5 jobs as well so that we can see
> whether how much difference this makes?
> 
> Also, are those data points multiple runs or just one?
> 

The plot was generated from the results collected so far.

# more ./timelog
building with nice=0, j=1, run=1
real 5254.18
user 4335.86
sys 965.83
building with nice=0, j=2, run=1
real 5047.78
user 4343.25
sys 979.63
building with nice=0, j=3, run=1
real 3668.48
user 4439.19
sys 1044.94
building with nice=0, j=4, run=1
real 3535.87
user 4459.59
sys 1061.05
building with nice=0, j=5, run=1

Yes, the script is currently running with 5 jobs for the first run.

Once the script runs to completion it would have had 5 runs and 5 jobs per run.

i'll generate a plot and post it out, once the 5th job for the first run is through.


thanks
Saifi.




[Date Prev][Date Next]  [Thread Prev][Thread Next]  [Date Index][Thread Index]