lkml.org 
[lkml]   [1998]   [Sep]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: Interesting scheduling times - NOT
Date
"A month of sundays ago Richard Gooch wrote:"
>
> As I've already said, you're probably not seeing the variance because
> you don't run with RT priority. If I run my test with SCHED_OTHER then


Naively (and that's my chief virtue :-) I see

a) Larry's tests show 100% variance between _different_ runs,
each run with different numbers of background processes.

b) Richard's tests show 100% variance during _one_ test run, and
he sees unpredicatbly varying numbers of processes on the run queue
during the run. Variance drops when he tries to stabilize
those numbers.

and I would venture to suggest that's about 70% of the cause. I.e. a hidden
variable. I'd like to see the full spectra of results and do my own stats
on it. Why are you all bleating about min/median when you have the full
distribution available? I don't know the dist of a min offhand, but
it strikes me as being intrinsically highly variable on a normal dist.


Peter

> It was not uncommon to start the benchmark with 2 processes on the run
> queue and finish with 10.
> With the settling-down delay, I'm now finishing the benchmark with 2
> or 3 processes on the run queue (rarely 4).
> This has made the variance in the median come down to 10% in the case


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:44    [W:0.070 / U:1.692 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site