Re: Interesting scheduling times - NOT

Larry McVoy (lm@bitmover.com)
Thu, 24 Sep 1998 16:40:05 -0700


Oliver Xymoron <oxymoron@waste.org>:
: > : This is not really true, especially for something like make. If you
:
: > The claim was that cache interference would explain variations in run
: > times of 70-100%.
:
: I agree that this is not likely the proper explanation.

That's cool, that was all I wanted to establish.

: > My counter claim was that if that were true, things
: > like make would not be deterministic.
:
: But this is also not likely true. As long as the number of cache accesses
: (hits or misses) is relatively high, ie high enough that the cache gets
: turned over fairly frequently, the variances of individual cache access
: timings tend to cancel each other out.

Yes, that's true - and also holds true for the benchmark. I think what you
are saying here is that you shouldn't be seeing high variance in either
the make or the context switch 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/