Messages in this thread | | | From | "Peter T. Breuer" <> | Subject | Re: Interesting scheduling times - NOT | Date | Thu, 24 Sep 1998 14:25:13 +0200 (MET DST) |
| |
"A month of sundays ago Oliver Xymoron wrote:" > > > to say "time make" and expect to get anything like the same number two times > > in a row, yet people do that all the time. > > This is not really true, especially for something like make. If you > combine n events of average length t and std dev u, you get an event of > average length n*t, but a smaller std dev - if you multiply a bunch of > bell curves together, you get a tighter curve. The possible range becomes > larger, sure. > > Not that you would really expect most of events you'd benchmark on a > computer to have a very bell-shaped distribution.
It doesn't matter. The result you quote is true for any distribution with bounded mean and variance. It's one of the basic theorems in probability. I.e. Larry's argument is wrong here. Well spotted.
I get the impression that lots of people are shaky on statistics. So am I - but I am a strong mathematician and I can guarantee to detect a wrong mathematical argument :-).
I haven't seen anything to convince me that Richards benchmark is not simply measuring a bi-distributed variable with number of running processes as a hidden variable. That makes it perfectly valid. We can get the two distributions out of that fairly simply. I don't know what all this nonsense is about measuring medians and minimums.
Larry published the test results he got, and they were clearly from a single distribution. I haven't seen that same basic data from Richard. That's all that's required to extract the statistical information and stop the argument (sic).
Peter
- 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/
| |