Messages in this thread | | | Date | Sun, 25 Apr 1999 22:34:51 -0700 (PDT) | From | Linus Torvalds <> | Subject | Re: bad lmbench numbers for mmap |
| |
On Mon, 26 Apr 1999, Chris Wedgwood wrote:
> > IIRC, Larry caught Solaris cheating and not actually mapping the > > pages at all until the first access. > > Surely in many cases this is a smart thing to do? Has anyone actually > profiled code to see if its better one way or the other? Lazy setup > doesn't seem an entirely stupid idea to me...
Oh, it wasn't Larry who caught anybody cheating, it was us who caught ourself just getting ridiculous numbers. It was very clear that without touching the pagces some way the benchmark was worthless - is basically opened itself up to being optimized away. I was the one (or at least one of the ones) that asked for the benchmark to be changed for that reason (it does have my name in it now).
It' sjust that when I say "touch", I don't mean "write"..
Linus
- 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/
| |