Messages in this thread | | | From | (Larry McVoy) | Subject | Re: [PATCH] scheduler patch, faster still | Date | Tue, 29 Sep 1998 00:29:04 -0600 |
| |
Richard Gooch <rgooch@atnf.csiro.au>: : Go read: http://www.atnf.csiro.au/~rgooch/benchmarks/ and you'll see : that I don't claim the RT performance is hurt by some unknown : variance. I make the point that RT performance is hurt by long run : queues. This is something that I've managed to pin down quite well, : despite spurious claims (without foundation) to the contrary.
Err, Mr Gooch, what you have done is shown that a benchmark sort of sees some difference. Benchmarks are basically noise unless they can be shown to have correctly captured some application's behaviour. The core of the problem with your claims is that you have not shown that any application actually sees any benefit from your proposed changes.
So you are correct (sort of, leaving aside the variance issues) if you claim that your benchmarks see some effect from your changes. You're just guessing when you claim that translates to significant RT performance gains. Do you see the difference?
: Finally, I will again point out to all the naysayers out there, that : the change I propose: : : - adds very little extra code : - simplies existing code paths : - improves RT performance under all conditions : - improves non-RT performance under all conditions : - properly isolates RT processes from normal processes in the : scheduler : - reduces the scope for bugs in the scheduler code.
Mr Gooch, these are just unsubstantiated claims. That's not science or engineering. I could say
- the sky is yellow with black polka dots - the world is flat - squares are round
and I would be no more correct. I'm sorry to seen as a naysayer, but as long as you are making claims that seem to have no basis in fact, I'm going to stick around and be one of the people that insists on data which supports your claims.
- 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/
| |