Re: 2.2.10-ac4 / knfsd 1.4.1 poor parallel reading performance

Thierry Danis (danis@sagem.fr)
Mon, 28 Jun 1999 10:09:15 +0200


On Fri, Jun 25, 1999 at 02:35:50PM -0700, Thomas Davis wrote:
> Thierry Danis wrote:
> >
> > Any idea ?
>
> How many nfsd threads?
>
> did you try changing it to say, 24? 32?
>
> --
> ------------------------+--------------------------------------------------
> Thomas Davis | PDSF Project Leader
> tadavis@lbl.gov |
> (510) 486-4524 | "Only a petabyte of data this year?"
>

I have tried with 1, 8, 16 and 32 threads, and it has exactly
the same bad rate (~ 1,7 Mo/s with 2 clients). HJ L gave a
try and he reported really crazy numbers. I guess it has
so much memory that most of the disk accesses are avoided.

In fact, I am quite surprised that nobody else is able
to reproduce it. Except the fact that I could not apply
the whole patch coming with knfsd 1.4.1 (as a result, I
did not apply it at all) because of conflicts with 2.2.10-ac4,
I think I have not done anything fancy.

A+,

-- 
	Thierry Danis
	Poste : 53 53	danis@spmo.sagem.fr

- 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/