2002-07-09 09:06:26

by Anders Fugmann

[permalink] [raw]
Subject: Chatserver workload simulator by Bill Hartner?

Hi.

I'm looking for the chatserver workload simulator made by Bill hartner,
which was used to test the O(1) scheduler by Ingo Molnar.

Does anyone know where to find it? - All I can find is the VolanoMark,
but I guess that this is not the one used, since the command used by
Ingo Molnar when benchmarking the O(1) scheduler is: './chat_c 127.0.0.1
10 100'.

Thanks in advance.
Anders Fugmann


2002-07-09 16:25:47

by Dave Hansen

[permalink] [raw]
Subject: Re: Chatserver workload simulator by Bill Hartner?

Anders Fugmann wrote:
> I'm looking for the chatserver workload simulator made by Bill hartner,
> which was used to test the O(1) scheduler by Ingo Molnar.
>
> Does anyone know where to find it? - All I can find is the VolanoMark,
> but I guess that this is not the one used, since the command used by
> Ingo Molnar when benchmarking the O(1) scheduler is: './chat_c 127.0.0.1
> 10 100'.

Volanomark is probably the one, although I didn't know that Bill
Hartner made it. It is one of the benchmarks that we use quite a bit.

BTW, why didn't you ask Bill Hartner?

--
Dave Hansen
[email protected]

2002-07-09 16:32:32

by Perches, Joe

[permalink] [raw]
Subject: RE: Chatserver workload simulator by Bill Hartner?

It's on the Linux Benchmark Suite site with several
other useful tools...

http://lbs.sourceforge.net/

2002-07-09 18:08:14

by Anders Fugmann

[permalink] [raw]
Subject: Re: Chatserver workload simulator by Bill Hartner?

Many thanks. Just what I was looking for.
(Strange that this did not show up on my google searches)

Regards
Anders Fugmann

Perches, Joe wrote:
> It's on the Linux Benchmark Suite site with several
> other useful tools...
>
> http://lbs.sourceforge.net/
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/



2002-07-10 20:15:26

by James Cleverdon

[permalink] [raw]
Subject: Re: Chatserver workload simulator by Bill Hartner?

On Tuesday 09 July 2002 09:35 am, Perches, Joe wrote:
> It's on the Linux Benchmark Suite site with several
> other useful tools...
>
> http://lbs.sourceforge.net/
> -

The chat-1.0.1.tar.gz on that page still has a memory free/use bug with the ti
array. I sent the patch to bhartner, but maybe he's not maintaining it
anymore. It only seems to cause trouble when running heavy loads. (Maybe
large blocks of memory get coalesced, or something.) Anyway, here it is:

--- chat_s.c.df Tue Jun 4 17:37:03 2002
+++ chat_s.c Thu Jun 20 15:18:51 2002
@@ -515,7 +515,6 @@
int exit_rc = 0;
int rc = 0;

- free(ti);
free(s_send_stack);
free(s_receive_stack);

@@ -533,6 +532,8 @@
}
}

+ free(ti);
+
return(exit_rc);
}



--
James Cleverdon
IBM xSeries Linux Solutions
{jamesclv(Unix, preferred), cleverdj(Notes)} at us dot ibm dot com

2002-07-10 20:32:53

by Anders Fugmann

[permalink] [raw]
Subject: Re: Chatserver workload simulator by Bill Hartner?

Thanks - You just saved me a couple of hours there.

I've also located a showstopper bug in the chat server code.
If anyone steps forward as the maintainer, I will gladly send a patch.

The bug is when using accept, the last argument should be the size of the 2. argument but
it is zero in the chat server.

Regards
Anders Fugmann

James Cleverdon wrote:
> On Tuesday 09 July 2002 09:35 am, Perches, Joe wrote:
>
> The chat-1.0.1.tar.gz on that page still has a memory free/use bug with the ti
> array. I sent the patch to bhartner, but maybe he's not maintaining it
> anymore. It only seems to cause trouble when running heavy loads. (Maybe
> large blocks of memory get coalesced, or something.) Anyway, here it is:
<patch cut out>

2002-07-11 04:04:21

by Rusty Russell

[permalink] [raw]
Subject: Re: Chatserver workload simulator by Bill Hartner?

On Tue, 09 Jul 2002 09:27:56 -0700
Dave Hansen <[email protected]> wrote:

> Anders Fugmann wrote:
> > I'm looking for the chatserver workload simulator made by Bill hartner,
> > which was used to test the O(1) scheduler by Ingo Molnar.
> >
> > Does anyone know where to find it? - All I can find is the VolanoMark,
> > but I guess that this is not the one used, since the command used by
> > Ingo Molnar when benchmarking the O(1) scheduler is: './chat_c 127.0.0.1
> > 10 100'.
>
> Volanomark is probably the one, although I didn't know that Bill
> Hartner made it. It is one of the benchmarks that we use quite a bit.
>
> BTW, why didn't you ask Bill Hartner?

Hmm, I wrote an even more cut-down one called hackbench (I really should have
called it schedbench). Basically, if you can do well on this, you should
do well on Volanomark, since at the core it's the same scheduler-stressing
load. Look for the linearity of "hackbench 10" to "hackbench 100".

Google is your friend,
Rusty.
--
there are those who do and those who hang on and you don't see too
many doers quoting their contemporaries. -- Larry McVoy