headers
Dmitri Nikulin | 9 Feb 2005 18:00
Picon

Re: benchmark

The apology would be because you accused Hubert of lying, which he 
simply could not even possibly have done from his role as a messenger. 
Do you call your browser a liar if you read something on Slashdot you 
don't like?

The paper was done very professionally. It gave clean facts with 
supporting visuals, left no ambiguity with regards to environment, and 
disclaimed its own validity as a measure of real-world performance. 
Hubert's exact wording of "NetBSD... beats FreeBSD" may seem like a 
hasty conclusion, but it's hard to argue that the benchmarks gave a lot 
to suppor that (say what you want, but on a UP i386 machine, NetBSD will 
run loops around FreeBSD 5.3, and non-i386 is hardly worth comparing - 
devs themselves admit FreeBSD 5 has lost a lot of performance ground in 
micro and macro).

Even if Hubert had done the benchmarks himself, even written them 
himself, and then drew conclusions on his own, his establishment as part 
of the NetBSD project (including his often overlooked but very useful 
blog, http://www.feyrer.de/NetBSD/blog.html) is a lot more than any 
credibility you can bring to the table. He's supporting a project which, 
by all rights, DESERVES to be on top because it would take the least 
amount of work to get it there (support a couple of new processors to 
catch up with Linux, add some fine-grained [but does it have to be 
Solaris-like? Isn't that what killed FreeBSD 5 to begin with?] SMP, and 
develop a more convenient update mechanism for pkgsrc, and suddenly 
NetBSD would be a world leader). I can sympathise with his efforts 
entirely. I do my part by speaking out (some call it trolling, that's up 
to them) on Slashdot, and I find a lot of people who agree.

You're quite right that BSDs should work together and not try to 'beat'
(Continue reading)

Permalink | Reply | 12 comments |
headers
Joel Macklow | 9 Feb 2005 20:38
Picon
Favicon

NetBSD 2 vs the rest with MySQL

Anyone seen this?

http://software.newsforge.com/article.pl?sid=04/12/27/1243207&from=rss

We do really well with the select-key test on a single CPU.  Any 
comments regarding the other tests from those who know? This is the kind 
of info the advocacy group needs - as long as we look good in it ;)
Permalink | Reply | 0 comments |
headers
Frank van der Linden | 9 Feb 2005 20:48
Picon

Re: NetBSD 2 vs the rest with MySQL

On Thu, Feb 10, 2005 at 08:38:36AM +1300, Joel Macklow wrote:
> We do really well with the select-key test on a single CPU.  Any 
> comments regarding the other tests from those who know? This is the kind 
> of info the advocacy group needs - as long as we look good in it ;)

Hm.. is this a threaded test? It looks like it. If so, the lack of
PTHREAD_CONCURRENCY > 1 explains the crappy 2-CPU results.

- Frank
Permalink | Reply | 0 comments |
headers
Jeff Rizzo | 9 Feb 2005 21:21
Favicon

Re: NetBSD 2 vs the rest with MySQL

Joel Macklow wrote:

> Anyone seen this?
>
> http://software.newsforge.com/article.pl?sid=04/12/27/1243207&from=rss
>
> We do really well with the select-key test on a single CPU.  Any 
> comments regarding the other tests from those who know? This is the 
> kind of info the advocacy group needs - as long as we look good in it ;)

I wonder if the crappy performance on the 10M row test would be affected 
by the tuning of vm.* sysctls, especially since Solaris problems on the 
same test were mitigated by forcedirectio...

Hm.

+j
Permalink | Reply | 0 comments |
headers
Hubert Feyrer | 9 Feb 2005 22:16
Picon
Favicon

10M row test in that MySQL benchmark


Re: http://software.newsforge.com/article.pl?sid=04/12/27/1243207&from=rss

I've talked to the author of the article, and he aparently really didn't 
set PTHREAD_CONCURRENCY (caused by it being not documented in NetBSD 
2.0); He wants to create some addendum, so we can look forward to an 
update.

A different question is, why NetBSD performed so bad in the 10M Rows test
(http://www.newsforge.com/blob.pl?id=82311a9e7896a841032c395f270d6a0f).
According to the article it's not a softdep or atime issue.

Any guesses?

  - Hubert

--

-- 
NetBSD - Free AND Open!      (And of course secure, portable, yadda yadda)
Permalink | Reply | 7 comments |
headers
Jason Thorpe | 9 Feb 2005 22:20

Re: NetBSD 2 vs the rest with MySQL


On Feb 9, 2005, at 11:48 AM, Frank van der Linden wrote:

> On Thu, Feb 10, 2005 at 08:38:36AM +1300, Joel Macklow wrote:
>> We do really well with the select-key test on a single CPU.  Any
>> comments regarding the other tests from those who know? This is the 
>> kind
>> of info the advocacy group needs - as long as we look good in it ;)
>
> Hm.. is this a threaded test? It looks like it. If so, the lack of
> PTHREAD_CONCURRENCY > 1 explains the crappy 2-CPU results.

I would definitely suggest to the author of the article to increase 
PTHREAD_CONCURRENCY for his tests and run them again.

-- thorpej
Permalink | Reply | 0 comments |
headers
Frank van der Linden | 9 Feb 2005 22:45
Picon

Re: NetBSD 2 vs the rest with MySQL

On Wed, Feb 09, 2005 at 01:20:38PM -0800, Jason Thorpe wrote:
> >Hm.. is this a threaded test? It looks like it. If so, the lack of
> >PTHREAD_CONCURRENCY > 1 explains the crappy 2-CPU results.
> 
> I would definitely suggest to the author of the article to increase 
> PTHREAD_CONCURRENCY for his tests and run them again.

Yeah, I dropped him a note. I'm curious how well it will perform
(if stable). I'm glad to see that we kick ass in the single-CPU
case.

- Frank
Permalink | Reply | 0 comments |
headers
Hubert Feyrer | 9 Feb 2005 22:46
Picon
Favicon

Re: NetBSD 2 vs the rest with MySQL

On Wed, 9 Feb 2005, Frank van der Linden wrote:
> (if stable). I'm glad to see that we kick ass in the single-CPU
> case.

Um, have you looked at the 10M Row test?
http://www.newsforge.com/blob.pl?id=82311a9e7896a841032c395f270d6a0f
Why so we not kick ass there at all?

  - Hubert

--

-- 
NetBSD - Free AND Open!      (And of course secure, portable, yadda yadda)
Permalink | Reply | 0 comments |
headers
Frank van der Linden | 9 Feb 2005 22:58
Picon

Re: NetBSD 2 vs the rest with MySQL

On Wed, Feb 09, 2005 at 10:46:30PM +0100, Hubert Feyrer wrote:
> Um, have you looked at the 10M Row test?
> http://www.newsforge.com/blob.pl?id=82311a9e7896a841032c395f270d6a0f
> Why so we not kick ass there at all?

I was talking about the first test, and the 1- to 2-CPU comparison,
which is interesting to look at if PTHREAD_CONCURRENCY=2 should work,
since we do so well in the 1-CPU case.

The I/O problem is a whole different matter, I don't know what's going
on there.

- Frank
Permalink | Reply | 0 comments |
headers
Jeff Rizzo | 9 Feb 2005 23:10
Favicon

Re: NetBSD 2 vs the rest with MySQL

Hubert Feyrer wrote:

> On Wed, 9 Feb 2005, Frank van der Linden wrote:
>
>> (if stable). I'm glad to see that we kick ass in the single-CPU
>> case.
>
>
> Um, have you looked at the 10M Row test?
> http://www.newsforge.com/blob.pl?id=82311a9e7896a841032c395f270d6a0f
> Why so we not kick ass there at all?
>
>
>  - Hubert
>

Well, I'm not really sure, since this is pretty far outside my area of 
expertise, but since he claims the Solaris numbers on the 10M row test 
were also horrid until turning off some FS cacheing, I wonder if tuning 
the VM system might help some.  I just sent the author a link to Arto 
Selonen's VM tuning page 
(http://www.selonen.org/arto/netbsd/vm_tune.html) in hopes that he might 
be able to squeeze some better numbers out of it...

Of course, someone who knows better than I should feel free to let me 
know if I'm full of it...

+j
(Continue reading)

Permalink | Reply | 0 comments |

Gmane