Hi, I posted on here last month that a client of mine wished to benchmark Suse 9.1/SEL9 Vs Red Hat ES4. We ran 6 identical 64 bit machines load ballenced running a high volume website with Apache and Informix. The Suse boxes outstripped the Red Hat ones by 2-1; the response time is almost twice as quick and the CPU/MEM usage is much lower. Both boxes were exactly the same spec running linux built with the same packages and configs. I would have thought that they would have been almost exactly the same. Can anyone explain why there would be such a difference? I've always waved the Suse flag but this is ridiculous! Matthew

On Thursday 12 May 2005 01:32 pm, Allen wrote:

...

The Suse boxes outstripped the Red Hat ones by 2-1; the response time is almost twice as quick and the CPU/MEM usage is much lower.

What do you mean by that exactly? The way you've worded this you could mean either SUSE or RedHat was the fastest...

The sentence was quite unambiguous to me. What part did you find confusing? -- _____________________________________ John Andersen

On Sunday 15 May 2005 06:40, Anders Johansson wrote:

On Saturday 14 May 2005 21:58, John Andersen wrote:

...

On Thursday 12 May 2005 01:32 pm, Allen wrote:

...

...

The Suse boxes outstripped the Red Hat ones by 2-1; the response time is almost twice as quick and the CPU/MEM usage is much lower.

What do you mean by that exactly? The way you've worded this you could mean either SUSE or RedHat was the fastest...

The sentence was quite unambiguous to me. What part did you find confusing?

For me it was "the response time is almost twice as quick". Time is quick?? :) You silly boy Anders. Don't you watch TV advertisements? Surely you've heard claims like ".. gets bathrooms twice as clean" ? Smile Anders, smile. Regards, Colin

On Sunday 15 May 2005 06:19, Randall R Schulz wrote:

Presumably the author thought of "quick" as more or less synonymous with "fast." The response time was shorter. The system responded more quickly.

Yeah, it was kinda a joke :) Those alternative phrasings are both correct, but unless you are moving at relativistic speeds or attending a university lecture, a second is always a second. Time can't be faster or quicker, Apple movie formats notwithstanding.

Presumably.

The rest of the tone was also favorable to SuSE in the SuSE vs. RedHat comparison, so this interpretation is at least consistent.

Sure, I agree. I would just love to see the details of the test. If it's true it's an incredible marketing plus

On Sunday 15 May 2005 12:46, Matthew Stringer wrote:

...

The rest of the tone was also favorable to SuSE in the SuSE vs. RedHat comparison, so this interpretation is at least consistent.

Didn't think my English was that bad.

It wasn't. But you don't seem to want to answer my plea to see the details of your benchmark. That is bad

On Sunday 15 May 2005 23:20, Peter Van Lone wrote:

On 5/15/05, Matthew Stringer wrote:

...

Didn't think my English was that bad.

Please post your setup and methods and the results?

Or, at least tell us that you cannot due to some company proprietary issue, or some such!

Matthew did reply to me off-list to say that there was no special setup at all. Both the SLES9 and RHES4 setups were totally bog standard out-of-the-box setups, with no modifications or tuning made. Which makes the results even more extraordinary. If it's replicable it will be a tremendous plus when pushing SLES9

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi! Matthew Stringer wrote:

I posted on here last month that a client of mine wished to benchmark Suse 9.1/SEL9 Vs Red Hat ES4.

Finally! I was waiting for somebody to do such a comparison. Will you be able to publish the results and the methology you used?

We ran 6 identical 64 bit machines load ballenced running a high volume website with Apache and Informix.

The Suse boxes outstripped the Red Hat ones by 2-1; the response time is almost twice as quick and the CPU/MEM usage is much lower.

Good to hear :)

Both boxes were exactly the same spec running linux built with the same packages and configs.

That should be a given, if you want to do a fair comparison. Did you do any kernel tweaking? Or did you try to keep them as "out-of-the-box" as possible?

I would have thought that they would have been almost exactly the same. Can anyone explain why there would be such a difference?

I've always waved the Suse flag but this is ridiculous!

Both SUSE and Red Hat apply quite a number of patches to the Linux Kernel - none of them can really be considered a "vanilla" kernel. Take a look at the patches that are included in the Kernel source RPMs. It seems like SUSE did a better job of enhancing the Kernel. One assumption would be that they have made changes to the VM and scheduling. After all, Andrea Arcangeli, who is one of the kernel VM hackers, works for SUSE... I have always assumed that the SUSE kernel would outperform the one used by Red Hat. I would love to see some facts that actually proof this. Bye, LenZ - -- - ------------------------------------------------------------------ Lenz Grimmer -o) [ICQ: 160767607 | Jabber: LenZGr@jabber.org] /\\ http://www.lenzg.org/ V_V -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFChEcNSVDhKrJykfIRAtdoAJsGzDoJvdeCf37qlmR+WI7/e/e9uACfUrMU uiYxoO6OUtThf3RA8MM9eX8= =0S7g -----END PGP SIGNATURE-----

On Thu, May 12, 2005 at 05:54:35PM +0100, Matthew Stringer wrote:

Hi,

I posted on here last month that a client of mine wished to benchmark Suse 9.1/SEL9 Vs Red Hat ES4.

We ran 6 identical 64 bit machines load ballenced running a high volume website with Apache and Informix.

The Suse boxes outstripped the Red Hat ones by 2-1; the response time is almost twice as quick and the CPU/MEM usage is much lower.

Both boxes were exactly the same spec running linux built with the same packages and configs.

I would have thought that they would have been almost exactly the same. Can anyone explain why there would be such a difference?

I've always waved the Suse flag but this is ridiculous!

The part that caused me confusion was here. It was as if he was saying he always liked SUSE but this was ridiculous, like SUSE had done so bad although he'd always loved it. That's what had me thinking what he mean it ... Ah well I hadn't slept.

Matthew

-- Check the headers for your unsubscription address For additional commands send e-mail to suse-linux-e-help@suse.com Also check the archives at http://lists.suse.com Please read the FAQs: suse-linux-e-faq@suse.com