Peer review of AMD64/FreeBSD article

[Peter Wemm]

On Friday 12 March 2004 04:08 am, Jem Matzan wrote:
> I've just finished writing this article comparing performance between
> an Athlon64 in 32-bit and 64-bit mode using FreeBSD:
>
> http://www.thejemreport.com/lab64/amd64vsi386.php
>
> (this is a temporary address which will later redirect to the
> published article)
>
> I've checked it over twice for fact accuracy, but I would like other
> eyes to look at it before it goes to press. I haven't spell-checked
> it yet, so don't worry about that... I just want to make sure I
> haven't made any factual errors.

There's a couple of quick comments I'd like to make.

The OpenSSL stuff in the FreeBSD source tree has hand-optimized i386 
assembler code in it.  The amd64 configuration of the integrated 
OpenSSL uses the generic C code.  Given this, I wouldn't expect them to 
be even close.

In the Stream section, you mention that the P4 has a faster 
front-side-bus..  Faster than what?  The AMD64 system doesn't have a 
frontside bus inbetween the cpu and memory.  The big difference though 
is that your AMD64 machine has a single 3.2GB/sec max memory 
controller, while the P4 has a dual-channel memory controller.  In 
other words, the AMD64 machine is getting half of the memory bandwidth 
of the P4 due to the DIMM configuration.

I'd be interested to see what happens if you compared the dual-channel 
Pentium-4 configuration to a dual-channel Athlon-FX, or a single 
channel Pentium-4 configuration to a single channel Athlon64.
I'd be fascinated to see what happened if you ran a second set of tests 
with a single memory DIMM for a more level playing field.

At the very least you should make a mention of this.  It should be 
rather interesting to point out that the single-channel athlon64 had 
the upper hand in those tests compared to the dual-channel 
configuration P4.

-- 
Peter Wemm - peter at wemm.org; peter at FreeBSD.org; peter at yahoo-inc.com
"All of this is for nothing if we don't go to the stars" - JMS/B5