Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Stata (gllamm) benchmarks for different platforms?


From   Michael Ingre <Michael.Ingre@ipm.ki.se>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   Re: st: Stata (gllamm) benchmarks for different platforms?
Date   Thu, 22 Apr 2004 12:22:46 +0200

Dear list

I would like to thank Buzz Burhans for his very informative e-mail. He came
through to me in my private e-mail while I was on travel and we have some
communication off list. Buzz has also  been kind to let me try a model on
his Opteron for comparison with my PowerBook. I will take his offer soon.

However I would like to make an on-list comment as well.

>> Buzz Burhans:
>>
>> I was fortunate to have some comparisons available where a -gllamm-
>> model was run on different processors.  At the time I was (and still do
>> part of the time) running a Pentium 4 at 2 GHz.  The comparison
>> processors were an single proc Opteron64 at 2.0 GHz, a dual G5 at 2.0
>> GHz, and a dual proc Pentium at 2.8GHz.
>>        The G5 and my P4, running at the same speed, ran essentially
>> similar times. The faster Pentium (2.8 GHz) ran almost exactly
>> proportionately faster than my P4 at 2.0GHz, which was a significant
>> increase, but not enough to make a hardware upgrade worth it to me.
>>        The Opteron 64 ran 3.7 times faster than my P4, although the
>> clockspeed on both my P4 and the AMD Opteron were both 2.0 GHz
>> 

> David Airey:
>
> Seems Stata is slower on the Mac OS X. That's been my perception. It's
> kind of strange considering other tests show the Opteron and G5 as a
> pretty close match:
> 
> http://www.barefeats.com/g5op.html

I agree with David on this. The almost 4 times improvement in speed for the
Opteron compared to the G5 cannot be explained by hardware differences
alone. The Opteron and G5 are both high end 64-bit processors and
benchmarchs (like the on David supplied above) suggest a 10-20% difference
with no clear winner. Any doubt about the G5 performance should be by the
fact that the G5 is used in the worlds third fastest (ever) supercomputer at
Virginia tech. They squeeze out 10.28 Tflops from a cluster of 1100 G5s
http://www.top500.org/

I would suggest that the differences in speed between the two is due to the
fact that Stata on the Opteron/Linux64 system benefits from native 64-bit
code and Stata on the G5 does not. The heavy use of "double" 64-bit floats
in -gllamm- makes all the difference.

Some time ago I asked about benefits from 64-bit G5 code on this list and
the answer was that Stata already benefits from "greater double precision
perfomance" on the G5 and that any improvement in speed from optimized code
would be minimal.

http://www.stata.com/statalist/archive/2003-11/msg00049.html

Now, I would like to challenge that statement. With Native 64-bit code on
the G5, I would suggest the G5 to perform in par with the Opteron.

Most of you have already guessed that I'm biased towards the Mac (to say the
least). However, since I moved from the PC world (some 8 years ago) into the
world of Macs, I have developed a very personal relation to my Mac and I
would like it to continue developing it. The switch has however, been made
at a cost: the Macs have previously lagged in processor speed.

The lesser speed in Macs has not been a huge problem until now (-gllamm-).
However, with the G5, Macs are not behind in speed anymore. If anything, I
would suggest they are in the very front!

What we need mow, is that developers of computationally demanding software,
make full use of the power from the next generation of computers that the G5
represent. I'm hoping for Stata to take the lead soon.

Michael Ingre


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index