Re: [ccp4bb]: Version 7.2 of RedHat / Intel's ifc for Linux

[kay <kay.diederichs@uni-konstanz.de>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***

Boaz Shaanan wrote:
> 
> Hi Kay,
> 
>  In relation to your message concerning the ifc compiler: do you have any
> opinion/exprience on how the AMD cpu (say XP 1.8Ghz) compares with the p4 when
> it comes to crystallographic programs (cns, ccp4, etc.) ?
> 
>   Best regards,
> 
>              Boaz

Hi Boaz,

I have Dual-MP1800+ on order so I can't currently say how XP1800+ compare with P4. 

I can however compare P4 1.7GHz with Athlon (TBird) 1100MHz: P4 is a bit faster (like 10-20%) for
g77-compiled crystallographic programs (CCP4). For an ifc-compiled program (CNS) the difference is
on the order of around 50%. The TBird-Athlons do run well with ifc-produced code.

I expect the XP1800 to be faster than the P4 1.7GHz for both g77 and ifc executables, because it
features SSE (like the P-III) and "prefetch" - the CPU is "Palomino", not "Thunderbird" ("TBird"). 

The P4 2.0GHz is supposed to be about as fast the XP1800, and wins in floating point if compiling w/
ifc (which can use its SSE2 instructions). This is the upshot of
http://www.heise.de/newsticker/data/ciw-28.10.01-003/ which has a nice up-to-date summary, with
pointers to SPEC-pages (the page is in German but you could use a web-based translation service).

I CC to the CCP4BB - hope you don't mind.

Hope that helps,

Kay
-- 
Kay Diederichs                    
email: Kay.Diederichs@uni-konstanz.de   Tel: +49 7531 88 4049  Fax: 3183
Fakultät für Biologie, M656, Universität Konstanz, D-78457 Konstanz