mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Hardware

Reply
 
Thread Tools
Old 2004-02-17, 20:00   #1
ixfd64
Bemusing Prompter
 
ixfd64's Avatar
 
"Danny"
Dec 2002
California

11·211 Posts
Default Hz vs. FLOP/s?

It seems that Hz are not always the same as FLOP/S...

For example, I've heard that some AMD 2.x GHz processors are faster than the P4 3.x GHz processors...

Is there a FLOP/s specification for processors? Just wondering. If so, where can I find reliable data?

Thanks.
ixfd64 is offline   Reply With Quote
Old 2004-02-18, 10:00   #2
TauCeti
 
TauCeti's Avatar
 
Mar 2003
Braunschweig, Germany

3428 Posts
Default

Quote:
Originally Posted by ixfd64
It seems that Hz are not always the same as FLOP/S...
For example, I've heard that some AMD 2.x GHz processors are faster than the P4 3.x GHz processors...
Is there a FLOP/s specification for processors? Just wondering. If so, where can I find reliable data?
To specify the measuring of the number of floating-point operations per second one has to agree _what_ to measure.

You can measure the theoretical peak rate of common floating point ops like FADD, FSUB, FMUL and FDIV without memory interaction (i think that was the 'original' MFLOP rating).

Or you could agree on a specific workload like solving some equations with LINPACK or the SPEC CFP

Tau
TauCeti is offline   Reply With Quote
Old 2004-02-19, 06:44   #3
E_tron
 
E_tron's Avatar
 
Sep 2002
Austin, TX

3·11·17 Posts
Default

It is hard to put AMD and Intel's processors side by side with the same measuring system. AMD processors perform FPU operations faster while Intel Processors favor integer. There are more complications like FSB, Stages, ect....

Personally I prefer AMD K7 processors, because they are cheap and perform. Intel P4’s perform too, but their price tag is higher and you are directly support what could become an oligopoly.
E_tron is offline   Reply With Quote
Old 2004-02-19, 09:46   #4
aaronl
 
aaronl's Avatar
 
Aug 2003

24·3 Posts
Default

Quote:
Originally Posted by E_tron
It is hard to put AMD and Intel's processors side by side with the same measuring system. AMD processors perform FPU operations faster while Intel Processors favor integer. There are more complications like FSB, Stages, ect....
I'm not an expert but I believe you've reversed the two. Athlons are known for very strong integer performance. Their FPU is also respectable, but Pentium 4's can be much faster with code that is optimized for SSE2. The best evidence of this is GIMPS performance on Pentium 4s as compared to Athlons.
aaronl is offline   Reply With Quote
Old 2004-02-19, 14:19   #5
dsouza123
 
dsouza123's Avatar
 
Sep 2002

66210 Posts
Default

Athlons are faster using the FPU, P4 have a very slow FPU which is why SSE2
instructions that do floating point math are used instead.

The P4 integer unit (for some code) runs at twice the clock speed.

Athlon64s have speed/access issues with SSE2.
dsouza123 is offline   Reply With Quote
Reply

Thread Tools


All times are UTC. The time now is 18:51.

Tue Dec 1 18:51:07 UTC 2020 up 82 days, 16:02, 2 users, load averages: 1.62, 1.65, 1.89

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.