Quick info on what's faster at folding...

GigaHertzAddict · Nov 1, 2006

Both will be using a SINGLE channel DDR motherboard . 512MB

AMD 64 3000+ Socket754 @ stock speed which is 2GHz

or

Intel Celeron 352 512K Cache @ overclocked to 3.6GHz

which will be faster? Thanks for the quick info..

P.S. Overclocking the AMD is not an option.

Junebug · Nov 1, 2006

My rule of thumb is a 2.0 Athlon is about equivalent to a 2.8 pentium, so at 3.6, the pentium will typically be about 30% faster (of course this is WU dependant).

Junebug

godofgorks · Nov 1, 2006

http://www.ocforums.com/showthread.php?t=462713

Celerons fall under "Intel Netburst not using HT"

At you quoted speeds with ChasR data... the winners per core
Core Difference CPU
Ambers: +4 ppd Celeron
Gromacs: +24 ppd Celeron
DGromacs: +32 ppd Celeron
GBGromacs: +6 ppd A64
Tinkers: +6.4 ppd Celeron
Big Gromacs (600 pointers): +56 ppd Celeron
Big Gromacs (364 pointers): +6 ppd A64

Looks like Celeron has just a little bit of an edge

ChasR · Nov 1, 2006

Running a single instance:

Gromacs: AMD=120ppd Celeron=140ppd
Ambers: AMD=140ppd Celeron=140ppd
GBGromacs: AMD=150ppd Celeron=162ppd
DGromacs: AMD=310ppd Celeron=342ppd
Big Gromacs: (600) AMD=220ppd Celeron=306ppd
Big Gromacs: (364) AMD=370ppd Celeron=360ppd

You'll have to figure out what WUs you'll get, Gromacs being the most plentiful.

godofgorks · Nov 1, 2006

ChasR, your Production by Core data for A64's in Faq's, is that single or dual? I've always presumed dual.

Also would, going to single channel, decrease both CPUs production proportionally?

ChasR · Nov 1, 2006

On most WUs dual channel makes no difference. It will make a difference on p1495/97 and perhaps on the ribos. Where you will see a big difference is when there isn't enough ram bandwidth and then it will be quite large. You'd only get caught on p1495/97 and on two instances of ribos. I'm going to update my data over the next few weeks, seperate the Dothan/Yonah, Allendales, and Conroes. And update for newer WUs which aren't quite the same as the ones the data is based on. I updated the GB Gromacs on netburst in the Sticky today to 45ppd/GHz/core.

GigaHertzAddict · Nov 1, 2006

thanks, you guys are great... if I can get the Celeron to 4GHz it should kill the stock AMD... 512K on a celeron is great..

godofgorks · Nov 1, 2006

ChasR said:
On most WUs dual channel makes no difference. It will make a difference on p1495/97 and perhaps on the ribos. Where you will see a big difference is when there isn't enough ram bandwidth and then it will be quite large. You'd only get caught on p1495/97 and on two instances of ribos. I'm going to update my data over the next few weeks, seperate the Dothan/Yonah, Allendales, and Conroes. And update for newer WUs which aren't quite the same as the ones the data is based on. I updated the GB Gromacs on netburst in the Sticky today to 45ppd/GHz/core.

Awesomeness.. can't wait to see the update, going to add GPU cores?

stratcatprowlin · Nov 2, 2006

GigaHertzAddict said:
thanks, you guys are great... if I can get the Celeron to 4GHz it should kill the stock AMD... 512K on a celeron is great..

I think it would be better to OC the 2ghz AMD to where it is even with the CeleronD @ 3.6 or 4.0 for that matter.It will at least run cooler than the Celeron D.Overall the A64 is a much more efficient and in my opinion "better" chip.
Your call though.

Quick info on what's faster at folding...

GigaHertzAddict

Member

Junebug

Member

godofgorks

Member

ChasR

Senior Member

godofgorks

Member

ChasR

Senior Member

GigaHertzAddict

Member

godofgorks

Member

stratcatprowlin

Senior Member

Similar threads