Has anyone gotten the F@H 1487 unit?

I've just got a 1479, 364 points at 488 ppd on my opty 146:D:D cetainly beats the 772, 44 pointers that have been at 130 ppd:(

Project : 1479
Core : Gromacs
Frames : 100
Credit : 364

Min. Time / Frame : 10mn 44s - 488.35 ppd
Avg. Time / Frame : 10mn 44s - 488.35 ppd



Project : 773
Core : Gromacs
Frames : 100
Credit : 44

Min. Time / Frame : 4mn 39s - 136.26 ppd
Avg. Time / Frame : 5mn 09s - 123.03 ppd
 
Last edited:
Project : 2500
Core : GBGromacs
Frames : 100
Credit : 208

Min. Time / Frame : 42s - 4278.86 ppd (hahaha if only it we're true :D)
Avg. Time / Frame : 58mn 58s - 50.79 ppd
 
I've just scanned through all the benchmarks I've got and found this one....

Project : 1163
Core : Tinker
Frames : 400
Credit : 241


-- BLUE --

Min. Time / Frame : 9mn 07s - 95.17 ppd
Avg. Time / Frame : 9mn 15s - 93.79 ppd


The same opty @ 2.7GHz and it only gets 93 ppd :confused:

I'm sure those 241 tinkers were always worth getting on my old XP2500!
 
rich99million and Mikey1280, you both mentioned project 992. Seems I ran this at some point on my single core opty
Code: 
Project : 992  
Core    : Double Gromacs  
Frames  : 100  
Credit  : 130       
Min. Time / Frame : 5mn 29s - 341.40 ppd  
Avg. Time / Frame : 5mn 29s - 341.40 ppd
The core used was double gromacs? It's only single core though :confused:
 
joeyjojo said:
rich99million and Mikey1280, you both mentioned project 992. Seems I ran this at some point on my single core opty
Code: 
Project : 992  
Core    : Double Gromacs  
Frames  : 100  
Credit  : 130       
Min. Time / Frame : 5mn 29s - 341.40 ppd  
Avg. Time / Frame : 5mn 29s - 341.40 ppd
The core used was double gromacs? It's only single core though :confused:
Click to expand...
that's some very tasty points - what speed is your opty running at?

It's called double-gromacs because it uses double precision floating point values for the calculations, they do very well on the opty/A64 etc as they take advantage of SSE2 :)
 
rich99million said:
that's some very tasty points - what speed is your opty running at?

It's called double-gromacs because it uses double precision floating point values for the calculations, they do very well on the opty/A64 etc as they take advantage of SSE2 :)
Click to expand...

Are there any units that can use sse3? Are they going to make some that can? Is there a new version of fah coming out soon as well or something?
 
rich99million said:
that's some very tasty points - what speed is your opty running at?

It's called double-gromacs because it uses double precision floating point values for the calculations, they do very well on the opty/A64 etc as they take advantage of SSE2 :)
Click to expand...
Opty 148 at 261*11 = 2871Mhz with 1.55V DFI Expert, very tasty :D
Mikey1280 said:
Are there any units that can use sse3? Are they going to make some that can? Is there a new version of fah coming out soon as well or something?
Click to expand...
These FAH fiends will tell you more ( :p ) but there was some licensing situation with intel and compiling with sse3, so amd systems couldnt use it

//edit linky
 
There haven't been any QMD units in about six months. They did use SSE2/SSE3. AFAIK it was the only core that used SSE3. In Intel's FORTRAN compiler ande math library EULAs they state that the code produced by the compiler should be optimized for Intel only. It detects if it sees the string "GenuineIntel." If, perchance, it sees "AuthenticAMD" it defaults to a state where the optimizations are not on. To hack around this is a violation of the license agreememt.

Running QMD WUs on an AMD is possible, but legally they cannot use SSE2/3 optimizations. Without them it's not even worth doing as it's so slow. Folding@Home could switch to a different compiler and math library but Intel's is significantly better than all others in terms of both speed and accuracy.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom