Stanford introduces new points reference PC

[Allan]

Stanford are now kicking the old points reference machine out, and in comes a new P4 2.8 GHz. According to Stanford, this should provide more standardization for points and deadlines for the folding@home project.

On a side note, Tinker scoring is about to go up (finally) so it won't matter much if you run Gromacs or Tinker for points.


Read it all here: http://forum.folding-community.org/viewtopic.php?t=7665

[SteveI]

Interesting....
Old machines folding tinkers will have MUCH better production...

It should make it much easier for people to catch me.

[Daniel ~]

Anyone catches you Steve and they will have had to of earned it! LOL

[keithwalton]

cool just today i cranked up my folding power quite abit. You should notice a rather large spike on my nick (TuViE) i was about to be overtaken, no chance was i going to let that happen, roll on my highest point scoring day ever (allready beaten my prev record)
Though it will drop off abit as a 160pt'er just went through and the systems got another todo but it'll take 92 hours

I have seen a unit recently using sse2 anyone the wiser on this as it flew by and scored well.
Although the p4 munched several small units today (3-4 hours each)

K

[dod]

Originally Posted by keithwalton

cool just today i cranked up my folding power quite abit. You should notice a rather large spike on my nick (TuViE) i was about to be overtaken, no chance was i going to let that happen, roll on my highest point scoring day ever (allready beaten my prev record)
Though it will drop off abit as a 160pt'er just went through and the systems got another todo but it'll take 92 hours

I have seen a unit recently using sse2 anyone the wiser on this as it flew by and scored well.
Although the p4 munched several small units today (3-4 hours each)

K

Soooo, the details of this crank in power are....???

[keithwalton]

There in the sig which i've enabled at last :-)
Basically was just the tbird for a while recently (used to be just the p3-m). If you follow the link in the sig you'll see a rather nice spike :-)

K

[Allan]

Cool folding output increase, Keith!!!

[Ebbe Bech]

Isnt it a problem if the tinkers gets more valuable as the gromacs offers better ressearch results.
We do more gromacs WU as they are faster to complete, but if the tinkers let u earn more points that way, wont some people do tinkers instead?
Not that Im against that tinkers gets more valuable. I remember the enghome_proteins. OMG, I hated them!

[keithwalton]

Just my 2p really
in my opinion i think points should be scaled as per time it would take them to complete so that points are racked up for the number of hours of work put and irrelevent of what the actual work is so that everything gets done and nobody does anything selectively just to score better.
The project isnt about who's got the highest score, it's giving what you can to help fellow mankind. I'm putting in everything i can afford todo at the mo and have spread the word amoung friends.
Yes competition helps motivate people todo better.

[SteveI]

Originally Posted by keithwalton

Just my 2p really
in my opinion i think points should be scaled as per time it would take them to complete so that points are racked up for the number of hours of work put and irrelevent of what the actual work is so that everything gets done and nobody does anything selectively just to score better.
The project isnt about who's got the highest score, it's giving what you can to help fellow mankind. I'm putting in everything i can afford todo at the mo and have spread the word amoung friends.
Yes competition helps motivate people todo better.

With all due respect, the actions taken by Pandegroup to change the scoring is indicative of the the vocal attempts by those only caring about score... and to make them shut up and fold.

The newer gromacs core uses 3dnow+, SSE, or SSE2 to be able to calculate 60% - 90% more efficiently. The result is that from a productive point of view, the gromacs units are in fact much more valuable... Coincidentally, the celeron 500s used to benchmark do not have SSE so took much longer to complete, ergo the higher point value.

The new benchmark machines is a P4 2.8. Because it has SSE, and SSE2, it will not give the gromacs a much higher value, and the tinkers get more points since they take so long. However, in terms of usefulness, they are still worth much less for research.

In the end, I am not trying to judge, just lay out the facts. I couldn't care less... Scoring has been highly subjective, and if it bothered me, I would have quit long ago. I want higher participation, and if a new scoring system helps, I'm all for it. I'd give back every point if it meant there were that more people participating to earn an additional 300,000.

Previous changes have hurt my scoring potential significantly. Not that anyone cares but, currently I'm ranked in the top 200 in the world... A couple of years ago, prior to some significant upgrades I might add, I was number 9 in the world.

EDIT: One more thing... Tinkers is much less stressful than Gromacs, and runs a bit cooler. For those concerned about the longevity of their machines running full bore... or those with stability issues running Bartons o/c'd to the moon, change your preference to Tinkers to improve stability.

[keithwalton]

Just got a new unit on the p4 rig :-) 107 points for 8 hours worth of folding, its a sse2 gromacs unit :-)