Proteins losing points?

[SmokeRngs]

I just finished a 600 pointer earlier today and sent it back. While checking my stats it seems I was only credited for 500 of the points. I also finished a 364 pointer today and was only credited with 292 points.

Has anyone else seen this happening or know of any announcement about WU's getting their points cut? Since they are still using a P4 2.8 as the benchmark machine, I don't see why these proteins would have their points changed.

 

[marty9876]

Bonus on them WU's reduced ~ 50% I think it was.

364 point WU's now = 292

600 point WU's not = 500

 

[[H]ugh_Freak]

Correct me if I'm wrong, but this basically means if you aren't running the SMP or GPU clients you are screwed in the points department?

 

[marty9876]

Basically, I can link you a fuckton of threads at the official forums on all this but...

Kasson said screw you all and reduced the points on those projects. Why? No formal reason posted as of yet to my knowledge but there has been a overload of talk/action on dumpers/blockers/screwed up points system.

My personal official position on all this is I don't care. 2 points for Kasson.

I think the other 805 point tet dropped to 755 points too.

 

[SmokeRngs]

Those damn 805 pointers should be worth even more considering they take up a hideous amount of RAM. The only thing I've seen comparable on those are QMDs. I don't remember exactly since it's been a little while since I got one, but I think they take up almost 300 meg of RAM.

I am annoyed about the loss of points on the 600 and 364 pointers. I can live with it, but it would be nice to see an announcement about it and the reason for it. It's not that they take a long time to crunch or anything, but they do use up a good bit of RAM depending on which ones you get. I've had some take up 100 meg while others have gotten closer to 200.

Personally, I see no reason to change the amount of points they are worth unless something major changes. The WU's aren't using less RAM and they didn't change their benchmark system so I don't see what could have caused the change.

If I actually went to their forums I might have know this, but it's so godawfully slow. I've gotten better speed on response on dial up. Then again, they may have changed some things since the last time I was there. It's been a while especially since they canceled my registration considering I hadn't been there within a certain period of time.

 

[ICE_9]

It used to take 30 seconds for me to log in there. They seemed to have gotten faster since.

 

[Bill Clo]

There was and is a LOT of pissing and moaning about bonus points, so the bonuses were reduced on some work units. People ought to be careful about what they wish for, they just may get it.

Being as I go heavy on RAM for my systems, especially the ones I do -advmethods flags on, the RAM requirements haven't caused any problems for me. If you were on a system with 256-512mb RAM and running a large work unit, I could see it being sluggish.

 

[DriveEuro]

Correct me if I'm wrong, but this basically means if you aren't running the SMP or GPU clients you are screwed in the points department?

Seems to be right. The SMP and GPU points haven't changed and they still yield pretty high PPD.

 

[Tormond]

I actually agree with Marty (/shiver) on the post he made on the stanford forums about switching back to the 1WU/1 Point deal that was original. I would be raking in the points with all these 3302s that count but EUE after 30 min

 

[[H]ugh_Freak]

There was and is a LOT of pissing and moaning about bonus points, .. etc....

Bill, I wasn't pissing and moaning about bonus points. Well ok, kinda.

When I first started folding, the highest point value WU was, I think, 77 points. However, those also took the longest to fold. I'm just of the opinion that if a WU takes 3 days to complete, it should be worth more then one that takes less then half the time (on the same system of course). The bonuses make no sense to me, and probably everyone else. Some gromacs get them while others don't, and amber core proteins must get a negative bonus cause they fold slower then crap on every system (even slower on AMDs) and aren't worth squat. If Stanford actually benchmarks them all on the same system, then there should be some resemblance to an order and there just isn't. Sorry if I think a guy with a quad core and a video card shouldn't be able to outproduce 10 or 15 average systems.