My opinion on the Rosetta@home credit issue

[TallBill]

Why not go with the system of measuring credit done based on cpu and performance as planned, but instead of lowering scores, set the benchmarks up to be equal to what the optimized clients do now.

As far as I know, it was never against the rules to be using an optimized client, so instead of punishing the people who have, instead use those levels to set the standard of what a WU is worth, and take the data back to february to RAISE everyone up to an equal level, instead of lowering everyone to an equal level.

The only problem left here is that credits will be worth more then other boinc projects, but this is an independant project that can measure its work however it wants. Statistic sites will just have to skew results as they see fit.

I hope this makes sense, and really wouldn't piss people off or have issues with a second data column or separate scoring.


--------------------------------------------------------------------------------------------------------------------------

 

[BlackMountainCow]

I've read through the Rosey and Ralph sites for quite a while now and to be honest, I don't get the new system either.

As far as I understand it, they introduce these two new colums as an addition to the old ones, but they won't replace the old ones.

As far as the backdating is concerned, I got it as if they only want to do that to extrem optimizers, not to say cheaters. There were examples were people's PCs reported some (only example) 100,000 million flops, but their CPU is physically only able to do some 50,000 flops. Those they will backdate, but not the "normal" or "regular" optimizers, if I got it right.

 

[TallBill]

I posted this on the rosetta boards Here, and its picking up some steam in the discussion.

The idea is split in two, raising levels up the the optimized client measurement either with or without the backtracking. Without would be fine, since it would give everyone the chance to perform well.

 

[Fardringle]

I've been using optimized clients on most of my computers because I didn't really see any reason not to and because I didn't like the benchmark scores I got from the client. For example, my P4 3.2 duallie server only got about 15% better benchmark scores than a P4 2.0 desktop and that really annoyed me. Besides, points is always better.

As long as the Rosetta team is consistent in the way they award points for everyone, I don't really care if they disregard the client benchmark scores completely and give points based on the actual work units. Personally, I'd prefer it that way (how Seti gives points now) so fast computers get rewarded for being fast instead of having very similar benchmark scores to slow machines in the official clients.

 

[networkman]

In the Einstein project, the optimized clients were actively embraced which solved quite a few problems.

 

[Insidious]

Points?

We can get points for crunching? :Q

I like this idea!

/me wants some points

-Sid

 

[petrusbroder]

:beer: me 2! :Q