DistrRTgen generation speed has increased by 50% in little less than a day! We are now doing 15 bil links/sec. See https://www.freerainbowtables.com/phpBB3/topic2952.html for a new graph on the project website!
More...
Printable View
DistrRTgen generation speed has increased by 50% in little less than a day! We are now doing 15 bil links/sec. See https://www.freerainbowtables.com/phpBB3/topic2952.html for a new graph on the project website!
More...
Now, if that isn't COMPELLING EVIDENCE that higher credits awarded sends a project's research power and participation throught the roof, then I don’t know what is. WAKE UP D.A. and anyone else controlling BOINC and open up your eyes and minds to reality!
Ah. But the problem is this boost in production is solely because of the fictional credits, and is basically from people chasing credits and dropping whatever project they used to be running. DA wants the science to matter, not the points. This proves his point as well.
Maybe but since the projects DEPEND on OUR systems doing THEIR work, well....... SCREW DA.Quote:
Originally Posted by Mumps
So a simple solution along DA's lines of thought; A CPU gets 50 credits from a project for crunching 24 hours (WuProp) regardless of speed or # of cores. A GPU gets 100 credits per 24 hours regardless of speed etc. Then all projects pay the same and everyone will crunch purely for the science.
Of course then all the folks with farms will take a look at their utility bills and wonder if it is worth it -or- there would be a run on Atom processors. :))
Understood. My point ;) is that his point is over rated. His way is not the only good way, or even the only "valid way". People will make their presence felt in a project that offers a certain level of "perceived value" to the crunchers. There are multiple ways to achieve "perceived value" in the market place. In the BOINC marketplace, one way is to do as D.A. wishes, and that is make the project valuable on a scientific basis. Fair enough. Projects like WCG and Einstein are prime examples of pitiful credit return, and yet attract major computing resources.Quote:
Originally Posted by Mumps
Another way is to offer BIG credits - especially if the project is not especially appealing (wouldn't you rather search for Pulsars than find prime numbers with CUDA cards? I certainly would.) Now, the question becomes, "Why does D.A. feel he must wipe out the second approach?" It's obviously working beyond even his own wildest dreams to attract the masses to projects. How (or why) does he feel threatened by this?
In fact, the best approach would be to combine a fairly valid science project with a nice credit return, and BOOM - you have a top tier BOINC project, ala Milkyway. Before PG's (admittedly insane, but very nice!) big credit jump on that GPU sub-project, MW was easily on track to become the #1 project. I don't crunch it much these days, but that's a fact. It will pass SETI before too long, although PG is just on a breakneck pace to beat all.
There must be something we're all missing to this. The marketplace is seldom wrong. In fact, some would say it is never wrong. Both approaches are equally valid, so we must figure out why he wants the one eradicated. A monopoly on anything is not real healthy in the long run. I don't think we really want that long term. The system needs to be flexible.
If DA ever gets his will - I will leave BOINC and Distributed Computing. It´s that simple. I will then dig out my Atom board and install Linux, so that I can read my mail, surf and talk to the daughter through skype on the school.
DA needs to understand that one hand washes the other. If he pay the credit, the crunchers WILL come and do his work. Everybody happy.
Beer, you are right; I like that saying, "one hand washes the other." In the U.S.A. we say, "don’t bite the hand that feeds you."
Some could say it's wrong way more than it's right but like religion or politics it's best not to go there on these boards.Quote:
Originally Posted by DrPop
For our purposes here, for now, enjoy the credit boost. Might not be there forever.