BPU is working like a charm. I am getting a mix of 30 second and 900 second tasks.
Edit: 19k tasks ready to send. I think they are back to 2 tasks per GPU, no limit on the total in progress.
BPU is working like a charm. I am getting a mix of 30 second and 900 second tasks.
Edit: 19k tasks ready to send. I think they are back to 2 tasks per GPU, no limit on the total in progress.
Last edited by zombie67; 02-21-12 at 05:08 PM.
They are going to zero the high paying credits tomorrow Dang
Here is what GDF said when I pointed out they were cutting the credits to levels below Moo:
My replay which he hasn't responded to yet:Maybe, I have no idea of moo wrapper. The ratio is about right. We are going to align to gpugrid acemd2 credits.
gdf
Then I'm afraid the project is doomed! IMHO of course.
If I can run an established project, Moo Wrapper, that pays higher credits, is stable, gets its credits reported on ALL major stats sites why exactly would I choose to run Donate?
Maybe, just maybe, I would do it to help get a researcher for GPUGrid BUT I think the project will never realize the full potential available from ATI crunchers.
I know they fixed the 30 second tasks. Changed them to something like 250 each. That's better than nothing. These were really failed tasks anyway.
But have they started lowering the credits awarded for real tasks?
The problem was with a bug in the 6.12 app, and it was fixed within 24 hours. Wouldn't they have let go on longer to attract more people, if it was intentional?
This sort of credit over-awarding problem and subsequent credit-correcting is fairly common with BOINC projects.
*shrug*
They said they were going to roll back the credits for the wu that took 30s and were paying full credits.
I had looked at the top Linux box when they had come out with the new app and it was taking almost double the normal time and awarding the same credits. If you look at the "New Application" thread they said the new app was being tied to the credit level of acmed versus acmedlong. When I questioned them on it they replied that that was correct. I then asked them why would someone run the project and get less credits than Moo which is an established project.
They didn't reply to that and they haven't dropped the credits on the Windows app. Maybe they won't. You can no longer see the top computers to check what the Linux box is doing since they are only showing the top 5. That machine was #3 until the credit boom happened.
INTERESTING: The last post I made on that thread is no longer there. That is the one I asked why would I run donate when I could get higher credits on Moo I quoted that post on this thread a couple of days ago. But you can see their response to the dropped credits (which hasn't happened yet at least on Windows boxes).
http://donateathome.org/forum_thread.php?id=47
I don't think it is a DB problem. It is a PBCAK (problem between the chair and keyboard) of the person who set up the main page. I think that person changed the queries on the stats pages to limit it to a few users so he/she could use those queries on the main web page without comprehending that those queries are used on the stats pages also. That's my theory anyway.
Credits are still at 6250 per WU, but the WUs have roughly doubled in size. Still, a 6970 can get over 500K per day when the WUs don't crash causing a pop-up window which halts all GPU processing until you clear it. So, just to get as many credits as possible while the going is good, I swapped out a 5450 with a 6850 since it is rated very high on the bitcoin/$$$ charts. After upgrading the AMD drivers, I found out the 5450 will work as well. I've even got the old 4830 on it. Hoping to see 1M per day if I babysit the boxes and close the pop-ups when the WUs crash.
Interesting side note... They hope to earn 18K Euros a year. The only problem with that is that the number of bitcoins found is constantly decreasing. The more people the "mine" them, the faster they will decline. So, this may not be all that long lived. (e.g. 18% in harder to find Oct, 8.5% harder in Nov, 5.9% in Dec, 7.8% in Jan). Eventually, like the Alaskan Gold Rush, it will get to the point where very few are found and those doing the mining to earn money will quit because the electricity cost outweighs the value of the bitcoins. Then again, if they keep paying us to do the mining for them and trade us useless credits for bitcoins.....
Spring 2008 Race: (1st Place)