Credit
Message boards :
Number crunching :
Credit
Message board moderation
| Author | Message |
|---|---|
|
Send message Joined: 3 Oct 19 Posts: 43 Credit: 40,548,179 RAC: 0        
|
My total credit seems to have jumped a tad ! Was expecting about 65k by now but showing at 818,923 as I write. Credit on tasks in my list are showing normal values so not sure where the extra credit is coming from, do have loads of download errors as noted in the other thread. |
|
Send message Joined: 3 Oct 19 Posts: 1 Credit: 210,041 RAC: 0
|
Me too. Was at less than 10K yesterday, but today at 733K. I note that I've finally got some validations on my linux host but have wu's there that failed after days of running, so perhaps credit has been given for some of those manually. |
|
Send message Joined: 23 Jul 19 Posts: 289 Credit: 464,119,561 RAC: 0    |
I think we do something wrong with trickle... |
|
Send message Joined: 3 Oct 19 Posts: 43 Credit: 40,548,179 RAC: 0
|
Guessing you granted credit for every single trickle that was sent up by each user for all past tasks ? Definitely something goofy going on ;) |
|
Send message Joined: 23 Jul 19 Posts: 289 Credit: 464,119,561 RAC: 0 |
We try a rollback. We gave credit pour all messages, but I guess messages where not deleted... |
|
Send message Joined: 3 Oct 19 Posts: 43 Credit: 40,548,179 RAC: 0
|
Are you still rolling back ? There are some big discrepancies from the last stats update recorded at Free-DC before the jump in credits...  |
|
Send message Joined: 7 Oct 19 Posts: 8 Credit: 2,016,759 RAC: 0
|
My Credits didn't change at all ... |
|
Send message Joined: 3 Oct 19 Posts: 43 Credit: 40,548,179 RAC: 0
|
My Credits didn't change at all ... Must try harder ! |
|
Send message Joined: 23 Jul 19 Posts: 289 Credit: 464,119,561 RAC: 0 |
Yes, we did not solve the entire problem. Doing a 24h rollback, would cancel out the other improvements to the application. I'm probably going to have to leave it that way..... |
|
Send message Joined: 23 Nov 19 Posts: 2 Credit: 8,081 RAC: 0 |
tbh, looks to be a new project in development so not the top priority; but the credits seem WAY too low. We're talking about 400 to 500 credits for 6 days crunch time. That is out of par with practically every other project out there... When time allows, might want to take a look into that, as for the science value the project is a good idea; but per incentive if SETI is paying a LOT more per task done, it might dis-incentivize some crunchers from chipping in... |
|
Send message Joined: 23 Jul 19 Posts: 289 Credit: 464,119,561 RAC: 0 |
I already look at credits, but I still don't understand how to increase them... If a project manager see this thread, maybe he could help me |
|
Send message Joined: 3 Oct 19 Posts: 153 Credit: 32,412,973 RAC: 0
|
tbh, looks to be a new project in development so not the top priority; but the credits seem WAY too low. We're talking about 400 to 500 credits for 6 days crunch time. If you are crunching for 6 days, that is over 5 days too long. https://quchempedia.univ-angers.fr/athome/forum_thread.php?id=23&postid=195#195 I find that setting "Max # CPUs 1" works best for me, and also limiting them to two at a time using an "app_config.xml" file, as explained in that post. |
|
Send message Joined: 23 Nov 19 Posts: 2 Credit: 8,081 RAC: 0 |
I have limited them per WU suspension, course an app_config file is another option, though haven't spent a lot of time mucking with that as I tried the project and also am working 11.5 hour work days now pretty much through the rest of the year... I do run with all 8 cores going, aka hyperthreading, and this could definitely make a difference as some projects benefit from it, and others do not (here there are VMs so that can impact things). Good for people to know, this said 500 credits over the course of a day is still below SETI even, and as such a bump in credits could definitely help with incentivizing the project especially with many crunchers and some of the heavier hitting teams. But all in all, credits can wait priority wise when the fundamentals of a project are in the process of being setup, especially if they're having to negotiate the science within the algorithm at this juncture. It is good to know he's looking at this, and tbh I'm not sure as to his question as I haven't run a project; unless it's something as simple as applying a multiplication factor to the credit granted based on the number of cobblestones completed or what have you... |
|
Send message Joined: 23 Jul 19 Posts: 289 Credit: 464,119,561 RAC: 0 |
the official documentation is vague and proposes to use the available and recommended methods to avoid projects that were abusively rewarding.... So, if possible, to add a coefficient you have to find and understand a piece of code that has been designed to be not easily changed. Probably in other projects they have managed to get around this and I think I can do it too with a little bit of time |
|
Send message Joined: 23 Jul 19 Posts: 289 Credit: 464,119,561 RAC: 0 |
the quick fix I can try, is to change from credit _runtime to credit_new (default). If I find a way to put a coefficient, what should I put ? x2 ? |
|
Send message Joined: 3 Oct 19 Posts: 14 Credit: 32,908,253 RAC: 0
|
Please don't change to Credit Screw. That is a random credit generator and pays horribly. I DO NOT support projects running credit new unless I have no alternative. This is my personal opinion and others may view it differently. This is NOT meant as a threat BTW, I'm just stating facts. :) |
 Send message Joined: 13 Oct 19 Posts: 87 Credit: 6,026,455 RAC: 0
|
I agree, Credit_New is horrible. I would rather things stay as they are before doing that. |
©2024 Benoit DA MOTA - LERIA, University of Angers, France