Index  | Recent Threads  | Unanswered Threads  | Who's Active  | Guidelines  | Search
 

Quick Go ยป
No member browsing this thread
Thread Status: Active
Total posts in this thread: 109
Posts: 109   Pages: 11   [ 1 2 3 4 5 6 7 8 9 10 | Next Page ]
[ Jump to Last Post ]
Post new Thread
Author
Previous Thread This topic has been viewed 160730 times and has 108 replies Next Thread
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
BOINC credits vs WGC

I know that the rate was set at 1:7 when WCG joined BOINC. However now it would seem that this rate is very low, in my BOINC@Austrailia team they have a regular (ummm three monthly?) Aussie Assult and choose a BOINC project for as many as possible of the team members to crunch for three weeks. The forums often are fairly anti WCG purely from a credit point of view as folks don't want to leave their favorite high paying projects for one that does give low credit. In the same forum they also rave about the fantastic work being done, so there is this massive split of opinions. Credits don't cost WCG anything so a slight change in policy say for future credits only to 1:5 or 1:4 would greatly increase the numbers crunching and doing this fantastic work.
I know it would be a pain to technically achieve with databases keeping track of pre and post some changeoever date but worth the pain for the greater good.

My two cents worth anyways...
[Nov 29, 2009 9:13:19 PM]   Link   Report threatening or abusive post: please login first  Go to top 
roundup
Veteran Cruncher
Switzerland
Joined: Jul 25, 2006
Post Count: 841
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: BOINC credits vs WGC

The forums often are fairly anti WCG purely from a credit point of view as folks don't want to leave their favorite high paying projects for one that does give low credit. In the same forum they also rave about the fantastic work being done, so there is this massive split of opinions.


Thanks for this thread, acrux!
I made similar experiences in my team. All the teams that are crunching multiple projects (like all the BOINC-Teams) have clear statistics regarding runtime / credit ratio and in all those statistics WCG does not perform that well.
On the other hand it is hard to find a BOINC project that is more meaningful for some of the most urgent challenges of mankind (like HIV / AIDS, Nutrition).
In fact WCG is in a competitive situation with other BOINC projects. Some Crunchers look at the meaning of a project, some like badge hunting, some like to push their BOINC Statistics.
WCG TEAM: Make a step forward, compare the granted credits with other BOINC projects and do some adjustment. It does not cost that much, but we would have a chance to get more crunchers to WCG.
[Nov 29, 2009 9:51:05 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: BOINC credits vs WGC

I have absolutely no idea what projects get what credit........I just crunch the work I believe in......heck if all folk want is credits there are plenty of dumb math projects for gpu's apparently giving out ridiculous credits.

I think Collatz is the current popular stats hoaring fave. laughing
[Nov 29, 2009 11:47:27 PM]   Link   Report threatening or abusive post: please login first  Go to top 
roundup
Veteran Cruncher
Switzerland
Joined: Jul 25, 2006
Post Count: 841
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: BOINC credits vs WGC

......heck if all folk want is credits there are plenty of dumb math projects for gpu's apparently giving out ridiculous credits.

I think Collatz is the current popular stats hoaring fave. laughing

Right, vaio. There are some projects out there which I have my doubts that they really bring mankind forward.
And yes, Collatz Conjecture seems to be one of the projects that is very attractive for credit hunters compared to other projects. Our team observes this closely and with a daily output of more than 10.000.000 (ten million) credits and the world wide #3 of the BOINC combined statistics (more than 3,5 Billion credits) we also have collected some experience with most of the BOINC projects (see attached picture that would not be suitable for a forum signature).

Maybe it would help if WCG granted more credits. I will keep crunching for WCG - the quality of the projects here is world class and the support in the forum is excellent. However, maybe the granted credits could be a simple way to get even more crunchers to WCG.


[Nov 30, 2009 12:43:12 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: BOINC credits vs WGC

Apparently my favourite project is one of the worst for stats but it doesn't bother me.

Faith in the project is more important (to me).

I guess we all have our own reasons for what we do but I suppose that doesn't matter.
That folk take part is all that matters whatever the motivation.

Never understood why credit isn't the same for all projects, then the only criteria for choosing would be (subjective) scientific merit smile
[Nov 30, 2009 12:55:29 AM]   Link   Report threatening or abusive post: please login first  Go to top 
roundup
Veteran Cruncher
Switzerland
Joined: Jul 25, 2006
Post Count: 841
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: BOINC credits vs WGC


Never understood why credit isn't the same for all projects, then the only criteria for choosing would be (subjective) scientific merit smile


I also never understand that. Can someone bring us some clarification on the BOINC credit system?
[Nov 30, 2009 7:41:39 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Sekerob
Ace Cruncher
Joined: Jul 24, 2005
Post Count: 20043
Status: Offline
Reply to this Post  Reply with Quote 
Re: BOINC credits vs WCG

Does the granted credit (ignoring 64 bit overclaiming for 32 bit sciences) differ over a longer period much at all? It did hardly on my 32 bit quad then, and it does hardly on my 64 bit quad now, so here you are, a project that stays close to the original intend of the BOINC Cobblestone system... 100,000 for 1 Teraflop of calculations.

At any rate, WCG position is and mine too, the inter-project comparisons has long been thoroughly lost and has long been broken. Why my quad gets almost 2.5 times for a Docking@Home job than WCG or 3x more at QMC than what's CLAIMED is a mystery to me and seeing just now on a SIMAP test 25% uplift, for a 64 bit science, which their scientists write that this version is barely faster than 32 bit... well we knew that the FLP registries are 32 bit anyhow... and on and on the discussion goes. In all that, it has been realized that the inter-project credit ratings is thoroughly broken so a BOINC wide redesign is being worked on... so credits will represent the real work done and this system will include the GPU credit grants.

My one and only injection on this series.

PS: 1:7, 1:5 or 1:4 whatever ratio, in the first instance WCG records and grants credits for everything, then multiplies the grants by 7 to arrive at the legacy WCG points, not the other way around.
----------------------------------------
WCG Global & Research > Make Proposal Help: Start Here!
Please help to make the Forums an enjoyable experience for All!
[Nov 30, 2009 9:51:53 AM]   Link   Report threatening or abusive post: please login first  Go to top 
roundup
Veteran Cruncher
Switzerland
Joined: Jul 25, 2006
Post Count: 841
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: BOINC credits vs WCG

In all that, it has been realized that the inter-project credit ratings is thoroughly broken so a BOINC wide redesign is being worked on... so credits will represent the real work done and this system will include the GPU credit grants.


Thanks, Sekerob!
This is good news. It would not be helpful if all the BOINC Projects started a credit inflation to pull more people to their specific projects. This would disperse the world wide BOINC Community.
Furthermore a BOINC wide fair redesign based on real work done could focus the crunchers more on the meaning and signification of a project rather than on the credits. This would be helpful for WCG wink
[Nov 30, 2009 10:54:05 AM]   Link   Report threatening or abusive post: please login first  Go to top 
Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline
Reply to this Post  Reply with Quote 
Re: BOINC credits vs WCG

One need only look at the Boinc project Collatz Conjecture.

How reasonable is it that a project can have less than 1% of the number of Active Host computers than WCG yet generate 2.3X the number of credits as WCG. Even allowing for the possibility of that project having faster computers, it appears unbelievable that the project is reporting 557 Tflops using 5,500 computers while WCG is reporting 343 Tflops on 233,000 computers?
[Nov 30, 2009 2:57:35 PM]   Link   Report threatening or abusive post: please login first  Go to top 
knreed
Former World Community Grid Tech
Joined: Nov 8, 2004
Post Count: 4504
Status: Offline
Project Badges:
Reply to this Post  Reply with Quote 
Re: BOINC credits vs WCG

We will be participating in this effort:

http://boinc.berkeley.edu/trac/wiki/CreditNew

In addition to improving the accuracy of 'claimed credits', this new system should improve the consistency of credit between BOINC projects.

Extensive testing of this new system is being performed. That link is pretty detailed. The quick summary of the system is that it will track the average difficulty for each result returned at the host-application version level and the over all application level. This will give a ratio which represents the efficiency of a specific host at processing work using a specific application version. Once that efficiency is measured, the system should be able to determine appropriate claimed credits more accurately. Additionally, it will make it more obvious to users which applications their systems are more efficient at relative to the overall set of systems.

In reverse, this will also be used to provide more accurate estimates of time to completion when workunits are sent to clients.
[Nov 30, 2009 4:56:42 PM]   Link   Report threatening or abusive post: please login first  Go to top 
Posts: 109   Pages: 11   [ 1 2 3 4 5 6 7 8 9 10 | Next Page ]
[ Jump to Last Post ]
Post new Thread