| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Suggestions / Feedback
Thread: Project differences, benchmarks and points |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 24
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
For my computers, the claimed points per hour is consistent at (FP MIPS + Int MIPS)/X, where X is 463+/-3. They are all P3 or Athlon machines.
----------------------------------------However, there are large differences between in points awarded (and therefore in points calculated by other people's machines) depending upon the project. eg. Project %awarded/claimed*Does anyone know what behaviour causes the huge differences between projects? * = For projects which only have a quorum of 2, I've used "claimed by my machines/claimed by other machines". Crediting may be the average, or one result or the other, which causes enormous variance. [edited to include all current projects] [Edit 4 times, last edit by Former Member at May 15, 2008 4:09:04 AM] |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Kremmen, discussed ad nauseam in other support and chat threads, really. 64 Bit client particularly overclaim up to 2x what it's worth.
----------------------------------------Have a nice Sunday, (though i think you're in ozzieland, so it will soon be over).
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I may have missed something, but I've not noticed any authoritative comment on the applications themselves. Some issues, such as 64-bit vs 32-bit clients, are likely to affect all projects. If some projects behave very differently to others, I would expect there might be significant differences in their underlying programming and/or type of data organisation between projects.
If different families of processors have significant advantages in some kinds of projects, maybe it would be useful to work out what those are and make recommendations? |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges: 
|
As discussed very often the benchmarking feature does not give equivalent results from one processor to another or from one OS to another.
----------------------------------------In addition it is also not "proportionally" different, i.e. one benchmarking might be 50 % better for integer operations and only 15 % better on floating point operations (percentages given only to illustrate the explanation). Once this is stated it is clear that different applications will not react identically versus the benchmarking depending on the proportion between integer and floating point operations in their total process. And other factors may be important like the way they need/use memory which may be more or less ideal regarding the physical characteristics of the machine (processor architecture, RAM, etc...) which runs them. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Once this is stated it is clear that different applications will not react identically versus the benchmarking ... When one application consistently grants almost 2.2 times the points of another, saying that they aren't identical is a severe understatement. It would be useful to know (to know, not to aimlessly speculate) what is causing the differences. I'd think perhaps that those who wrote and administer the applications would know the answers. Does WCG have any statistics on what families of machines work efficiently on which projects? |
||
|
|
retsof
Former Community Advisor USA Joined: Jul 31, 2005 Post Count: 6824 Status: Offline Project Badges:
|
Once this is stated it is clear that different applications will not react identically versus the benchmarking ... There are priority and adjustment differences depending on what the techs-that-be want to accomplish. When one application consistently grants almost 2.2 times the points of another, saying that they aren't identical is a severe understatement. Why not run that one then?
SUPPORT ADVISOR
Work+GPU i7 8700 12threads School i7 4770 8threads Default+GPU Ryzen 7 3700X 16threads Ryzen 7 3800X 16 threads Ryzen 9 3900X 24threads Home i7 3540M 4threads50% |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Why not run that one then? I intend to. I did run as many AC@H jobs as I could. I was just bringing the issue up because plenty of machines are sure to be crunching less than optimal jobs and it would seem to me that having some information from the techs about what projects are best suited for what CPUs might be useful. |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: 
|
Kremmen,
The question you ask in regards to differential point awarding is a interesting question and something I will have to take a look at. Ideally it will be that some projects just work better on some machines then on others but I do want to make sure that for some reason some projects are simply awarding less credit. We will take a look at that. We are currently up to our eyeballs in a few things at the moment so it might be a bit before I can take a look. thanks, Kevin |
||
|
|
GIBA
Ace Cruncher Joined: Apr 25, 2005 Post Count: 5374 Status: Offline |
 Kremmen, The question you ask in regards to differential point awarding is a interesting question and something I will have to take a look at. Ideally it will be that some projects just work better on some machines then on others but I do want to make sure that for some reason some projects are simply awarding less credit. We will take a look at that. We are currently up to our eyeballs in a few things at the moment so it might be a bit before I can take a look. thanks, Kevin Kevin, despite be a priority or no, once I guess that you have a lot of pendencies to solve before (or at least at sametime), I agree that maybe Kremmen point of view and questions could improve quickly WCG projects. His considerations could change the machines settings, machines best configurations to run each project or at least change the minimum requirements for each project in WCG. With multi cores technologies evolving very quickly as we are watching, maybe WCG need investigate it asap... I agree with Kremmen considerations and I have interest in allocate the right machines just in right projects to optmize my participation, including in terms of efforts and energies spent for all projects (I have many machines crunching WCG, as many crunchers). I will wait with lot of curiosity your inputs about this investigation to clarify us what happens. Thank you in advance for your interest in take a deep look on that. Best Regards. Giba. 
Cheers ! GIB@
 Join BRASIL - BRAZIL@GRID team and be very happy ! http://www.worldcommunitygrid.org/team/viewTeamInfo.do?teamId=DF99KT5DN1 |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
We are currently up to our eyeballs in a few things at the moment so it might be a bit before I can take a look. No worries, Kevin. The really ironic thing is that it's only now that HCC is finally fixed and runs on my machines that we discover that it's really not an efficient thing for me to run. :) |
||
|
|
|