| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: The Clean Energy Project - Phase 2 Forum
Thread: Questions about participation drive for CEP2 |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 152
|
|
| Author |
|
|
anhhai
Veteran Cruncher Joined: Mar 22, 2005 Post Count: 839 Status: Offline Project Badges: 
|
CEP2 scientist say they are trying to increase the number of people who participate in CEP2. However, I was wondering a few things.
----------------------------------------How much more can harvard's server really handy (50% increase? 100%? 200%?) How much of the "allocated" crunching power that WCG allows CEP2 to have is being used by CEP2? Personally, I find that whenever I finish a CEP2 WU (thus freeing up one of my limited slots for CEP2) I will always get another CEP2 WU. This tells me that there are plenty of CEP2 WU available in the feeder, probably more then the expected ratio. (ie not enough cruncher for the project) Just wondering, if anyone has an answer, please reply. Thx  |
||
|
|
gb009761
Master Cruncher Scotland Joined: Apr 6, 2005 Post Count: 2955 Status: Offline Project Badges: 
|
What I'd like to see, is a follow-up e-mail sent out to all the Grid members, expanding upon the announcement made when CEP2 launched for Windows/Mac back in November, stating that it IS an opt-in project, the reasons WHY it's an opt-in project, along with simple instructions on a) how to opt-in, and b) how to run more than 1 CEP2 WU at any one time (with, of course, the possible drawbacks of doing this).
----------------------------------------We've got to make this as simple and as foolproof as possible, and also to realise that, due to it's special nature, it'll probably never get up to the same/similar throughput as all the other, non opt-in/special requirement projects. Edit : Is there some way of fully automating the special requirements (other than the bandwidth/memory usage etc.), i.e., checking/ensuring that the computer is on a certain number of hours in a day/week? That way, the opt-in option could be removed, defaulting to 1 CEP2 WU/computer - that is, unless the user has manually selected to accept more. This would then open up CEP2 to all those who have selected to crunch all projects.  [Edit 1 times, last edit by gb009761 at Jan 11, 2011 4:44:56 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Dear anhhai, gb009761, and others,
we have a lot of capacity to spare - we could theoretically accomodate an increase of 2000% of the current level. You are right, that we don't have enough crunchers to fill up our quota. We believe that this is mostly due to the opt-in requirement which many casual users probably don't want to be bothered with. This is unfortunate, but we also don't want to jam up low end machines that will not be able to handle CEP2 work. A follow-up eMail is a great idea (perhaps including our youtube installation/setup clips), but this is something IBM/WCG has to decide. I think one could formulate optimum settings for individual machines, but I cannot imagine that this can be integrated in BOINC. Best wishes from your Harvard CEP team |
||
|
|
David_L6
Senior Cruncher USA Joined: Aug 24, 2006 Post Count: 296 Status: Offline Project Badges: 
|
I don't run this project because:
----------------------------------------1) Had a LOT of problems with the first phase. 2) I come here and see that there are problems with this phase. I'll keep checking back now and then (just as I have been all along) to see if the bugs are ironed out. Until they are, I won't run this project as I don't want my computers wasting energy producing errors when they could be doing some good on another project.  [Edit 1 times, last edit by David_L6 at Jan 13, 2011 11:20:17 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
David_L6,
----------------------------------------Have you got any new observations beyond repeating long expired experiences you had in a previous decade? It would help to know your specifics so that if fixable and significant, they can be addressed by the techs and the Harvard team. FAIK the only ''regular'' problem left and being tackled in a current beta is the upload issue. Beyond that, yes this project is demanding and therefor it's opt-in. cheers PS: One member quite recently reported to run CEP2 on a P4 with HT... 2 concurrently, whilst his spouse is using the computer, without complaints reaching his ears... what more testimony does one need? [Edit 1 times, last edit by Former Member at Jan 13, 2011 1:15:05 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi David_L6,
Many of the early problems of CEP2 got ironed out by now (and there is no software overlap with CEP1). So you may want to give it another shot - if it doesn't work out for you it's perfectly fine to donate your time to other projects. As SekeRob said - we appreciate all reports of concrete problems, so that we can improve the project. Best wishes, Your Harvard CEP team |
||
|
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges:
|
cleanenergy, the fundamental issues are the amount of RAM and bandwidth required. Many dual core systems have 2GB RAM and many quads have 4GB. With the operating system overhead, this means limiting to less than 2 or 4 CEP2 tasks per system.
Even for more powerful systems RAM can be an issue; more cores/threads requires even more RAM. Many high end systems are restricted to 16GB RAM due to the motherboard. RAM expense is also an issue for many crunchers. In Q3 8core/16thread systems will hit the mainstream market. If it were possible to multi-thread these tasks that might solve the RAM issue, albeit at some server communication overhead. I think one could formulate optimum settings for individual machines, but I cannot imagine that this can be integrated in BOINC. It would be useful to have a Boinc opt-in so that set it and leave crunchers could allow scientists or CA's to configure/optimize crunchers settings. Elsewhere this is even more of a problem; crunchers with really stupid settings end up getting no tasks or continuous failures. I'm talking 20% of crunchers for some projects. Good luck, |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
cleanenergy, the fundamental issues are the amount of RAM and bandwidth required. The science requires what it requires. RAM and bandwith requirements are determined by the scientists,ans as such are not an issue, but a specification.  I think one could formulate optimum settings for individual machines, but I cannot imagine that this can be integrated in BOINC. I am sure that the CAs would post the page of specs that you could develop.  So if I can run 4 CEP2 WU simultaneously on an i5, for a month with no errors, who believes the problem is with Boinc? Does it help or hurt to know that I also have had an AMD tri-core and 6 C2D running flawlessly while loaded to the gills with CEP2. Change Boinc!! Change Boinc!! The new 'number of wu' option solves everyone's problem. I am at 1-2 depending on the CPU and totally in 'set and forget' mode.  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Dear skgiven,
yes, you are right - the inherent demand of CEP2 is still there and will not go away in the future. One possibility is to pair CEP2 with a less demanding project to fully utilize your hardware. We were thinking of an MPI version at some point which is well supported by Q-Chem, but it is quite time consuming to port for WCG, so it is (unfortunately) not on the immediate agenda. Best wishes, Your Harvard CEP team |
||
|
|
sk..
Master Cruncher http://s17.rimg.info/ccb5d62bd3e856cc0d1df9b0ee2f7f6a.gif Joined: Mar 22, 2007 Post Count: 2324 Status: Offline Project Badges:
|
"The new 'number of wu' option solves everyone's problem. I am at 1-2 depending on the CPU and totally in 'set and forget' mode".
The issue mentioned above concerns the sort of 'set and forget' crunchers who never read any threads, or at least not for a long time. Some of these crunchers set and forgot months or years ago. Most didn't opt into this project when it was a new opt-in only project and don't know how to select to run one or two tasks. |
||
|
|
|