Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
![]() |
World Community Grid Forums
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
No member browsing this thread |
Thread Status: Locked Total posts in this thread: 48
|
![]() |
Author |
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
You really have missed the point.
We are having some serious miscommunication here, and if I have been in any way unclear, I apologise.
|
||
|
courine
Master Cruncher Capt., Team In2My.Net Cmd. HQ: San Francisco Joined: Apr 26, 2007 Post Count: 1794 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
This requires you know to opt-out. The part about bias I think is the most missed.
----------------------------------------The point is better driven home with this example: Let's pick a team like, for this example Gay USA, who would have a bias for their attack on AIDS, to fight for my team's bias to end cancer in this lifetime. What they do is take all the single core machines, as mentioned, and point them at HCC w/ the "crunch anything" switch set on. So if there is no HCC wu's to crunch, keep busy. What I do is run the other way around with dual cores. Pointing at FAAH only. Both teams win out by a good 7%. Not just points, but efficiency. This is done to actually avoid the balancing of the grid, as multi core systems would still be sent HCC wu’s. What this thread really addresses is the lack of civic awareness to the HCC problem in general. But for now, let’s just handle this. ![]() ![]() ![]() |
||
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3715 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Hi Courine!
----------------------------------------Your reasoning would be good if the assumptions which support it were right. Unfortunately not all multicore machines are affected by this problem, and not all single core machines are exempt. If you had read the thread that you sent off track last week (that can work both ways ![]() Currently I have only one machine and it is very close to what you describe: a P4 at 2.4 GHz with 1.5 GB RAM. And I experienced the same problem with billion page faults on HCC. Maybe (not sure) because it is a HT processor, as many machines of this size. I have simply gone back to DDD-T after testing HCC. Sincerely I think it is reasonable to simply let people know that some projects might not run efficiently on their machines and that they are welcome to opt them out until a solution is found because the corresponding work will be redistributed to other machines. Trying to force drastic "solutions" not supported by clear evidence could be worse eventually. And please don't misread me: I have no doubt that your intention was good. ![]() Cheers. Jean. |
||
|
courine
Master Cruncher Capt., Team In2My.Net Cmd. HQ: San Francisco Joined: Apr 26, 2007 Post Count: 1794 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Dear Jean,
----------------------------------------Be sure. If you are running anything here, HT is just overhead if it's a dedicated 24/7 cruncher. If you load 2 HCC wu's at once, you will see the problem. Realize, the side product of this thread is now the main issue. I did not side track the other thread, it was dead by request. I just waited a month before saying something. But look what happens when I do. The point now is to see how long it takes for you guys to figure out what the new point is. This will make the letter that explains the whole affair quite simple. As it will be this thread. This is not a threat, for if I wanted to be a problem, I would say nothing and just go around this to those who it really effects. There is something else, but let's get past this one first. ![]() ![]() ![]() |
||
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3715 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Courine,
----------------------------------------As reported then in this post this case is not straight forward and is not simply a matter of two heavy tasks fighting for resources. In case you don't want to jump to this post my HCC task was running alone in a 1.5 GB machine and starting another process after watching TaskManager for a while did not increase or lower the page faults rate at all. What I don't figure out at all is what you exactly want from a practical viewpoint. - Tell people that some of the machines running HCC are not performing as well as we would like it? Nobody has tried to hide it and anybody who cares knows it. - Stop running HCC until the problem is identified and fixed? That would be losing 60,000 work units a day for nothing. Remember that these WUs are not erroring, they are only taking a little too long in some machines, and nobody can even say which percentage of these WUs have been really slowed down. So what? If that makes you feel better to express your frustration, please do, but be not surprised if you don't move crowds behind you. Cheers. Jean. |
||
|
courine
Master Cruncher Capt., Team In2My.Net Cmd. HQ: San Francisco Joined: Apr 26, 2007 Post Count: 1794 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
If that makes you feel better to express your frustration, please do, but be not surprised if you don't move crowds behind you. Dear Jean, The reason why you don’t see the crowd, it is my email. None that have accepted what I suggest will come speak here because of the sandbagging. Quite frankly, they understand and have accepted my offer, but they don’t want to be targeted. But then again, "why am I being targeted?" The same reason why this "world community" is so small. See. I know the reason, but it is best for this forum to figure out why for itself. So the best I can do is pose the clues that would bring the needed change. Want a clue? Most of the questions I bring up are from "guest" emails. ![]() ![]() ![]() [Edit 1 times, last edit by courine at Feb 7, 2008 12:53:51 AM] |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
What emails?
Jean, your post was most to-the-point. |
||
|
courine
Master Cruncher Capt., Team In2My.Net Cmd. HQ: San Francisco Joined: Apr 26, 2007 Post Count: 1794 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Dear D,
----------------------------------------Welcome back. So the question is not "what emails", but how do I get them? ![]() ![]() ![]() |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Ah, back to riddles!
I think I know this one. |
||
|
courine
Master Cruncher Capt., Team In2My.Net Cmd. HQ: San Francisco Joined: Apr 26, 2007 Post Count: 1794 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Well. how do I get them? Or better, Why?
----------------------------------------![]() ![]() ![]() |
||
|
|
![]() |