2 CPU crunching question
6 Sep 2006 19:33:48 UTC
Topic 191801
(moderation:
I added this computer to my account in the past day:
http://einsteinathome.org/host/729324
It is a friend's computer and has 2 CPUs.
However, my friend reports that while it DOES crunch 2 work units at the same time, one moves much slower.
Here's what she sees (sorry for the fuzziness):
She says the CPU time of the upper one (90%) only ticks off one second for every 6 real seconds or so, while the bottom one (33%) runs at normal speed.
Is that normal? I thought 2 CPUs meant both would run at the same speed. This is when the computer's idle when she's gone at school all day.
2 CPU crunching question
)
My first question would be was else is running on the slower BOINC core.
Alinator
If the computer is doing some
)
If the computer is doing some other tasks then it is common that only one of the E@H threads will suffer. I notice that all the time, for example when i play games one of the Einstein threads will almost stand still while the other continues as normal.
There must be some other tasks keeping one of her CPUs busy, check in task manager, maybe its something less important =)
Team Philippines
I just talked to my friend,
)
I just talked to my friend, and she says it's 1 CPU with 2 cores. Would that make a difference?
edit: Can you tell that I don't know anything about processors? :)
RE: I just talked to my
)
No, i think that there are background-processes running. Like the infamous Windows Indexing Service... Or some AV-software. Just guessing though.
HTH, Alex.
"I am tired of all this sort of thing called science here... We have spent
millions in that sort of thing for the last few years, and it is time it
should be stopped."
-- Simon Cameron, U.S. Senator, on the Smithsonian Institute, 1901.
I don't think the fact it's a
)
I don't think the fact it's a dual core as opposed to a dual cpu SMP box makes a difference. One factor is a lot of software isn't multithreaded effectively, let alone SMP aware, so some research into the routine tasks you have running would be a good place to start looking.
Alinator
I don't want to burden this
)
I don't want to burden this poor girl with all sorts of work, so I'll just leave it be. It's still a fast cruncher.
I was just worried that something might have been broken, or maybe BOINC had broken something, and if that was the case I'd have done something about it.
Thank you for everyone's help.
Nothing seems terribly
)
Nothing seems terribly unusual with the four completed results for this PC.
The three long WUs completed within a 20% margin [15,5K min to 18,9K max]. There is only one shorty at 2K.
All seems well and no big CPU core to CPU core differences are really apparent from those three results.
42208437 13538863 6 Sep 2006 8:10:35 UTC 6 Sep 2006 23:25:48 UTC Over Success Done 15,507.63 111.22 pending
42170177 13520433 6 Sep 2006 2:19:22 UTC 6 Sep 2006 8:25:07 UTC Over Success Done 18,875.59 111.22 pending
42170160 13520425 6 Sep 2006 1:46:01 UTC 6 Sep 2006 17:36:37 UTC Over Success Done 17,969.53 111.22 111.22
42155535 13513387 6 Sep 2006 1:44:51 UTC 6 Sep 2006 2:39:34 UTC Over Success Done 2,022.39 12.37 pending
RE: Nothing seems terribly
)
Those times are fine but don't represent REAL time :)
For example, one of those work units took 17,969 CPU seconds, but in real time took much longer, since the CPU TIMER clicked a second off only once every 6 real seconds.
Go to Task Manager, under
)
Go to Task Manager, under View check "Show Kernal Times", then for CPU History check "One Graph Per CPU". It will give you at least some idea as to what is going on in the background. Or, you can download Task Info from zdnet. It's a great little piece of software that allows you to track everything that's happening on a per cpu basis.
The other thing I noticed
)
The other thing I noticed when looking at the results for this PC both today and yesterday is that it currently has only one result outstanding for Einstein. As a 2 CPU machine it should have at least 2 results in progress. It doesn't right now and didn't yesterday.
In other words currently only one of the 2 cores has a E@H WU to crunch on.
I missed the wall clock vs CPU time angle of your question. Sorry about that.