Error while computing
15 Sep 2015 13:43:23 UTC
Topic 198229
(moderation:
Three work units I received on 9 Sept and one on 24 Aug had error while computing. I looked and didn't see any other threads re this. It seems that three units the same day is not a coincidence--or is it? Has any others had this problem? What is going on?
Red Thumb
Error while computing
)
Got a private message from "Tom". I couldn't figure out how to answer what he said. I consider myself semi-literate with word. Anything beyond that is Greek (or is it Geek). could somebody tell me in "kindergarten" how to fix this. It looks like the problem started about Aug 31. I am on suspended status right now. If I can't get it fixed...
Red Thumb
RE: Got a private message
)
It seems as though you problem is this: "exceeded elapsed time limit" for each of your invalid units, other than the ones it says you aborted on your own. Do you leave a cpu core free, ie NOT crunching Boinc, just to feed the gpu? If not you might want to try that and see if that helps, it should.
Either that or your gpu is busy doing something else and since Boinc is set to run at a low clock priority, meaning most other things take priority, ie surfing the web, playing movies, etc., your gpu just doesn't have enough time to finish them before the deadline.
Another option could be that you are physically using your pc alot and you are using the default Boinc settings of having 'the gpu stop crunching while the pc is in use' or even 'suspend Boinc when the pc is in use'.
To change the setting go to Your Accout on the Einstein home page, then 'When and how BOINC uses your computer Computing preferences' and look at your settings. These are the ones I am referring to:
Suspend work while computer is in use? no
Suspend GPU work while computer is in use?
Enforced by version 6.6.21+ no
'In use' means mouse/keyboard activity in last 3 minutes
Suspend work if no mouse/keyboard activity in last
Needed to enter low-power mode on some computers --- minutes
Those are my settings and you can see Boinc is setup to run at 100% 24/7 on my pc's. That works for ME but may not for YOU!! So go thru them thoughtfully as any changes could affect how your pc reacts when you want to use it for other things.
Couple of comments to
)
Couple of comments to "Mikey"
1) Did you read the original message? I was asking about ERRORS WHILE COMPUTING!
2) Did you read my reply to "Tom"? Every thing you talked about has NOTHING to do with errors while computing. They had to do with units that I aborted as I could see that I could not get them done in time, and thought that it best to get them back into the system so they could be completed in a timely manner. BTW For 90%++ of the time I am crunching numbers of mt 1 computer
RT
RE: 1) Did you read the
)
I'm sure he did.
And he gave you the answer for that. If you still don't realise this, you'd better go back and read the very first sentence again.
He said,
... problem is this: "exceeded elapsed time limit" for each of your invalid units ...To exceed the time limit that the project considers adequate to crunch the task is classified as 'error while computing'. The most likely reason why your host is taking so long to crunch these tasks - to the point that they exceed this time limit - is probably something to do with how your hardware is configured or is being used. I know nothing about APUs but since you are trying to crunch GPU tasks on one, I imagine your problem is that you haven't 'reserved' a CPU core for GPU support duties. You don't currently have any Einstein CPU tasks running but if you do have CPU tasks from other projects, this is probably what is crippling the GPU. My guess is that CPU support is rather more critical for the GPU component in APUs than it would be for discrete GPUs on an external card.
So how did mikey know what caused your problem. He clicked on your name to view your details and he clicked on the 'view' link to see the list of your computers. You have one, so he clicked on the 'tasks' link to see the list of tasks. Then he clicked on the name of one of your 'error' tasks to see what caused the error. This is something you should do for yourself rather than being rude to someone who did it for you.
If you want to solve your problem, I suggest you try reducing the number of CPU cores that are also crunching - also something that mikey suggested which you don't seem to have picked up. Your computer shows as "AMD A6-5350M APU with Radeon(tm) HD Graphics", a dual core notebook APU. Are you using both CPU cores for crunching? If so, you should change your BOINC preferences locally through BOINC Manager to use 50% of the cores and see if this gives you a big speedup in GPU crunch times. There are other ways to reserve a CPU core for GPU support duties but this is the quickest/easiest way to see if it will make a difference.
Cheers,
Gary.
Hey guys,All of my CPU
)
Hey guys,
Sorry for bumping the thread, but it seemed to be the most relevant one to my problem.
All of my CPU tasks die after 4-5 seconds (0 sec CPU) of runtime. Am I missing some libraries?
Now I'm on Arch linux, earlier I didn't have this issue with Ubuntu 18.10 :D
Hi Tamas Most likely you
)
Hi Tamas
Most likely you need to enable "Run Linux app versions built with LIBC 2.15" in the project preferences.
Oh, I remember I saw that
)
Oh, I remember I saw that option when I signed up for the project and crawled through all the menus, but couldn't figure what that would be good for :D
I already enabled it on the go 2 hours ago, we'll see what happens in the next week or so, but I suspect there must be something else behind this, since SETI x86_64 tasks also behave the same way while there are no such switches on their preferences site.