Posts by AllenIN


log in
1) Message boards : Number crunching : Average Credit Decreasing? (Message 1783046)
Posted 28 Apr 2016 by Profile AllenIN
From April 3rd (22,000 RAC) I was on a steady uphill to 24,200 RAC on April 10th, but on April 14th I was on a fast downhill to 22,100 RAC by April 24th.

Now it seems to be steady at this point not moving up or down more than about 70 RAC. This is for 5 machines, 4 with GPU's.

I was thinking of suspending all of the guppis and see if the RAC starts to climb again.
2) Message boards : Number crunching : Lunatics Help (Message 1783041)
Posted 28 Apr 2016 by Profile AllenIN
Wow, I should have known that.

Thanks a bunch Hal9000 and Stephen!
3) Message boards : Number crunching : Lunatics Help (Message 1782720)
Posted 26 Apr 2016 by Profile AllenIN
Does anyone know if there is a way to prevent Astropulse from running on the cpu and only run on the gpu?

Thanks,

Allen
4) Message boards : Number crunching : Question on EVGA GTX650ti (Message 1775444)
Posted 31 Mar 2016 by Profile AllenIN
Can't figure it out. I came down to check on the computer and it was running the one wu with no trouble...

If the computer clock/time go back tasks (app processes) stop (exit) but BOINC continue to show them "Running"
(at least this happens on BOINC 6.10.58 (and less), I don't know if this happens on current versions of BOINC)

Usually the scenario is:
- computer clock go too fast and after some time is too far in the future
- then Windows syncs time (from CMOS or Internet) to the correct (= for all programs this looks like "go back in time")

In this case:
- tasks (app processes) stop (exit) after 30 seconds
- BOINC continue to show them "Running"

- BOINC restarts the tasks after the "missing" time period expires


Can somebody test this on new BOINC?
E.g. set the computer clock 3 minutes back (and wait/look what happens in the next 5 minutes)
 

Interesting. I do remember when this used to happen in the past, as Richard has pointed out. It was quite strange. Apparently it doesn't happen anymore though. I do wonder if somehow this one system didn't get caught in the twilight zone of the time change we just went through. I just checked on it and it is running 3 wu's and running them all properly. However, it was locked up this morning when I checked on it. Gave it a cold boot and it's been working fine since.....knock on wood!
5) Message boards : Number crunching : Question on EVGA GTX650ti (Message 1775443)
Posted 31 Mar 2016 by Profile AllenIN

BTW, why does your medal show that you are in the upper 5% of work done when mine shows 1% and you have done 7 times more work than I have??


Medals are linked to RAC, not to credit.

Sleepy


Boy, it's great to keep on learning everyday. Thanks sleepy!!
6) Message boards : Number crunching : Question on EVGA GTX650ti (Message 1775096)
Posted 30 Mar 2016 by Profile AllenIN
I do have another question. The gpu usage is very strange compared to the way it used to run under v7. While it sometimes runs at 65-90%, sometimes it stays at 0% for 15 or 20 seconds and then bounces up to 80% or there abouts for a couple of seconds and then takes a break again. Right now it just jumped to 38% and then stopped for a few seconds and then bounce back to 80%. Now it's at 14%.
Is this normal???

Allen

BTW, why does your medal show that you are in the upper 5% of work done when mine shows 1% and you have done 7 times more work than I have??
7) Message boards : Number crunching : Question on EVGA GTX650ti (Message 1775014)
Posted 30 Mar 2016 by Profile AllenIN
Hal,

Can't figure it out. I came down to check on the computer and it was running the one wu with no trouble. I returned it to running 3 and it is now doing that with no trouble. I can only figure that either some hardware is going bad or I had a glitch in the system and it is clear now.

I did go into the bios and looked around. I did check on the gpu pci-e and made sure it was set properly and then saved and rebooted.

I did experienced a few lock-ups of the system before this odd occurrence reared it's ugly head. Now running three wu's and temp is 56c and usage is 65-90% so things are back to normal for now.

Thanks for explaining the .04cpu, I had no idea, I just figured it needed that much cpu for the wu to run on the gpu....stupid me.

Thanks for now! I hope this adventure is over but if it is not, I'll be sure to look you up again. BTW, all the work showed up in the task manager. By the time I got to the GPUz, the machine had locked up and when I rebooted everything was working fine again.
8) Message boards : Number crunching : Question on EVGA GTX650ti (Message 1775007)
Posted 30 Mar 2016 by Profile AllenIN
Suddenly my gpu is not functioning as it was. I am running a gadget called GPU Meter and it shows that the GPU Clock is at 928MHz but that the usage is at 0%. Normally it is bouncing around at about 70 to 90 percent. I have changed nothing and not even done any MS updates lately. I've suspended a few of the wu to see if another would run and I have the same results.

I had had the cfg file set to run 3 tasks at a time and that wasn't working so I changed it to 2 tasks with the same results. Now I have it at one task and still nothing. If it weren't for the .04 cpu, the wu would never progress at all.

I've checked to make sure that everything is clean and the temp is running at 45C which is about 10 to 15 degrees cooler than normal.

Does anyone have any ideas what might be wrong?

Allen

Are you saying that you have no GPU tasks running or there is 1 when there should be more? The 0% load on the GPU would seem to indicate you have no tasks running.
If you have any tasks running on the GPU. Then you might want to start by opening Task Manager and check the Processes tab to make sure the app is actually running.
If it is you might want to try GPUz and see what it displays for the GPU usage. Maybe your gadget is broken.

When you see Running (0.04 CPU + 1 GPU) that doesn't mean the CPU is being used. The 0.04 is a percentage of CPU time the BOINC client uses when deciding how much work to request.

Hal,
Yes, I am running (trying to) one task now, but when this slowdown occurred I was running 3 tasks with no trouble. I will try the GPUz and see what it says and I will also make sure that all tasks are running in task manager and I'll get back to you when I find out.

Thanks for the response!
9) Message boards : Number crunching : Question on EVGA GTX650ti (Message 1775001)
Posted 29 Mar 2016 by Profile AllenIN
Suddenly my gpu is not functioning as it was. I am running a gadget called GPU Meter and it shows that the GPU Clock is at 928MHz but that the usage is at 0%. Normally it is bouncing around at about 70 to 90 percent. I have changed nothing and not even done any MS updates lately. I've suspended a few of the wu to see if another would run and I have the same results.

I had had the cfg file set to run 3 tasks at a time and that wasn't working so I changed it to 2 tasks with the same results. Now I have it at one task and still nothing. If it weren't for the .04 cpu, the wu would never progress at all.

I've checked to make sure that everything is clean and the temp is running at 45C which is about 10 to 15 degrees cooler than normal.

Does anyone have any ideas what might be wrong?

Allen
10) Message boards : Number crunching : WU's timing out........ (Message 1772261)
Posted 18 Mar 2016 by Profile AllenIN
Hal and all you helpers,

Yep, that was it....a bunch of ghosts. I guess I'll just turn my head and let Boinc/Seti figure it out for me.

BTW, is detaching and reattaching a good idea?

Thanks all!
11) Message boards : Number crunching : WU's timing out........ (Message 1772030)
Posted 17 Mar 2016 by Profile AllenIN
With the limits of 100 tasks per CPU/GPU & a deadline of about 6 weeks how many hours a day do you let your machines run tasks?

The only things I can think of are:
1) Adjust your queue size.
2) Stop fiddling with things & let BOINC handle running tasks instead of trying to micromanage what it does. BOINC will run tasks that are near the deadline sooner if it thinks there is a chance of missing them.

EDIT: A single core CPU machine that takes around 10 hours to complete a task would be able to complete a queue of 100 tasks before the deadline if it ran 24/7.


I run 24/7 on all of my machines and I didn't start "playing around" until Boinc was unable to finish all of the tasks it had at it's command. As I said, never had this problem before the change of versions.
Thanks for the input though.

Allen

Are you looking at the tasks in BOINC Manager and seeing they are close to their due date or are you looking at the error tasks list on the website and seeing "Timed out - no response"?
If you are seeing "Timed out - no response" in your errors those tasks are most likely from ghosts that were never on your system.

BOINC runs tasks FIFO (First In First Out) unless it thinks there will be a problem. Then it will run tasks in order by due date. How your cache settings are set can effect how BOINC determines this.


Wow! Now that is interesting. I did see the time outs that got me interested in doing something about it, but I never checked to see if they were actually ever on my system or at least present at some time. I just assumed (stupidly) that they were there and they were skipped for other wu's. They could have been ghosts as you said, since I didn't know anything about 'ghosts'.

I guess at one time or another I could have lost some wu's during the change over of versions, but I didn't think it would affect all of the machines in relatively the same way since some of them didn't lose any wu's.

After reading some of the give and take from Richard and Mr. Gee, I was wondering if changing the number of gpu units running at one time could affect the time it would take to complete a batch of wu's and maybe throw off the timing a bit. BUT....it seems to be a very large number of timed out units for that to be the case.

Thanks for all the info.... didn't know about ghosts....hmmmm

Sometimes when a host requests work the server selects tasks and assigns to the host, but the tasks never reason the host. So we call them "ghosts". The BOINC server does have an option "resend lost tasks". However SETI@home often disables that options because it can add extra strain to the servers. The only option at that point is having them timeout. Which isn't a problem as the server reassigns them.

As other mentioned some of your machines show more than 200 tasks in progress. SETI@home has a limit of 100 CPU tasks and then 100 per GPU. This can make it easy to tell if you have a system with ghosts tasks. Looking at your host 6335328. It shows In progress (224) if we select next until we get past the first 200. Then we see the sent date is January 21st to the 24th for the last 24 tasks. If you don't see those tasks on your system then they are ghosts & will time out in the next few days. Which means everything is basically running fine on your end.


Okay, got it. I will check it out and let you know what I find, but right now I'm guessing that you are right. Bed time for me. Thanks for the help.
12) Message boards : Number crunching : WU's timing out........ (Message 1771980)
Posted 16 Mar 2016 by Profile AllenIN
I see the imfamous "finish file present too long" error. We really must all gang up on David to get that one fixed sometime.


Over 10k V8 tasks done now (CPU and GPU), no invalids, 3 errors all of the above:-(



As Indiana Jones said, "Don't get cocky kid."....grin

I was in the same boat a few weeks back and then everything changed..grrrr!
13) Message boards : Number crunching : WU's timing out........ (Message 1771979)
Posted 16 Mar 2016 by Profile AllenIN
With the limits of 100 tasks per CPU/GPU & a deadline of about 6 weeks how many hours a day do you let your machines run tasks?

The only things I can think of are:
1) Adjust your queue size.
2) Stop fiddling with things & let BOINC handle running tasks instead of trying to micromanage what it does. BOINC will run tasks that are near the deadline sooner if it thinks there is a chance of missing them.

EDIT: A single core CPU machine that takes around 10 hours to complete a task would be able to complete a queue of 100 tasks before the deadline if it ran 24/7.


I run 24/7 on all of my machines and I didn't start "playing around" until Boinc was unable to finish all of the tasks it had at it's command. As I said, never had this problem before the change of versions.
Thanks for the input though.

Allen

Are you looking at the tasks in BOINC Manager and seeing they are close to their due date or are you looking at the error tasks list on the website and seeing "Timed out - no response"?
If you are seeing "Timed out - no response" in your errors those tasks are most likely from ghosts that were never on your system.

BOINC runs tasks FIFO (First In First Out) unless it thinks there will be a problem. Then it will run tasks in order by due date. How your cache settings are set can effect how BOINC determines this.


Wow! Now that is interesting. I did see the time outs that got me interested in doing something about it, but I never checked to see if they were actually ever on my system or at least present at some time. I just assumed (stupidly) that they were there and they were skipped for other wu's. They could have been ghosts as you said, since I didn't know anything about 'ghosts'.

I guess at one time or another I could have lost some wu's during the change over of versions, but I didn't think it would affect all of the machines in relatively the same way since some of them didn't lose any wu's.

After reading some of the give and take from Richard and Mr. Gee, I was wondering if changing the number of gpu units running at one time could affect the time it would take to complete a batch of wu's and maybe throw off the timing a bit. BUT....it seems to be a very large number of timed out units for that to be the case.

Thanks for all the info.... didn't know about ghosts....hmmmm
14) Message boards : Number crunching : WU's timing out........ (Message 1771975)
Posted 16 Mar 2016 by Profile AllenIN
The only real change that I have made since v7, is that I upgraded all of my machines to the latest Boinc version. Maybe that's the problem.

"The latest"? Could we have an actual number on that, please? (Yes, I know I can look it up - v7.6.22 - but you might have meant you're testing v7.6.29). In general, it's better to use absolute data than relative terms which might get overtaken by events.

Speaking of which, could you post the actual numbers you're using for cache size settings - both of them, please. I personally didn't have any problem with BOINC v7.6.22 maintaining the cache sizes I requested, but that tends to be an absolute maximum of 1 day, more often 0.4 or 0.5 days.


Hi Richard. Sorry, I am running v7.6.22 on all of my computers except the tablet. I have quads with gpu's on 3 of them and one dual with gpu and one without. (It's a laptop.)

I have it set for 10 and 10 since I hate to run out when Seti goes down for awhile. I've never had trouble with it running newer wu's and skipping over the one's that are due soon and that's why I asked if anyone else was having this problem.

I thought that Boinc or Seti would decide how many wu's of whichever kind I should get to be sure to be able to accomplish finishing them on time. Usually I never mess with them, but when I saw so many of them being timed out, I thought I would try to fix it. It did get better for a day, but then it would grab units that weren't due for weeks instead of one that was due in a couple of days and then it would be late eventually.

Thanks for helping!
15) Message boards : Number crunching : WU's timing out........ (Message 1771853)
Posted 16 Mar 2016 by Profile AllenIN
If you are still actively crunching Seti Beta as stock there, there have been some hints there *Might* be some problems with multiple project resource share management, to be further investigated. [possibly further complicated by Einstein fixed credit]. If so, selecting among your hosts to dedicate to given projects (setting no new tasks on the others) *might* let things recover.

No, *actual* credit (fixed or otherwise) has no effect on resource share - only the simplified REC is used.

But Einstein has much shorter deadlines than here, and that can play badly with very large cache settings. Although, since he doesn't seem to have actually crunched for Einstein for over 2 months (last contact 10 January), that's unlikely to be the cause currently.


Thanks Richard. As you stated, no Einstein for quite awhile. I can obviously change the size of my cache, but I thought Boinc was suppose to figure out how many wu's you need based on how long it takes you to run wu's, so I've just been leaving it up to Boinc to decide.

Thanks again, I just can't figure out what is wrong. The only real change that I have made since v7, is that I upgraded all of my machines to the latest Boinc version. Maybe that's the problem.

Allen
16) Message boards : Number crunching : WU's timing out........ (Message 1771852)
Posted 16 Mar 2016 by Profile AllenIN
Which of your computers is having this problem?
One of them ( http://setiathome.berkeley.edu/results.php?hostid=6335328 ) has a single GPU and over two hundred tasks, which means either it has a lot of "ghosts", or you've been playing with re-scheduling, either of which can give this problem.
First thing to try on that computer is to "detach" wait a couple of minutes then "re-attach" to the project. This will help to kill off ghosts, and it will have the "real" tasks list.
Secondly, as others have said, don't micro-manage, BOINC does very well when left to do its own thing, but tends to make a mess when we fiddle.


As I said in the previous note, I didn't try to micro manage until Boinc was unable to monitor things correctly and run the older wu's first. I think it pays to much attention to which units are completed by other people than just doing them by their due dates. I didn't just jump in and start making changes, it has been failing to get the job done for some time now. As for the single CPU machine, even the tablet has more than one core and the real systems have at least two.

Thanks,

Allen
17) Message boards : Number crunching : WU's timing out........ (Message 1771851)
Posted 16 Mar 2016 by Profile AllenIN
With the limits of 100 tasks per CPU/GPU & a deadline of about 6 weeks how many hours a day do you let your machines run tasks?

The only things I can think of are:
1) Adjust your queue size.
2) Stop fiddling with things & let BOINC handle running tasks instead of trying to micromanage what it does. BOINC will run tasks that are near the deadline sooner if it thinks there is a chance of missing them.

EDIT: A single core CPU machine that takes around 10 hours to complete a task would be able to complete a queue of 100 tasks before the deadline if it ran 24/7.


I run 24/7 on all of my machines and I didn't start "playing around" until Boinc was unable to finish all of the tasks it had at it's command. As I said, never had this problem before the change of versions.
Thanks for the input though.

Allen
18) Message boards : Number crunching : WU's timing out........ (Message 1771772)
Posted 15 Mar 2016 by Profile AllenIN
I've been having a lot of WU's that run out of time before they are run. I don't remember this being a problem in the past.

Is anyone else having this problem??

If not, is there anything I can do to fix it? I've tried suspending work until the older ones are through running but this is very time consuming and sometimes I don't get back in time to resume them and nothing gets done during that time.

Thanks for any help you can offer.

Allen
19) Message boards : Number crunching : Anything relating to AstroPulse tasks (Message 1766738)
Posted 21 Feb 2016 by Profile AllenIN
I have 38 AP's across 4 machines, but I accept any kind of work.
20) Message boards : Number crunching : Anything relating to AstroPulse tasks (Message 1763642)
Posted 9 Feb 2016 by Profile AllenIN
-use_sleep -unroll 10 -oclFFT_plan 256 16 512 -ffa_block 6144 -ffa_block_fetch 1536


Jim try these. Got this from the old 750Ti thread

http://setiathome.berkeley.edu/forum_thread.php?id=75919#1588675

this as the last recommendation on a 750Ti

-use_sleep -unroll 6 -oclfft_plan 256 16 256 -ffa_block 2830 -ffa_block_fetch 2830 -tune 1 64 4 1 -tune 2 64 4 1

See if it helps


I was just wondering, would the same commands work on a GTX650ti or are they specific to only the 750?


Try -use_sleep -unroll 8 -oclfft_plan 256 16 256 -ffa_block 2830 -ffa_block_fetch 2830 -tune 1 64 4 1 -tune 2 64 4 1.


Thanks Mike, I'll give it a try.


Next 20

Copyright © 2016 University of California