Open Beta test: SoG for NVidia, Lunatics v0.45 - Beta6 (RC again)

Message boards : Number crunching : Open Beta test: SoG for NVidia, Lunatics v0.45 - Beta6 (RC again)
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 7 · 8 · 9 · 10 · 11 · 12 · 13 . . . 32 · Next

AuthorMessage
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14679
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1795097 - Posted: 10 Jun 2016, 15:44:21 UTC

The installer is designed to shut down the BOINC client - the bit that runs silently in the background - by itself. If it's running when the installer starts, you should get a warning that it is going to need to do that, and give you the opportunity to cancel.

It doesn't matter to the Installer if BOIMC Manager, BOINC Tasks, or any other user interface is running - nothing the installer does matters to Manager-like tools. In fact, the plan was to leave the Manager running, but close the client and start it back afterwards. If anyone sees anything different to that, please describe with as much detail as you can.

The only change between 0.44 and 0.45 in this area is the detection or 32-bit and 64-bit OSes - I might have broken something there.
ID: 1795097 · Report as offensive
Profile Bernie Vine
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 26 May 99
Posts: 9958
Credit: 103,452,613
RAC: 328
United Kingdom
Message 1795311 - Posted: 11 Jun 2016, 7:27:31 UTC
Last modified: 11 Jun 2016, 7:28:24 UTC

Well my slow but sure rise in RAC continues:



Just so you know it is this machine:

6851722

AuthenticAMD
AMD A8-3800 APU with Radeon(tm) HD Graphics [Family 18 Model 1 Stepping 0]
(4 processors)

NVIDIA GeForce GTX 970 (4095MB) driver: 368.22 OpenCL: 1.2

I run two tasks on the 970 and have two CPU cores free.

However it is also my daily machine and I am typing this on it. I watch streaming videos for several hours a day with no noticeable problems. I play games for around 4 hours per day and then Boinc is off.

Just for a historical perspective during last years WOW event this machine peaked at 15,799 RAC, dipped to 3,899 on the 22nd Jan and has not gone higher than 11,458 since.

Before this upturn the lowest point was 6,459. current RAC is 8969

Lets hope it continues.
ID: 1795311 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13854
Credit: 208,696,464
RAC: 304
Australia
Message 1795314 - Posted: 11 Jun 2016, 7:54:06 UTC - in response to Message 1795311.  

However it is also my daily machine and I am typing this on it. I watch streaming videos for several hours a day with no noticeable problems. I play games for around 4 hours per day and then Boinc is off.

Are you running with the defaults or have to made use of -use sleep?
Grant
Darwin NT
ID: 1795314 · Report as offensive
Rasputin42
Volunteer tester

Send message
Joined: 25 Jul 08
Posts: 412
Credit: 5,834,661
RAC: 0
United States
Message 1795342 - Posted: 11 Jun 2016, 11:52:48 UTC

I have noticed, that the 64 bit installer shows the warning about boinc running, even if it is not.
ID: 1795342 · Report as offensive
Profile Bernie Vine
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 26 May 99
Posts: 9958
Credit: 103,452,613
RAC: 328
United Kingdom
Message 1795345 - Posted: 11 Jun 2016, 12:13:34 UTC - in response to Message 1795314.  

However it is also my daily machine and I am typing this on it. I watch streaming videos for several hours a day with no noticeable problems. I play games for around 4 hours per day and then Boinc is off.

Are you running with the defaults or have to made use of -use sleep?

Sorry, yes forgot to mention the "-use sleep" is the only command I currently use.
ID: 1795345 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14679
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1795347 - Posted: 11 Jun 2016, 12:23:41 UTC - in response to Message 1795342.  
Last modified: 11 Jun 2016, 12:46:11 UTC

I have noticed, that the 64 bit installer shows the warning about boinc running, even if it is not.

Thanks, I'll go down and check that out now.

Edit: well, I've learned something new - but not what I expected.

At my first test attempt, I didn't get the warning at all, whether BOINC was running or not. Then I remembered that I had set that machine up to 'allow_multiple_clients': with that disabled, detection behaved as planned (warned that BOINC was running if it was, didn't warn if it wasn't). And having looked at the code, I can see why it might not work with multiple clients allowed - even if only one is running. The installer makes no attempt to enumerate and stop multiple clients - if you use that (somewhat obscure) facility, take care with the installer.

My test was with a 64-bit copy of Windows 7: could you describe the system you see this false detection on, please?

And please can you use a third-party tool, like Task Manager or Process Explorer, to check that BOINC is really stopped when you think it is? You would normally expect to see three BOINC components listed when it is fully up and running:

boinc.exe
boincmgr.exe
boinctray.exe

It's only the first one of those that we're testing: boincmgr and boinctray can stay running, as far as the installer is concerned. Let us know what you see in both 'running' and 'stopped' states.
ID: 1795347 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 66332
Credit: 55,293,173
RAC: 49
United States
Message 1795409 - Posted: 11 Jun 2016, 16:32:19 UTC

Of the three wu's I'm working on 2 are having trouble(I'm trying 0.45 again with 0.5/0.1 as suggested), the time left does not shrink at all, processing slows to a crawl, when I stop boinc, all 4 of the cpu wu's and 1 of 3 gpu's stop, the other 2 gpu wu's stay in Memory and continue, plus the cpu% keep dropping.

I have the cmdline set for: -use_sleep -instance_per_device 3
I've also had the cmdline set for 2 also, makes no difference.

Here's two screen grabs, note the time left, on 2 they don't change...


Savoir-Faire is everywhere!
The T1 Trust, T1 Class 4-4-4-4 #5550, America's First HST

ID: 1795409 · Report as offensive
Rasputin42
Volunteer tester

Send message
Joined: 25 Jul 08
Posts: 412
Credit: 5,834,661
RAC: 0
United States
Message 1795414 - Posted: 11 Jun 2016, 17:00:58 UTC

My test was with a 64-bit copy of Windows 7: could you describe the system you see this false detection on, please?


I redid it with no boinc(related) processes running---fine, no warning.
With just boinc.exe running---warning---correct..

I have another machine running boinc( older version), which i have not considered.
That appears to be fine.
I can log onto it, if necessary, but the link is bad, and i rarely connect to it.
ID: 1795414 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14679
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1795416 - Posted: 11 Jun 2016, 17:12:08 UTC - in response to Message 1795414.  

That's OK - I think we can put it down to a random failure to shut down boinc.exe, without spending too much time on it. But now you've alerted everyone who reads this thread to the possibility, I'm sure other people will have tested as well (and they know what to look out for).

If anyone reproduces the problem (consistently), we can follow it up then.
ID: 1795416 · Report as offensive
Rasputin42
Volunteer tester

Send message
Joined: 25 Jul 08
Posts: 412
Credit: 5,834,661
RAC: 0
United States
Message 1795418 - Posted: 11 Jun 2016, 17:18:07 UTC - in response to Message 1795416.  

Thanks for looking into it.
I am 90% certain, that boinc was fully shut down, but there are still these 10%...
ID: 1795418 · Report as offensive
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5517
Credit: 528,817,460
RAC: 242
United States
Message 1795419 - Posted: 11 Jun 2016, 17:18:31 UTC - in response to Message 1795409.  

With the use of -use_sleep the percentage of CPU usage should continue to drop as the work continues to be processed.

The failure of time left to change, might be an issue with the way the program works. Somewhere I seem to remember comments about the non-linear countdown of work units. I don't know if that is what is happening but might explain it.

As to why the work units continue to crunch after you have exited? I think someone had commented that with BOINCtasks, there was a command that needed to be checked to stop all work when exiting. Failing that, I think I remember something like that happening a while back.

Are you using an up to date version of BOINCtasks? Just trying to figure out what is happening.
ID: 1795419 · Report as offensive
Rasputin42
Volunteer tester

Send message
Joined: 25 Jul 08
Posts: 412
Credit: 5,834,661
RAC: 0
United States
Message 1795421 - Posted: 11 Jun 2016, 17:24:22 UTC

As to why the work units continue to crunch after you have exited? I think someone had commented that with BOINCtasks, there was a command that needed to be checked to stop all work when exiting. Failing that, I think I remember something like that happening a while back.


Correct.
There is a menu item to stop the client.
If you just exit "BoincTasks", the client remains running.
ID: 1795421 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14679
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1795427 - Posted: 11 Jun 2016, 17:41:07 UTC - in response to Message 1795421.  

As to why the work units continue to crunch after you have exited? I think someone had commented that with BOINCtasks, there was a command that needed to be checked to stop all work when exiting. Failing that, I think I remember something like that happening a while back.

Correct.
There is a menu item to stop the client.
If you just exit "BoincTasks", the client remains running.

Both BOINC Manager and BoincTasks have that behaviour, but in both cases, they either stop boinc.exe (and all tasks should stop with it), or they leave everything running.

Vic wrote:
all 4 of the cpu wu's and 1 of 3 gpu's stop, the other 2 gpu wu's stay in Memory and continue

That shouldn't happen with either Manager. It sounds more like the applications themselves failing to notice that BOINC wasn't there any more - something we had problems with a few years ago.
ID: 1795427 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 66332
Credit: 55,293,173
RAC: 49
United States
Message 1795434 - Posted: 11 Jun 2016, 18:15:52 UTC - in response to Message 1795419.  
Last modified: 11 Jun 2016, 18:42:51 UTC

With the use of -use_sleep the percentage of CPU usage should continue to drop as the work continues to be processed.

The failure of time left to change, might be an issue with the way the program works. Somewhere I seem to remember comments about the non-linear countdown of work units. I don't know if that is what is happening but might explain it.

As to why the work units continue to crunch after you have exited? I think someone had commented that with BOINCtasks, there was a command that needed to be checked to stop all work when exiting. Failing that, I think I remember something like that happening a while back.

Are you using an up to date version of BOINCtasks? Just trying to figure out what is happening.

I'm using 1.69 which is the latest BoincTasks and BT is set to have all work stop when Boinc exits.
Note: When Boinc was set to go to sleep as scheduled, all work ceased at the appointed time.

At the time of this post I was using Boinc v7.6.22, now I'm using v7.6.33, during the upgrade Boinc shut down, but 2 gpu wu's kept going, even after I put boinc to sleep, so I ended the programs, then restarted Boinc, then the 2 had lost ground and would now go to sleep.
Savoir-Faire is everywhere!
The T1 Trust, T1 Class 4-4-4-4 #5550, America's First HST

ID: 1795434 · Report as offensive
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5517
Credit: 528,817,460
RAC: 242
United States
Message 1795450 - Posted: 11 Jun 2016, 19:36:37 UTC - in response to Message 1795434.  
Last modified: 11 Jun 2016, 19:37:29 UTC

Zoom,

Just a thought.

Do you run ccleaner after your installs?

Sometimes things get left in the registry after updating and that can cause issues.

Just like what we do with clean install of drivers.

Might be worth a shot.

https://www.piriform.com/ccleaner/download


Zalster
ID: 1795450 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 66332
Credit: 55,293,173
RAC: 49
United States
Message 1795461 - Posted: 11 Jun 2016, 21:34:47 UTC - in response to Message 1795450.  
Last modified: 11 Jun 2016, 21:36:43 UTC

Zoom,

Just a thought.

Do you run ccleaner after your installs?

Sometimes things get left in the registry after updating and that can cause issues.

Just like what we do with clean install of drivers.

Might be worth a shot.

https://www.piriform.com/ccleaner/download


Zalster

No.

What Richard Haslegrove said sounds plausible.
Savoir-Faire is everywhere!
The T1 Trust, T1 Class 4-4-4-4 #5550, America's First HST

ID: 1795461 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1795540 - Posted: 12 Jun 2016, 3:45:15 UTC - in response to Message 1795419.  

With the use of -use_sleep the percentage of CPU usage should continue to drop as the work continues to be processed.

The failure of time left to change, might be an issue with the way the program works. Somewhere I seem to remember comments about the non-linear countdown of work units. I don't know if that is what is happening but might explain it.

As to why the work units continue to crunch after you have exited? I think someone had commented that with BOINCtasks, there was a command that needed to be checked to stop all work when exiting. Failing that, I think I remember something like that happening a while back.

Are you using an up to date version of BOINCtasks? Just trying to figure out what is happening.


. . I get a pop up window when exiting BOINC client confirming to stop running tasks.
ID: 1795540 · Report as offensive
Stephen "Heretic" Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Sep 12
Posts: 5557
Credit: 192,787,363
RAC: 628
Australia
Message 1795542 - Posted: 12 Jun 2016, 3:53:44 UTC - in response to Message 1794966.  


When 0.44 was installed, SoG went away, so did the 16 Sog wu's which were replaced with 18 guppis, the 16 were also guppis.


. . I think you made the same mistake I did, when "rolling back" from 0.45 running SoG I did an uninstall, this is why the downloaded "SoG" WUs went away, if I had done what I should have done and simply re-run the 0.45 install but selecting CUDA50 instead of SoG things would have been fine.

. . Another brilliant "D'oh!" moment.
ID: 1795542 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 66332
Credit: 55,293,173
RAC: 49
United States
Message 1795551 - Posted: 12 Jun 2016, 5:43:11 UTC

Still getting 'waiting to run, now with 4000+ day run times...

Savoir-Faire is everywhere!
The T1 Trust, T1 Class 4-4-4-4 #5550, America's First HST

ID: 1795551 · Report as offensive
Profile Brent Norman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 1 Dec 99
Posts: 2786
Credit: 685,657,289
RAC: 835
Canada
Message 1795553 - Posted: 12 Jun 2016, 5:52:23 UTC

Set your cache to 1 day and just let it run.

You have changed apps so much BOINC has no idea what the average times should be.

It will run high priority until it sorts it out.
ID: 1795553 · Report as offensive
Previous · 1 . . . 7 · 8 · 9 · 10 · 11 · 12 · 13 . . . 32 · Next

Message boards : Number crunching : Open Beta test: SoG for NVidia, Lunatics v0.45 - Beta6 (RC again)


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.