Advanced search

Message boards : Number crunching : Expired but still stucked WU in Boinc manager

Author Message
jlhal
Send message
Joined: 1 Mar 10
Posts: 147
Credit: 1,077,535,540
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 24848 - Posted: 9 May 2012 | 21:21:33 UTC

Hi, the WU here after is still showing in Boinc manager as "upload in progress", but is expired since 2 of May because it never began to upload when finished:

5295002 3376731 27 Apr 2012 | 3:09:33 UTC 2 May 2012 | 3:09:33 UTC Délai expiré - aucune réponse 0.00 0.00 --- Long runs (8-12 hours on fastest card) v6.16 (cuda31)

Translated:
5295002 3376731 27 Apr 2012 | 3:09:33 UTC 2 May 2012 | 3:09:33 UTC Expired delay - no answer 0.00 0.00 --- Long runs (8-12 hours on fastest card) v6.16 (cuda31)

What can I do in order to remove it from Boinc display ?
I rebooted more than once since, but no effect.
Thanks for your time...
____________
Lubuntu 16.04.1 LTS x64

jlhal
Send message
Joined: 1 Mar 10
Posts: 147
Credit: 1,077,535,540
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 24888 - Posted: 10 May 2012 | 18:04:49 UTC - in response to Message 24848.

Any idea/tip ?
____________
Lubuntu 16.04.1 LTS x64

Profile skgiven
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 24890 - Posted: 10 May 2012 | 18:55:06 UTC - in response to Message 24888.

Go to Transfers, and if it is listed there click Retry Now.
If it's not there Abort it.
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

jlhal
Send message
Joined: 1 Mar 10
Posts: 147
Credit: 1,077,535,540
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 24896 - Posted: 10 May 2012 | 19:31:06 UTC - in response to Message 24890.
Last modified: 10 May 2012 | 19:33:40 UTC

Go to Transfers, and if it is listed there click Retry Now.
If it's not there Abort it.


Hi skgiven !
Already tried what you are suggesting, but no success.
.This WU was never seen in transfers...
When I click 'abort' , nothing happens.
I've been thinking about looking in the files describing the WUs but I'm afraid to touch anything down there...
____________
Lubuntu 16.04.1 LTS x64

Profile skgiven
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 24990 - Posted: 12 May 2012 | 12:56:49 UTC - in response to Message 24896.

Maybe remove the GPUGrid project, delete the GPUGrid folder and then re-attach.
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

jlhal
Send message
Joined: 1 Mar 10
Posts: 147
Credit: 1,077,535,540
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25062 - Posted: 14 May 2012 | 8:12:24 UTC - in response to Message 24990.
Last modified: 14 May 2012 | 8:21:30 UTC

Maybe remove the GPUGrid project, delete the GPUGrid folder and then re-attach.


Resetting the project , solved this problem.

Note : Boinc 7.0.27 Beta (x86) on Linux x64 seems to run fine with 295.40 driver delivered with Precise (Xubuntu) 12.04 ONLY for ACEMD2 standard and Long Runs.
Until now betas all fail. Will try with newer driver 295.49 soon.
____________
Lubuntu 16.04.1 LTS x64

Dagorath
Send message
Joined: 16 Mar 11
Posts: 509
Credit: 179,005,236
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25064 - Posted: 14 May 2012 | 8:38:51 UTC - in response to Message 25062.

The 295.40 driver is working for me on Kubuntu 12.04 for Beta tasks. I downloaded the driver from NVIDIA and installed it myself. Are you sure Xubuntu installs the correct driver? Maybe it installs the proper driver but doesn't setup the driver correctly?

jlhal
Send message
Joined: 1 Mar 10
Posts: 147
Credit: 1,077,535,540
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25075 - Posted: 14 May 2012 | 17:25:42 UTC - in response to Message 25064.

The 295.40 driver is working for me on Kubuntu 12.04 for Beta tasks. I downloaded the driver from NVIDIA and installed it myself. Are you sure Xubuntu installs the correct driver? Maybe it installs the proper driver but doesn't setup the driver correctly?


Hi Dagorath !
What kind of setup parameters are you talking about ?
Can you be more...precise ;-) ?

____________
Lubuntu 16.04.1 LTS x64

Dagorath
Send message
Joined: 16 Mar 11
Posts: 509
Credit: 179,005,236
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25079 - Posted: 14 May 2012 | 19:08:40 UTC - in response to Message 25075.
Last modified: 14 May 2012 | 19:09:49 UTC

I'm not an expert on GPU computing so I am not sure what all the setup parameters are or where to locate them. I know that when I install the NVIDIA driver manually it makes big changes to /etc/X11/xorg.conf so my guess is many of the setup parameters are in that file. There may be 1000 other parameters stored in other files that I don't know about.

Here is my xorg.conf immediately after installing Kubuntu and before manually installing the NVIDIA driver:

Section "Device"
Identifier "Default Device"
Option "NoLogo" "True"
EndSection



Here is my xorg.conf after installing the driver manually. Do not replace your xorg.conf with mine as I can almost guarantee it won't work. There is only 1 line from my xorg.conf you should consider using and that is the red line. More on the red line below:

# nvidia-xconfig: X configuration file generated by nvidia-xconfig
# nvidia-xconfig: version 295.40 (buildmeister@swio-display-x86-rhel47-06.nvidia.com) Thu Apr 5 22:40:54 PDT 2012

Section "ServerLayout"
Identifier "Layout0"
Screen 0 "Screen0"
InputDevice "Keyboard0" "CoreKeyboard"
InputDevice "Mouse0" "CorePointer"
EndSection

Section "Files"
EndSection

Section "InputDevice"
# generated from default
Identifier "Mouse0"
Driver "mouse"
Option "Protocol" "auto"
Option "Device" "/dev/psaux"
Option "Emulate3Buttons" "no"
Option "ZAxisMapping" "4 5"
EndSection

Section "InputDevice"
# generated from default
Identifier "Keyboard0"
Driver "kbd"
EndSection

Section "Monitor"
Identifier "Monitor0"
VendorName "Unknown"
ModelName "Unknown"
HorizSync 28.0 - 33.0
VertRefresh 43.0 - 72.0
Option "DPMS"
EndSection

Section "Device"
Identifier "Device0"
Driver "nvidia"
VendorName "NVIDIA Corporation"
Option "Coolbits" "4"
EndSection

Section "Screen"
Identifier "Screen0"
Device "Device0"
Monitor "Monitor0"
DefaultDepth 24
SubSection "Display"
Depth 24
EndSubSection
EndSection


I added the Coolbits option myself. Coolbits allows manual control of the fan speed. I was forced to use manual control because my GPU doesn't increase the fan speed enough automatically to keep the GPU at the low temperature I want. The auto fan control allows the GPU to reach 90 C which is much too hot IMHO.

I learned about Coolbits and other interesting stuff from the Gentoo wiki .

Profile skgiven
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25086 - Posted: 14 May 2012 | 22:42:56 UTC - in response to Message 25079.

So you need,
Option "Coolbits" "4" for Kubuntu
rather than Option "Coolbits" "5" as in Ubuntu?
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Dagorath
Send message
Joined: 16 Mar 11
Posts: 509
Credit: 179,005,236
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25090 - Posted: 15 May 2012 | 1:38:06 UTC - in response to Message 25086.
Last modified: 15 May 2012 | 1:40:37 UTC

I'm not sure. I just saw Coolbits 5 recommended somewhere the other day for the first time and I passed it off as a typo. I'm using Coolbits 4 on Kubuntu and have never tried Coolbits 5. The Gentoo wiki article I mentioned on my previous post even mentions Coolbits 1 as being required for something to do with unlocking the clock settings though you put Coolbits 1 in a different section of xorg.conf, according to Gentoo wiki.

Maybe Coolbits 4 allows manual fan speed settings and 5 allows all that plus OCing? And just 1 allows OCing but not manual fan speed adjusts? I also find it odd that it won't allow me to set the fan speed % to just any number from 1 to 100, only 44 and 85 are accepted. If I use anything between 86 and 100 it falls back to 85. Maybe Coolbits 5 will allow more fan speed values? I dunno but now I'm curious.

I'll see if it will allow me to adjust a clock with Coolbits 4. If not then I'll try Coolbits 5. I'm not familiar with OCing so which setting(s) would you recommend as the one(s) to try first and how much should I adjust it? Currently I'm at:

Graphics: 742 MHz
Memory: 1900 MHz
Processor: 1484 MHz

Dagorath
Send message
Joined: 16 Mar 11
Posts: 509
Credit: 179,005,236
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25091 - Posted: 15 May 2012 | 5:36:12 UTC - in response to Message 25090.

Here's the doc that explains everything for the 295.40 driver for Linux. If it's not in that doc then it's not worth knowing ;) Change the 295.40 in the URL to match the driver you are using.

This page from that doc explains the Coolbits option and it turns out my wild guess was right. Coolbits 4 allows manual fan speed control and Coolbits 1 allows clocks adjustments. The integer argument to Coolbits is a bit mask so Coolbits 5 allows both clocks and manual fan control so 5 is probably what most crunchers want to use. BTW, Coolbits 2 attempts to setup SLI between cards with different amounts of RAM if anybody has run into that problem.

In addition to putting Coolbits 5 in xorg.conf:

1) To allow manual fan control you also need to start the nvidia-settings utility, go to the Thermal Settings page and check the Enable GPU Fan Settings box. Then you can set the fan speed manually. The setting does not stick between X sessions for my Asus card so I have to put the following statements in a startup script.

nvidia-settings -a "[gpu:0]/GPUFanControlState=1"
nvidia-settings -a "[fan:0]/GPUCurrentFanSpeed=85"


2) To allow clock adjustments, the Gentoo wiki says use the following statement to adjustment clocks:

nvidia-settings --assign "[gpu:0]/GPUOverclockingState=1" \
--assign "[gpu:0]/GPU2DClockFreqs=<gpu clock>,<mem clock>" \
--assign[color=red][b]=[/b][/color]"[gpu:0]/GPU3DClockFreqs=<gpu clock>,<mem clock>" &


I am quite sure the red equal sign is a typo. The single statement above is actually 3 statements rolled into 1. Broken into the 3 components and using -a (the abbreviation for --assign) we have:

nvidia-settings -a "[gpu:0]/GPUOverclockingState=1"
nvidia-settings -a "[gpu:0]/GPU2DClockFreqs=<gpu clock>,<mem clock>"
nvidia-settings -a "[gpu:0]/GPU3DClockFreqs=<gpu clock>,<mem clock>"


Of course you must substitute integers for <gpu clock> and <mem clock>.

Gentoo wiki says the clock settings do not stick between X sessions so again you need to put statements for overclocking into a startup script.

Does any of this actually work? Well, the manual fan speed adjustment works for me on my Asus GTX570 with the limitation that I can set the fan speed no higher than 85%. The clock adjustments don't work at all for me, maybe I need to RTFM again. Maybe clock adjustments will work for you. Maybe Asus has disabled clock adjustment to the point where I need to reprogram the BIOS.

Profile skgiven
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25095 - Posted: 15 May 2012 | 10:53:19 UTC - in response to Message 25091.

Never used a startup script for coolbits, but using coolbits 5 allows me to set the fan value to any number in Ubuntu (56%, 72, 81...). I found 77% keeps my GTX470 below 70°C whatever the task is (usually ~66°C). I did try other values, in addition to 5 but I have not been able to increase GPU clocks. Maybe I put them in the wrong place. Not really an issue for me as the card doesn't OC very well.
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Dagorath
Send message
Joined: 16 Mar 11
Posts: 509
Credit: 179,005,236
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25101 - Posted: 15 May 2012 | 23:30:17 UTC - in response to Message 25095.

I don't use the Coolbits option in a startup script either, if that's what you mean. The Coolbits statement definitely has to be inserted into /etc/X11/xorg.conf. Statements for setting fan and/or clock speed go into startup scripts.

Still no luck adjusting clocks here. I think I need to go to Asus forums and ask about setting clocks and fan speed.

@jlhal,

Sorry about hijacking your thread. Let us know how you're making out.

jlhal
Send message
Joined: 1 Mar 10
Posts: 147
Credit: 1,077,535,540
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25105 - Posted: 16 May 2012 | 5:51:18 UTC - in response to Message 25101.


@jlhal,

Sorry about hijacking your thread. Let us know how you're making out.


You're welcome ...
Will make a snapshot of my Nvidia X-server window setingd to let you see what is available without doing anything myself ...
____________
Lubuntu 16.04.1 LTS x64

jlhal
Send message
Joined: 1 Mar 10
Posts: 147
Credit: 1,077,535,540
RAC: 0
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25110 - Posted: 16 May 2012 | 16:55:28 UTC - in response to Message 25105.


Will make a snapshot of my Nvidia X-server window setingd to let you see what is available without doing anything myself ...


OOPS ! I know this is a silly newbie question but...

How do I insert an image (png) in a message ?
(I want to do it in a new thread in GPU CARDS main thread.)
[/b]
____________
Lubuntu 16.04.1 LTS x64

Profile skgiven
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 25111 - Posted: 16 May 2012 | 18:10:40 UTC - in response to Message 25110.
Last modified: 16 May 2012 | 18:12:39 UTC

The image needs to be online somewhere.

Put the link in your post, select the link and click Img from the above BBCode tags.
Alternatively use [img ] url [/img ], without the spaces.
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Post to thread

Message boards : Number crunching : Expired but still stucked WU in Boinc manager

//