Advanced search

Message boards : Number crunching : Discussion of Ubuntu 16.04-x64 LTS installation and configuration

Author Message
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 43504 - Posted: 23 May 2016 | 14:17:08 UTC
Last modified: 23 May 2016 | 14:18:09 UTC

This thread is to allow discussion of the following (locked) FAQ thread,

How to - install Ubuntu 16.04 x64 Linux & setup for GPUGrid

Post here about Ubuntu 16.04-x64 LTS installation and configuration.
Create a new thread for other distros, please.
____________
FAQ's

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

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1112
Credit: 6,162,416,256
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 43505 - Posted: 23 May 2016 | 14:20:33 UTC - in response to Message 43504.

This thread is to allow discussion of the following (locked) FAQ thread,

How to - install Ubuntu 16.04 x64 Linux & setup for GPUGrid

Post here about Ubuntu 16.04-x64 LTS installation and configuration.
Create a new thread for other distros, please.

Thank you VERY MUCH for doing this guide! It's greatly appreciated!

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 43506 - Posted: 23 May 2016 | 15:04:32 UTC - in response to Message 43505.

You are welcome and I hope you find it useful.

Feedback, experiences and discussion would be appreciated.
Will update the FAQ with any augmentations/corrections as needed.
____________
FAQ's

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

klepel
Send message
Joined: 23 Dec 09
Posts: 189
Credit: 4,195,561,293
RAC: 1,590,531
Level
Arg
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 43955 - Posted: 13 Jul 2016 | 22:39:45 UTC

I think, this is the more appropriate thread for the discussion I started over in COOLBITS as LUBUNTU and LINUX MINT are UBUNTU derivates.

My computer with the LUBUNTU 16.04 64-bit installation runs great for over 2 weeks now.

I installed LINUX MINT 18 64-bit on the computer with LINUX MINT 17.3 32-bit previously. Beside that the video card is too old to run GPUGRID, everything works very smooth.

So, I would like to say, that it is quite easy to use Linux for crunching with Nvidia GPUs since the UBUNTU 16.04 LTS release.

Matt
Avatar
Send message
Joined: 11 Jan 13
Posts: 216
Credit: 846,538,252
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44015 - Posted: 18 Jul 2016 | 21:47:05 UTC

THANK YOU so much for doing this.

Would you recommend a USB 3.0 drive, or is a 2.0 drive sufficient?

klepel
Send message
Joined: 23 Dec 09
Posts: 189
Credit: 4,195,561,293
RAC: 1,590,531
Level
Arg
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44017 - Posted: 19 Jul 2016 | 3:50:14 UTC

If you use the USB-stick for installation on a Hard Disk it is not so important if you use USB 2.0 or 3.0, although the latter is noticeable faster during installation or running LINUX from the USB-stick (The start-up process is much slower on the USB 2.0).

However I do recommend that you use an USB-stick of at least 16 GB for installation only and 32 GB if you like to run LINUX Mint from the stick itself. The best thing is to use a spare/cheap one, as I have burned several in the process, but I am not a computer or Linux specialist at all.

I always used Multiboot Yumi http://www.pendrivelinux.com/yumi-multiboot-usb-creator for the primary installation and then installed the LINUX Distro to second destination from there to either USB-stick or the Hard Disk.

Matt
Avatar
Send message
Joined: 11 Jan 13
Posts: 216
Credit: 846,538,252
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44028 - Posted: 21 Jul 2016 | 16:40:43 UTC - in response to Message 44017.
Last modified: 21 Jul 2016 | 17:25:49 UTC

If you use the USB-stick for installation on a Hard Disk it is not so important if you use USB 2.0 or 3.0, although the latter is noticeable faster during installation or running LINUX from the USB-stick (The start-up process is much slower on the USB 2.0).

However I do recommend that you use an USB-stick of at least 16 GB for installation only and 32 GB if you like to run LINUX Mint from the stick itself.


I have a 32GB USB 3.0 stick to which I was hoping to install Linux. Not run as a live distro (I'm assuming one cannot create, edit, and save persistent files within a live copy - is this incorrect?) or install from, but install Linux to the stick such that when the stick is plugged in, I will have a fully installed copy of Ubuntu, but when the stick is not plugged in, the PC will boot to Windows as normal from the internal drive.

After reading through skgiven's How To, I'm not sure whether this is the scenario in mind or not.

EDIT: I've tried twice to get Ubuntu up and running using UUI and also Rufus (as recommended on Ubuntu's site). Both times have resulted in absolutely nothing happening after the initial Install/Live selection screen. I would just get 20 minutes of black screen and the indicator light on the USB would not be flashing either. After 20 minutes I gave up - figured something should have loaded by then if it was working as it should. I hope I'm doing something wrong, but can't figure out what it would be.

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 44030 - Posted: 22 Jul 2016 | 10:45:22 UTC - in response to Message 44028.

My intention was to cover a basic/standard installation to an internal drive, rather than an external USB stick to create a multi-boot environment.
Overall, I found the OS and NV driver installations very easy but that might not be the case if you need to multi-boot &/or boot to a USB stick.

In your case I don't know what's going wrong, but you may need to select the drive you are booting to when you restart or remove the other USB drive (assuming you are installing the OS from one USB stick onto the other)?
____________
FAQ's

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

Matt
Avatar
Send message
Joined: 11 Jan 13
Posts: 216
Credit: 846,538,252
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44041 - Posted: 24 Jul 2016 | 22:38:49 UTC - in response to Message 44030.

My intention was to cover a basic/standard installation to an internal drive, rather than an external USB stick to create a multi-boot environment.
Overall, I found the OS and NV driver installations very easy but that might not be the case if you need to multi-boot &/or boot to a USB stick.

In your case I don't know what's going wrong, but you may need to select the drive you are booting to when you restart or remove the other USB drive (assuming you are installing the OS from one USB stick onto the other)?


I tried both the "Install" option and the Live session option, both just went to black screen immediately afterward.

When I saw your install instructions and you had us using USB sticks rather than optical discs I just assumed we were actually installing Linux TO the USB rather than just using it load the image for install to an internal drive.

I'd rather not set-up a dual boot situation due to bad past experiences, but I have an extra SSD I could swap in for the Linux install. Never tried it, but I'm assuming that unplugging the Windows drive and replacing it with a drive on which I would then install Linux would allow me to just switch which SATA cable I have plugged in to switch OS (and files, etc.).

nanoprobe
Send message
Joined: 26 Feb 12
Posts: 184
Credit: 222,376,233
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwat
Message 44103 - Posted: 8 Aug 2016 | 21:23:13 UTC

I've been slowly moving my DC machines to Ubuntu 16.04 LTS. There's been some trials along the way but I was able to search for answers. The only problem I have not been able to solve is getting the Boinc manager to connect to the boinc client again. It was working fine until I did a sudo service boinc-client restart command in a terminal. After that the manager is shown as disconnected and it will not reconnect even after a restart or a complete power down and restart. I think it may have something to do with the client_state.xml file. The is no ip addy listed but when I edit the file and add one it doesn't stick. When I restart boinc it's empty again. When I check that file on other Linux hosts it shows the ip 127.0.0.1. On my lone windows host it shows the assigned ip addy I gave it (192.168.1.xx). Have no idea what else to try and asking for help @ askubuntu.com has resulted in no responses.

Matt
Avatar
Send message
Joined: 11 Jan 13
Posts: 216
Credit: 846,538,252
RAC: 0
Level
Glu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44104 - Posted: 8 Aug 2016 | 21:28:44 UTC - in response to Message 44103.

The only problem I have not been able to solve is getting the Boinc manager to connect to the boinc client again.


This is the issue I've always had with BOINC in Ubuntu. I've never been able to get it working because the Manager and the Client don't seem to want to talk to each other.

captainjack
Send message
Joined: 9 May 13
Posts: 171
Credit: 2,321,329,288
RAC: 2,347,578
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44105 - Posted: 9 Aug 2016 | 0:12:13 UTC

nanoprobe,

Never saw the issue that you have, but here are some ideas that might help.

Start up System Monitor and see if the boinc client is running. The Process Name will be "boinc" and the User will be "boinc".

Try restarting the boinc client with this command: "sudo /etc/init.d/boinc-client restart". Then restart the boinc manager and see if it connects.

If it is still not connected, from the boinc manager, from the menu bar at the top of the window, click on "File", then "Select computer...". A small window will open up asking for host name and password. In the host name box, try "localhost" (without the quotes) no password and see if that connects. If that doesn't work, try host name "127.0.0.1:31416" (without the quotes) and no password.

If none of that gets you going, how did you install BOINC?



Jim1348
Send message
Joined: 28 Jul 12
Posts: 819
Credit: 1,591,285,971
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44107 - Posted: 9 Aug 2016 | 13:24:53 UTC - in response to Message 44104.

The only problem I have not been able to solve is getting the Boinc manager to connect to the boinc client again.


This is the issue I've always had with BOINC in Ubuntu. I've never been able to get it working because the Manager and the Client don't seem to want to talk to each other.

I have had this problem on every BOINC installation I have done in Ubuntu 16.04 on four machines.

The solution is:

Copy "gui_rpc_auth.cfg" from the /etc/boinc-client folder to the home directory and reboot to allow BM to connect.

Since you are only reading from the /etc/boinc-client it should work, but if you have a permission problem let me know. Why they did not get this right to begin with is beyond me, but I am new to Ubuntu myself and don't know what worked before.

nanoprobe
Send message
Joined: 26 Feb 12
Posts: 184
Credit: 222,376,233
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwat
Message 44112 - Posted: 9 Aug 2016 | 19:49:24 UTC - in response to Message 44105.
Last modified: 9 Aug 2016 | 20:33:56 UTC

nanoprobe,

Never saw the issue that you have, but here are some ideas that might help.

Start up System Monitor and see if the boinc client is running. The Process Name will be "boinc" and the User will be "boinc".

Try restarting the boinc client with this command: "sudo /etc/init.d/boinc-client restart". Then restart the boinc manager and see if it connects.

If it is still not connected, from the boinc manager, from the menu bar at the top of the window, click on "File", then "Select computer...". A small window will open up asking for host name and password. In the host name box, try "localhost" (without the quotes) no password and see if that connects. If that doesn't work, try host name "127.0.0.1:31416" (without the quotes) and no password.

If none of that gets you going, how did you install BOINC?




Thanks but none of your suggestions worked. The boinc client and manager were both installed from the repositories.

The system monitor did not show boinc, only boinc manager but I know the client is running because I can control it with BoincTasks and the sensors show temp and 100% core load. Maybe things are different in this version of ubuntu.

nanoprobe
Send message
Joined: 26 Feb 12
Posts: 184
Credit: 222,376,233
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwat
Message 44113 - Posted: 9 Aug 2016 | 19:54:00 UTC - in response to Message 44107.

The only problem I have not been able to solve is getting the Boinc manager to connect to the boinc client again.


This is the issue I've always had with BOINC in Ubuntu. I've never been able to get it working because the Manager and the Client don't seem to want to talk to each other.

I have had this problem on every BOINC installation I have done in Ubuntu 16.04 on four machines.

The solution is:

Copy "gui_rpc_auth.cfg" from the /etc/boinc-client folder to the home directory and reboot to allow BM to connect.

Since you are only reading from the /etc/boinc-client it should work, but if you have a permission problem let me know. Why they did not get this right to begin with is beyond me, but I am new to Ubuntu myself and don't know what worked before.


I copied the gui_rpc_auth.cfg from the /etc/boinc-client folder but I was not able to paste it into the home directory. The paste option was greyed out which means I have a permissions issue. Did a lot of searching and tried some terminal commands but have not found an answer yet.

Jim1348
Send message
Joined: 28 Jul 12
Posts: 819
Credit: 1,591,285,971
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44114 - Posted: 9 Aug 2016 | 20:50:26 UTC - in response to Message 44113.
Last modified: 9 Aug 2016 | 21:49:54 UTC

I copied the gui_rpc_auth.cfg from the /etc/boinc-client folder but I was not able to paste it into the home directory. The paste option was greyed out which means I have a permissions issue. Did a lot of searching and tried some terminal commands but have not found an answer yet.


Here is a more complete list of what I do to allow access to /etc/boinc-client.

Note that I copy a modified "gui_rpc_auth.cfg" into that folder, but you may not need to modify it. Also, it allows for creating (or changing) a "cc_config.xml" file, if you need to do that. I also use a "remote_hosts.cfg" file to allow access over the LAN, but you may not need it.

I have also included the commands to allow access to /var/lib/boinc-client, where the project folders are located. You will need that to copy an "app_config.xml" file into the GPUGrid project folder, for example.

-> Just replace "user_name" with you actual user name.

Join the root group: sudo adduser user_name root
Join the BOINC group: sudo adduser user_name boinc

Allow group to read, write and execute in /etc/boinc-client folder: sudo chmod -R g+rwx /etc/boinc-client
Allow group to read, write and execute in /var/lib/boinc-client: sudo chmod -R g+rwx /var/lib/boinc-client
Reboot

Copy “gui_rpc_auth.cfg” to /etc/boinc-client folder
Copy “remote_hosts.cfg” to /etc/boinc-client folder
Copy “cc_config.xml” to /etc/boinc-client folder
Copy "gui_rpc_auth.cfg" to the home directory
Reboot

Copy the "app_config.xml" files to the project folders in /var/lib/boinc/

nanoprobe
Send message
Joined: 26 Feb 12
Posts: 184
Credit: 222,376,233
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwat
Message 44119 - Posted: 10 Aug 2016 | 0:51:09 UTC - in response to Message 44114.

I copied the gui_rpc_auth.cfg from the /etc/boinc-client folder but I was not able to paste it into the home directory. The paste option was greyed out which means I have a permissions issue. Did a lot of searching and tried some terminal commands but have not found an answer yet.


Here is a more complete list of what I do to allow access to /etc/boinc-client.

Note that I copy a modified "gui_rpc_auth.cfg" into that folder, but you may not need to modify it. Also, it allows for creating (or changing) a "cc_config.xml" file, if you need to do that. I also use a "remote_hosts.cfg" file to allow access over the LAN, but you may not need it.

I have also included the commands to allow access to /var/lib/boinc-client, where the project folders are located. You will need that to copy an "app_config.xml" file into the GPUGrid project folder, for example.

-> Just replace "user_name" with you actual user name.

Join the root group: sudo adduser user_name root
Join the BOINC group: sudo adduser user_name boinc

Allow group to read, write and execute in /etc/boinc-client folder: sudo chmod -R g+rwx /etc/boinc-client
Allow group to read, write and execute in /var/lib/boinc-client: sudo chmod -R g+rwx /var/lib/boinc-client
Reboot

Copy “gui_rpc_auth.cfg” to /etc/boinc-client folder
Copy “remote_hosts.cfg” to /etc/boinc-client folder
Copy “cc_config.xml” to /etc/boinc-client folder
Copy "gui_rpc_auth.cfg" to the home directory
Reboot

Copy the "app_config.xml" files to the project folders in /var/lib/boinc/

That fixed it. Can't thank you enough. For someone who claims to be new at Linux you fixed something I've searching for weeks. Where did you come up with the info?
Thanks again.
David

Jim1348
Send message
Joined: 28 Jul 12
Posts: 819
Credit: 1,591,285,971
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44120 - Posted: 10 Aug 2016 | 2:07:56 UTC - in response to Message 44119.

For someone who claims to be new at Linux you fixed something I've searching for weeks. Where did you come up with the info?
Thanks again.
David

You are welcome. It is Googling and educated guessing. There is unfortunately no single source of information for BOINC on Linux. That seems to be true of other aspects of Linux too. You have to learn it yourself for each distro. But for my simple crunching needs, it is now actually faster to set up a machine on Ubuntu than for Windows. I did my fourth one today in under two hours from the point of wiping the SSD with a secure erase to crunching on BOINC, and shortly thereafter Folding too.

It is unfortunate that the learning curve is steep though. Linux is really designed to be a secure server operating system, not a desktop OS. The permissions and folder structure are quite different than Windows, as you know.

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 44123 - Posted: 10 Aug 2016 | 10:33:08 UTC - in response to Message 44120.

Each distribution/version of Linux has different features, terminal commands and add ons which makes it almost as difficult moving from one distribution to another as it is to move from Windows to Linux. However, Ubuntu 6.04-x64 is a Long Term Support distro, so it's something you can stick with for longer. While the command lines might have changed somewhat, again, the GUI updates (via the Ubuntu Sortware icon and Sortware & Updates in System Settings) are encouraging.
At present 6.04-x64 LTS works for GPUGrid crunching due to the driver support for GTX900 series GPU's. The question at the back of my mind is will it support NV's GTX1000 series GPU's?

http://www.omgubuntu.co.uk/2016/06/top-5-terminal-commands

http://manpages.ubuntu.com/manpages/xenial/man8/
____________
FAQ's

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

nanoprobe
Send message
Joined: 26 Feb 12
Posts: 184
Credit: 222,376,233
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwat
Message 44124 - Posted: 10 Aug 2016 | 12:29:10 UTC - in response to Message 44123.

Each distribution/version of Linux has different features, terminal commands and add ons which makes it almost as difficult moving from one distribution to another as it is to move from Windows to Linux. However, Ubuntu 6.04-x64 is a Long Term Support distro, so it's something you can stick with for longer. While the command lines might have changed somewhat, again, the GUI updates (via the Ubuntu Sortware icon and Sortware & Updates in System Settings) are encouraging.
At present 6.04-x64 LTS works for GPUGrid crunching due to the driver support for GTX900 series GPU's. The question at the back of my mind is will it support NV's GTX1000 series GPU's?

http://www.omgubuntu.co.uk/2016/06/top-5-terminal-commands

http://manpages.ubuntu.com/manpages/xenial/man8/


16.04 does support the 10xx series GPUs. It was a minor PITA(probably more so for me being the Linux noob) to get it working. I can post what I did later if that will help. Just waiting for GPUGrid to support them.
Thanks for the links.

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 44137 - Posted: 11 Aug 2016 | 20:31:40 UTC - in response to Message 44124.

Please do, it's likely to be useful to many.
Thanks,
____________
FAQ's

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

nanoprobe
Send message
Joined: 26 Feb 12
Posts: 184
Credit: 222,376,233
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwat
Message 44142 - Posted: 12 Aug 2016 | 1:26:18 UTC
Last modified: 12 Aug 2016 | 1:28:13 UTC

When I first installed 16.04 I was using a 750Ti and I installed the 361 driver from the repository. When I tried to install my 1060 I couldn't get it to boot. I put the 750 back in and started searching. The 367.27 driver is needed for the 10xx cards. I also found it helped me to install synaptic package manager first.
I used the following commands to get started with the video drivers.

sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt-get update
sudo apt-get install nvidia-367 nvidia-settings


I then reinstalled the 1060 and it booted but there were still some issues. Installing the 367 driver did not install the needed OpenCl or CUDA libraries. I used synaptic package manager to install them. Now we're almost done.

Add the unprivileged boinc account to the video group

sudo usermod -a -G video boinc

Move BOINC start-up to the end of the start-up processing.

sudo update-rc.d -f boinc-client remove
sudo update-rc.d boinc-client defaults 99


Without nvidia-modprobe installed boinc will not find the GPU

sudo apt-get install nvidia-modprobe <-- when I did this I found that it was already installed YMMV

sudo modprobe nvidia_uvm

sudo service boinc-client restart

That's it but in some cases a reboot is needed. Hope this will help someone. My 1060 now happily runs POEM and Einstein. Patiently waiting for support here
You can find more driver info here.

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 45348 - Posted: 21 Nov 2016 | 12:12:20 UTC

Tried to use the 375.20 repo drivers on Ubuntu 16.04_x64 LTS but could not get Boinc to recognise the GPU. Tried lots of tricks (repeatedly) but none worked. Reverted to 370.28 and things are working again. X server said the 375.20 driver was there but maybe it didn't fully install.
For here there is no known reason to upgrade - I was just looking to see if there was anything in it (library updates that might make things faster), on the off-chance. There usually isn't BTW. If anyone tries to use 375.20 and can't get it to work, don't spend too much time on it.
____________
FAQ's

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

Rion Family
Send message
Joined: 13 Jan 14
Posts: 21
Credit: 15,404,426,517
RAC: 0
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 45395 - Posted: 25 Nov 2016 | 16:53:42 UTC

I have a setup running


Boinc 7.6.31
GenuineIntel Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz [Family 6 Model 60 Stepping 3]
(8 processors) [2] NVIDIA GeForce GTX 780 (3052MB) driver: 375.20 Linux
4.4.0-21-generic 25 Nov 2016 | 16:44:32 UTC


I was doing an upgrade from Linux Mint 17.3 (ubuntu 14.04) to 18 (ubuntu 16.04) ,
I did test the 375.20 drivers with both versions. however to get the latest boinc client I had to get to mint 18.

Basic approach - your milage may vary :)

sudo apt-get purge nvidia*

Note - Once you reboot here you have to do the next steps from the Console as it will not boot into X
sudo reboot

sudo add-apt-repository ppa:graphics-drivers/ppa
sudo
apt-get update

sudo apt-get install nvidia-375 nvidia-settings libcuda1-375

sudo apt-get install boinc 7.6


I found these sites usefull -
http://askubuntu.com/questions/757080/why-is-it-so-hard-to-install-nvidia-drivers-what-is-the-most-fool-proof-method
https://sites.google.com/site/easylinuxtipsproject/12

Here is what appears to be the relevent packages - there may be more - these are the ones I touched during the upgrades

paz@stealth-mint ~ $ sudo dpkg -l | grep -i nvidia
[sudo] password for paz:
rc libcuda1-331 331.113-0ubuntu0.0.4 amd64 NVIDIA CUDA runtime library
rc libcuda1-346 346.22-0ubuntu1~xedgers14.04.1 amd64 NVIDIA CUDA runtime library
rc libcuda1-349 349.16-0ubuntu0~xedgers14.04.1 amd64 NVIDIA CUDA runtime library
ii libcuda1-375 375.20-0ubuntu0~gpu16.04.1 amd64 NVIDIA CUDA runtime library
ii nvidia-375 375.20-0ubuntu0~gpu16.04.1 amd64 NVIDIA binary driver - version 375.20
ii nvidia-settings 375.20-0ubuntu0~gpu16.04.1 amd64 Tool for configuring the NVIDIA graphics driver

paz@stealth-mint ~ $ uname -a
Linux stealth-mint 4.4.0-21-generic #37-Ubuntu SMP Mon Apr 18 18:33:37 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux

paz@stealth-mint ~ $ sudo dpkg -l | grep -i boinc
ii boinc 7.6.31+dfsg-6ubuntu1 all metapackage for the BOINC client and the manager
ii boinc-client 7.6.31+dfsg-6ubuntu1 amd64 core client for the BOINC distributed computing infrastructure
ii boinc-manager 7.6.31+dfsg-6ubuntu1 amd64 GUI to control and monitor the BOINC core client
ii libboinc7:amd64 7.6.31+dfsg-6ubuntu1 amd64 libraries of BOINC the client depends on

paz@stealth-mint ~ $

MarkJ
Volunteer moderator
Volunteer tester
Send message
Joined: 24 Dec 08
Posts: 738
Credit: 200,909,904
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 47071 - Posted: 23 Apr 2017 | 11:06:09 UTC
Last modified: 23 Apr 2017 | 11:07:00 UTC

I've been converting all my Windows machines over to Linux of late. I have been running Win7 but given I don't want Microsoft deciding how to manage my machines for me I went with Linux.

I've written a blog post with the instructions HERE describing how to get Debian going.

Why Debian? Its more up to date than Ubuntu - Ubuntu is based upon Debian. Also everything needed is in their standard repositories. No fiddling with driver versions and the like.

Regarding the problems getting the manager to connect with the core client I found that simply adding the password that's in gui_rpc_auth.cfg to the properties of the BOINC Manager icon was enough for it to work. No need to have two copies of gui_rpc_auth. I use a windows laptop that runs BOINC Tasks to control the crunchers but also have the manager on the desktop of each cruncher.
____________
BOINC blog

ALAIN_13013
Avatar
Send message
Joined: 11 Sep 08
Posts: 18
Credit: 1,547,879,462
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50743 - Posted: 27 Oct 2018 | 7:20:52 UTC

UBUNTU 18.04 LTS
BOINC 7.9.3
NVIDIA driver 390.77
GPU 980 TI

How configure driver and BOINC, to reach 100% GPU load ?

GPU load about 30% 60%
T° 79° speed fan 50%
____________

Erich56
Send message
Joined: 1 Jan 15
Posts: 1090
Credit: 6,603,906,926
RAC: 21,893,126
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwat
Message 50744 - Posted: 27 Oct 2018 | 7:40:28 UTC - in response to Message 50743.

I am not familiar with Linux, but to me it seems that there may be some GPU throtteling due to the high temperature (79° - which, running constantly, is definitely not good for the lifetime of the card).

Zalster
Avatar
Send message
Joined: 26 Feb 14
Posts: 211
Credit: 4,496,324,562
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 50745 - Posted: 27 Oct 2018 | 15:53:34 UTC - in response to Message 50743.
Last modified: 27 Oct 2018 | 16:10:35 UTC

UBUNTU 18.04 LTS
BOINC 7.9.3
NVIDIA driver 390.77
GPU 980 TI

How configure driver and BOINC, to reach 100% GPU load ?

GPU load about 30% 60%
T° 79° speed fan 50%



I agree with the previous statement. You need to look at increasing your fan speed as your card is getting hot. Do you have nvidia settings installed? Are you able to adjust the fan speed?

Edit.. OK, just saw your PM. I have responded with instructions for you

Z
____________

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50746 - Posted: 27 Oct 2018 | 18:32:52 UTC - in response to Message 50743.

UBUNTU 18.04 LTS
BOINC 7.9.3
NVIDIA driver 390.77
GPU 980 TI

How configure driver and BOINC, to reach 100% GPU load ?

You can't use the SWAN_SYNC environmental variable under Linux to dedicate a CPU thread for the GPUGrid app to increase GPU usage.
(It works only under Windows.)
BTW the GPUGrid app can't reach 100% GPU usage. The maximum is about 95% (under Windows XP with SWAN_SYNC on).

GPU load about 30% 60%

This is quite low.
There's one component you forgot to post:
Your host has an AMD Ryzen Threadripper 2990WX 32-Core Processor, which is hyper-threaded, so it can run 64 tasks at the same time. In my experience if BOINC is not limited to use only the 50% of a HT CPU, the tasks will hinder each other's performance (as these scientific applications usually do a lot of floating point operations, and there's only 1 such FP unit available per core on the CPU). Memory bandwidth could be a limiting factor too when running too many tasks simultaneously. Too many CPU tasks can lower GPU performance a lot (In my experience there can be only 1 CPU task running without much GPU performance loss).

T° 79° speed fan 50%

This is quite hot, perhaps your CPU & GPU heats each other too much. If the GPU cooler emits the heat inside the PC case, you'll need some extra fans to remove the excess heat.

Jim1348
Send message
Joined: 28 Jul 12
Posts: 819
Credit: 1,591,285,971
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50747 - Posted: 27 Oct 2018 | 20:12:06 UTC - in response to Message 50743.
Last modified: 27 Oct 2018 | 20:25:54 UTC

UBUNTU 18.04 LTS
BOINC 7.9.3
NVIDIA driver 390.77
GPU 980 TI

How configure driver and BOINC, to reach 100% GPU load ?

GPU load about 30% 60%
T° 79° speed fan 50%

I am not quite sure how you managed to get that, except insufficient CPU support as Zoltan suggested. But I have a GTX 980 running under Ubuntu 16.04 at the moment (PABLO_2IDP), supported by a single core of an i7-3770. (The other seven cores are running Universe.) And I have the 396.54 driver, though I doubt that matters. It runs at 91 to 95% utilization, averaging about 93%. So I think you need to do something with the CPU to support your card better.

As for the temps, they are high. Is your room warm? But I see your fan is running at only 50%, so maybe it is designed to run hot. Some cards are. (Mine runs cool, since it has three fans; currently 65C with 38% fan speed.) Good luck.

PS - I have a Ryzen 1700 machine, and it has no trouble supporting Nvidia cards. I think it does at least as well as my Haswell (i7-4770), so the Threadripper should do OK. But what other projects are you running? How many cores reserved?

PPS: I recall it being discussed a while ago that Nvidia cards down-clock above a certain temperature, but I don't know what it is. You might want to check your GPU clock speed.

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50748 - Posted: 28 Oct 2018 | 1:52:57 UTC - in response to Message 47071.

I use a windows laptop that runs BOINC Tasks to control the crunchers but also have the manager on the desktop of each cruncher.


I just run BoincTasks under Wine. I have to have BT to monitor and control my hosts but don't want to ever touch Windows again. BT runs fine under Wine.

tullio
Send message
Joined: 8 May 18
Posts: 190
Credit: 104,426,808
RAC: 0
Level
Cys
Scientific publications
wat
Message 50749 - Posted: 28 Oct 2018 | 1:59:51 UTC

My GTX 750 Ti on a SuSE Leap 15.0 Linux and an Opteron 1210 CPU runs at 57 C, GPU clock 1137 MHz, memory clock 2700 MHz. Driver is 390.87.
Tullio

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50750 - Posted: 28 Oct 2018 | 2:19:36 UTC - in response to Message 50748.

Yes, all the Nvidia cards down/up clock in 12 Mhz bins for every 5 degrees decrease/increase in temps. I think the temp where the card doesn't hit the thermal binning point is somewhere around 54°C. At that point it will run at max clocks. All my EVGA Hybrid cards always run at max clocks since they are always cooler than 54°C. The only card that even gets close or rarely goes past that point is the GTX 1080Ti in the summer. So the cooler you can run the card, the higher the clocks it will maintain.

I run multiple projects on the cards but only 1 gpu task at a time. So any card at any time might have a Seti, MilkyWay, Einstein or GPUGrid task on it. Yet any time I look at nvidia-smi I always see all cards always at 95 - 100% utilization.

With Ryzen or TR, try and keep cpu utilization below 75% to allow enough cpu support for the gpu tasks.

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50751 - Posted: 28 Oct 2018 | 2:29:56 UTC - in response to Message 50746.

You can't use the SWAN_SYNC environmental variable under Linux to dedicate a CPU thread for the GPUGrid app to increase GPU usage.
(It works only under Windows.)


Maybe you can explain why I have one host that is able to run all GPUGrid tasks with SPIN synchronization. All my other hosts run with BLOCKING.

All the hosts have identical hardware for mobo, cpu and memory. The only differences are in the mix of Nvidia gpus each host has.

All the hosts run Ubuntu 18.04 with Nvidia 410.66 drivers.

So how is this host any different than the others? https://www.gpugrid.net/show_host_detail.php?hostid=456812

The only obvious difference is this host has three identical 1070Ti's and the other hosts have a mix of 1070/1070Ti/1080/1080TI cards in each host.

I've been asked a couple times now how I managed this feat. I don't have a clue.

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50753 - Posted: 28 Oct 2018 | 23:09:11 UTC - in response to Message 50751.
Last modified: 28 Oct 2018 | 23:10:51 UTC

You can't use the SWAN_SYNC environmental variable under Linux to dedicate a CPU thread for the GPUGrid app to increase GPU usage.
(It works only under Windows.)

Maybe you can explain why I have one host that is able to run all GPUGrid tasks with SPIN synchronization. All my other hosts run with BLOCKING.

I wish I could explain that, but I'm glad your host proves that I am wrong.

All the hosts have identical hardware for mobo, cpu and memory. The only differences are in the mix of Nvidia gpus each host has.

All the hosts run Ubuntu 18.04 with Nvidia 410.66 drivers.

So how is this host any different than the others? https://www.gpugrid.net/show_host_detail.php?hostid=456812

The only obvious difference is this host has three identical 1070Ti's and the other hosts have a mix of 1070/1070Ti/1080/1080TI cards in each host.

I've been asked a couple times now how I managed this feat. I don't have a clue.

There's total confusion about it on the forum.
I've tried every setting I can think of.
I've restarted my host after each configuration change.
I've tried to change the user account under which boinc runs, by changing the BOINC_USER setting in /etc/default/boinc-client and in /etc/init.d/boinc-client to my user and to root, none of them took effect (boinc still runs under user "boinc").
I've changed the access permissions of /var/lib/boinc-client to full access.
I've put "export swan_sync=0", "export swan_sync=1" to ~/.bashrc, /etc/init.d/boinc-client, /etc/bash.bashrc, /etc/profile
I've put "swan_sync=0", "swan_sync=1" to /etc/default/boinc-client, ~/.profile, /etc/profile
None of the above worked. (I don't know how Linux works, so it's no wonder.)
BTW I'm using Ubuntu 18.04.1 LTS.

References:
ACEMD2 6.12 cuda and 6.13 cuda31 for windows and linux
GTX 460
New applications ACEMD2 6.07/6.08 for Win and Lin
Load balancing GPUGRID and other boinc projects on Linux.
Setting up BOINC/GPUGRID on Fedora
GT240 and Linux: niceness and overclocking
Just a few Linux/GPU questions
The punchline: (message 38355)
CUDA 6.5 app for Linux now available on acemdbeta and acemdshort

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50759 - Posted: 29 Oct 2018 | 22:32:24 UTC

Well mystery solved after reading your history of threads regarding swan_sync. I believe that host 456812 was the first host I brought up on Linux and the project.

Seems I somewhere in the past put the SWAN_SYNC=1 entry into the /etc/environment file.

I assume that is the reason that host has been running tasks with the SPIN synchronization parameter.

Even though everyone said their attempt with the same entry never worked, I wonder if the reason it works with my hosts is that I don't use the repository versions of BOINC.

I use a specially Seti@home user compiled version 7.4.44 with settings allowing 3000 tasks on board and also the developer compiles it on much older distributions for compatibility. I believe he makes it compatible with distribution as old as 12.04.

Anyway, I am now modifying the environment file for the other hosts and see whether they too start running in SPIN synch.

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50761 - Posted: 30 Oct 2018 | 15:41:25 UTC - in response to Message 50759.

So I experimented and now have changed to SWAN_SYNC=1 on all my hosts. I ran overnight on the odd BOINC 7.8.3 host to make sure that it wasn't just the BOINC 7.4.44 that was able to use the SWAN_SYNC environment variable. I have the tasks now marked with SPIN synchronization too and the cpu_times equal the run_times so I used 100% of a cpu core to support the gpu task. Looks like I shaved 30-60 minutes of the typical PABLO task.

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 511
Credit: 4,617,042,755
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 50762 - Posted: 30 Oct 2018 | 16:33:23 UTC - in response to Message 50761.

So I experimented and now have changed to SWAN_SYNC=1 on all my hosts. I ran overnight on the odd BOINC 7.8.3 host to make sure that it wasn't just the BOINC 7.4.44 that was able to use the SWAN_SYNC environment variable. I have the tasks now marked with SPIN synchronization too and the cpu_times equal the run_times so I used 100% of a cpu core to support the gpu task. Looks like I shaved 30-60 minutes of the typical PABLO task.

Do you think you could give a complete guide of how to accomplish this for the linux novice? I think everyone would benefit from this information.

Zalster
Avatar
Send message
Joined: 26 Feb 14
Posts: 211
Credit: 4,496,324,562
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 50763 - Posted: 30 Oct 2018 | 17:12:37 UTC - in response to Message 50762.
Last modified: 30 Oct 2018 | 17:32:12 UTC

It's fairly easy. Question is whether the version of BOINC manager you are using will allow you to use it.

Keith and I both run a BM that was compiled by one of the Seti Gurus.

First thing is to unhide your files. In the right upper corner you should have a button to see files, you want to click show hidden files.

Once you do that, open files. Find your computer, click on it. Find the etc folder, click on that and then locate the environment file. Check it with texteditor. Might say something with "games" in the name. All you are doing is looking at it. Then close the file

move mouse over the open etc folder and right click and open the terminal.

Next get to root, so

"sudo su"

hit return then it will ask you to enter your password

you should be at root

"sudo nano environment"

Now you can edit the environment file

Add the

"SWAN_SYNC=1"

then control o to write, hit enter. You have now overwritten that file

control x to exit the edit

close terminal.

now using textedit, recheck to see if the edit took. If it did then you should be good.

I don't think you need to reboot. That should be it. I've tested it on one of my system and it works.

Edit..

Ok, reboot then. I always reboot anytime I do anything to a computer. Thanks Keith.
____________

mmonnin
Send message
Joined: 2 Jul 16
Posts: 332
Credit: 3,772,896,065
RAC: 4,765,302
Level
Arg
Scientific publications
watwatwatwatwat
Message 50764 - Posted: 30 Oct 2018 | 17:24:19 UTC - in response to Message 50761.

So I experimented and now have changed to SWAN_SYNC=1 on all my hosts. I ran overnight on the odd BOINC 7.8.3 host to make sure that it wasn't just the BOINC 7.4.44 that was able to use the SWAN_SYNC environment variable. I have the tasks now marked with SPIN synchronization too and the cpu_times equal the run_times so I used 100% of a cpu core to support the gpu task. Looks like I shaved 30-60 minutes of the typical PABLO task.


Did the tasks change w/in the past day or have you not been getting the 24hr bonus? All your tasks are worth 73k, 92k or 120k. I had mainly been getting 110k and 180k credit tasks. I was hoping to compare vs my 1070 running 2x tasks at once but want to make sure I compare proper task length.

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50765 - Posted: 30 Oct 2018 | 17:28:59 UTC - in response to Message 50763.



I don't think you need to reboot. That should be it. I've tested it on one of my system and it works.


You do have to either reboot or logoff the system. /etc/environment is a system wide variable and is loaded for all users local or remote. So a logoff and login is mandatory at the least. Reboot to be sure.

I will be curious to see whether this works on anyone running a repository version of BOINC.

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50766 - Posted: 30 Oct 2018 | 17:34:16 UTC - in response to Message 50764.

So I experimented and now have changed to SWAN_SYNC=1 on all my hosts. I ran overnight on the odd BOINC 7.8.3 host to make sure that it wasn't just the BOINC 7.4.44 that was able to use the SWAN_SYNC environment variable. I have the tasks now marked with SPIN synchronization too and the cpu_times equal the run_times so I used 100% of a cpu core to support the gpu task. Looks like I shaved 30-60 minutes of the typical PABLO task.


Did the tasks change w/in the past day or have you not been getting the 24hr bonus? All your tasks are worth 73k, 92k or 120k. I had mainly been getting 110k and 180k credit tasks. I was hoping to compare vs my 1070 running 2x tasks at once but want to make sure I compare proper task length.

None of the recently completed tasks with swan_sync were completed within 24 hours as GPUGrid is not my main project and only has a small resource assignment. Its too bad there is no work now to test the 24 hour bonus theory. I chewed up 3 of my normal 6 task cache overnight for the test but received no replacements. Today would have been the best day to test too since it is Outage Tuesday at Seti and I run out of Seti work in the first hour the project goes down which allows my backup projects like GPUGrid to make hay for the day.

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50767 - Posted: 30 Oct 2018 | 18:27:15 UTC

I would appreciate those attempting the SWAN_SYNC=1 parameter in their environment file on repository versions of BOINC to report success or not. You can check if the parameter took or not with a

printenv SWAN_SYNC
in Terminal which should report a "1" if successful.

I also would like to hear which version people are using and reporting success or failure with.

So far all the positive reports are from two users using non-repository versions of BOINC.

ALAIN_13013
Avatar
Send message
Joined: 11 Sep 08
Posts: 18
Credit: 1,547,879,462
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50768 - Posted: 30 Oct 2018 | 18:29:40 UTC

OK merci beaucoup pour vos commentaires.
Je test pour vous dire si c'est bon.

OK thank you very much for your comments.
I test to tell you if it is good.

SWAN_SYNC=1
99% CORE CPU
GPU FAN Setting 60%
GPU load 60 to 90%
T° 77° to 82%

____________

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50770 - Posted: 30 Oct 2018 | 18:49:57 UTC - in response to Message 50768.

Alain, that was with repository BOINC version 7.9.3. Correct?

captainjack
Send message
Joined: 9 May 13
Posts: 171
Credit: 2,321,329,288
RAC: 2,347,578
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50772 - Posted: 30 Oct 2018 | 21:17:18 UTC

Keith Myers asked:

I would appreciate those attempting the SWAN_SYNC=1 parameter in their environment file on repository versions of BOINC to report success or not.

I am running Ubuntu 18.10 and BOINC 7.12.0 from the Ubuntu software installer.

I was able to get the SWAN_SYNC=1 set up in the environmental variables. When I request to printenv SWAN_SYNC, it answers back with a 1.

However, my GPU tasks are still running with Synchronization mode BLOCKING.

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50773 - Posted: 30 Oct 2018 | 23:18:27 UTC - in response to Message 50772.
Last modified: 30 Oct 2018 | 23:26:40 UTC

Did you logoff or reboot to enable the parameter to be read?
[Edit] Also any task started before the parameter was set already has the old BLOCKING parameter written into the header in the slot. So only tasks started after the parameter was set will have the new SPIN synchronization set in the stderr.txt output.

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 511
Credit: 4,617,042,755
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 50774 - Posted: 31 Oct 2018 | 0:50:21 UTC - in response to Message 50763.

Now you can edit the environment file

Add the

"SWAN_SYNC=1"

Do I add this after PATH=bin:? And do quotations matter?

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50775 - Posted: 31 Oct 2018 | 1:32:38 UTC - in response to Message 50774.

Just drop down to a new line below the PATH statement. And just put in SWAN_SYNC=1. No quotes. But is does have to be capitalized since Linux cares about upper/lower case in names.

captainjack
Send message
Joined: 9 May 13
Posts: 171
Credit: 2,321,329,288
RAC: 2,347,578
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50776 - Posted: 31 Oct 2018 | 1:40:52 UTC

Keith asked:

Did you logoff or reboot to enable the parameter to be read?
[Edit] Also any task started before the parameter was set already has the old BLOCKING parameter written into the header in the slot. So only tasks started after the parameter was set will have the new SPIN synchronization set in the stderr.txt output.

The PC was rebooted and the GPU tasks started after the reboot. You should be able to see my computers, the computer that I am testing is named Sledgehammer if you want to look.

Question for you, does the version of BOINC that you are using run under your id or the id of "boinc" or other? You probably know that the repository version of BOINC runs under the id of "boinc".

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50777 - Posted: 31 Oct 2018 | 2:43:33 UTC - in response to Message 50776.

Keith asked:
Did you logoff or reboot to enable the parameter to be read?
[Edit] Also any task started before the parameter was set already has the old BLOCKING parameter written into the header in the slot. So only tasks started after the parameter was set will have the new SPIN synchronization set in the stderr.txt output.

The PC was rebooted and the GPU tasks started after the reboot. You should be able to see my computers, the computer that I am testing is named Sledgehammer if you want to look.

Question for you, does the version of BOINC that you are using run under your id or the id of "boinc" or other? You probably know that the repository version of BOINC runs under the id of "boinc".

I have no way of knowing what your internal name for your computers are. That is not exposed to anyone but you. You have to identify computers by their host ID for anyone other than you to figure out which is which in your hosts list.

My and Zalster's BOINC accounts are "owned" by us since they reside in the /home directory.

Alain however is using the standard repository version 7.9.3 and the owner would be the standard "boinc" user. He has reported success in using the swan_sync variable.

There might a version level that the parameter no longer works. So far, the highest level BOINC that is successful is 7.9.3. I hope others report their results of either failure or success with the version level.

I was afraid that our lower level BOINC versions were the reason for success and the fact they are owned by the user and not boinc. His result proves that is not the case.

So either your higher level of BOINC is the issue or their is something different in your system compared to the 3 successful cases so far.

Zalster
Avatar
Send message
Joined: 26 Feb 14
Posts: 211
Credit: 4,496,324,562
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 50778 - Posted: 31 Oct 2018 | 3:20:39 UTC - in response to Message 50777.

.

So either your higher level of BOINC is the issue or their is something different in your system compared to the 3 successful cases so far.


I keep looking for Alain's results to see if it's working. The last one reported was done before he made the change so we wait to see what it says when his current 2 get done. Till then, we only know your and mine are working. Hopefully by tomorrow more people will have tried it and we will get a better idea.

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50779 - Posted: 31 Oct 2018 | 3:57:34 UTC - in response to Message 50778.

Of course I have no proof, since PM's leave no record, but I'm positive he said his gpu is running at 100%, thus he has a full cpu feeding it.

rod4x4
Send message
Joined: 4 Aug 14
Posts: 266
Credit: 2,219,935,054
RAC: 0
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 50780 - Posted: 31 Oct 2018 | 5:08:15 UTC

I wont get a chance to try this solution for another 6 hours, but...

If you check host id: 153837 and 423329. (both belong to same user, [not me])

Both running Boinc client 7.8.3 and have been using SPIN since at least 21st October (possibly longer)

ALAIN_13013
Avatar
Send message
Joined: 11 Sep 08
Posts: 18
Credit: 1,547,879,462
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50781 - Posted: 31 Oct 2018 | 5:16:31 UTC - in response to Message 50770.

Alain, that was with repository BOINC version 7.9.3. Correct?

Yes 7.9.3 version.
____________

Zalster
Avatar
Send message
Joined: 26 Feb 14
Posts: 211
Credit: 4,496,324,562
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 50782 - Posted: 31 Oct 2018 | 7:55:51 UTC - in response to Message 50781.

Alain, that was with repository BOINC version 7.9.3. Correct?

Yes 7.9.3 version.


For Alain,

1 of 2 just returned. The time to complete is less than other work units but they aren't the same type so can't tell if there is really any improvement due to the addition of the SWAN_SYNC

Here's part of the stderr

# CUDA Synchronisation mode: BLOCKING
# SWAN Device 0 :
# Name : GeForce GTX 980 Ti
# ECC : Disabled
# Global mem : 6075MB
# Capability : 5.2
# PCI ID : 0000:42:00.0
# Device clock : 1076MHz
# Memory clock : 3505MHz
# Memory width : 384bit
# GPU [GeForce GTX 980 Ti] Platform [Linux] Rev [3212] VERSION [80]
# SWAN Device 0 :
# Name : GeForce GTX 980 Ti
# ECC : Disabled
# Global mem : 6075MB
# Capability : 5.2
# PCI ID : 0000:42:00.0
# Device clock : 1076MHz
# Memory clock : 3505MHz
# Memory width : 384bit
# Time per step (avg over 6560722 steps): 3.356 ms
# Approximate elapsed time for entire WU: 33555.639 s
# PERFORMANCE: 24566 Natoms 3.356 ns/day 0.000 ms/step 3.356 us/step/atom
06:41:38 (8543): called boinc_finish
# CUDA Synchronisation mode: BLOCKING
# SWAN Device 0 :


According to this, it wasn't picking up it up. Will have to wait for the other work unit to return to see if that one has it on there or not.
____________

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 511
Credit: 4,617,042,755
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 50783 - Posted: 31 Oct 2018 | 10:48:57 UTC
Last modified: 31 Oct 2018 | 10:49:17 UTC

Hello, I followed the instructions and typed "printenv SWAN_SYNC" and it returned with 1. Below are my tasks from last night but it appears not to have synced the CPU time with the overall time. This computer's BOINC version is 7.6.31.

http://www.gpugrid.net/results.php?hostid=490728

kksplace
Send message
Joined: 4 Mar 18
Posts: 53
Credit: 1,398,076,749
RAC: 3,583,485
Level
Met
Scientific publications
wat
Message 50784 - Posted: 31 Oct 2018 | 11:24:59 UTC

Same thing here. printenv SWAN_SYNC returns a 1 now, but still showing Blocking on the WUs from overnight.

http://www.gpugrid.net//results.php?hostid=48715
(WUs completed on 31 Oct all started after SWAN_SYNC=1)

Boinc 7.9.3 from Linux Mint Software Manager.

(Thank you to Keith and Zalster for trying to figure this out.)

Zalster
Avatar
Send message
Joined: 26 Feb 14
Posts: 211
Credit: 4,496,324,562
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 50785 - Posted: 31 Oct 2018 | 13:08:19 UTC

ALAIN's second work unit returned also with BLOCKING still in the stderr. So it is not picking up the SWAN_SYNC=1

Thanks to everyone who is reporting back with their results.

I think it maybe necessary to use the BOINC Manager version compiled by TBar at crunchers anonymous in order to get it to work properly.


http://www.arkayn.us/lunatics/BOINC-7.8.3.7z

The default of the BM is set to connect to Seti@home but you can always add GPUGrid as a project and then delete the Seti@home folder.
____________

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50787 - Posted: 31 Oct 2018 | 16:46:37 UTC

OK, yes I agree with Zalster's assessment now too. It looks like only the TBar BOINC versions is able to run with SWAN_SYNC=1 and have it applied to the tasks.

To add to just the link that Zalster posted, I thought the full post describing the steps necessary to run TBar's BOINC versions.


The Latest version of BOINC for Testing in Linux. You should check the Dependencies and Execute bits before running. This version should work in Ubuntu 12.04 and above by installing libwebkitgtk from the Package Manager, however, Mint users should probably use the repository version of BOINC .

This version has Fixed the BOINC Manager Tasks view problem that resulted in the view being focused on the bottom of the Tasks page, and Simple View works. If you have an existing BOINC folder in your Home folder expand the download to your Downloads folder, and Stop all boinc Apps. Open the download, Copy the files and Paste them into your existing BOINC folder in Home. Double click boincmgr to launch the Manager.

To expand the file in Linux you need to have 7zip installed, in the terminal;
To install 7zip run, sudo apt-get install p7zip
To install libwebkitgtk run, sudo apt-get install libwebkitgtk-1.0-0


I also have had to install libcurl3 for the boinc client along with the libwebkitgtk-1.0-0 for the Manager. I did discover that Ubuntu 18.10 has removed libcurl3 from the repository and the solution is to add the curl34 ppa repository and then install the libcurl4 library from that ppa which has both libcurl3 and libcurl4 libraries in the libcurl4 package.

A major benefit of these BOINC versions are that you simply unpack them to location of your choice like /home. The owner is then just you with full permissions to do whatever you want with any file.

Based on comments from others at SETI that have previously run the repository versions of BOINC, it is best to purge/autoremove the repository versions and reboot before installing the TBar versions to prevent old symbolic links from interfering.

Zalster
Avatar
Send message
Joined: 26 Feb 14
Posts: 211
Credit: 4,496,324,562
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 50788 - Posted: 31 Oct 2018 | 18:45:50 UTC - in response to Message 50787.


To install libwebkitgtk run, sudo apt-get install libwebkitgtk-1.0-0


Thanks Keith for all that.

Correct me if I'm wrong but isn't it?

sudo apt-get install libwebkitgtk-1.0



Z
____________

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50789 - Posted: 31 Oct 2018 | 19:16:17 UTC - in response to Message 50788.

Yes, you should just drop the trailing zero.

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50791 - Posted: 31 Oct 2018 | 20:44:50 UTC - in response to Message 50780.

I wont get a chance to try this solution for another 6 hours, but...

If you check host id: 153837 and 423329. (both belong to same user, [not me])

Both running Boinc client 7.8.3 and have been using SPIN since at least 21st October (possibly longer)

Can't tell just from the Host details which and where from the hosts obtained the BOINC 7.8.3. It very well could be the TBar version so explains why it works.

This user participates in lots of projects, Seti being one of them and could very well have installed the TBar BOINC 7.8.3 version.

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50792 - Posted: 31 Oct 2018 | 23:47:30 UTC
Last modified: 31 Oct 2018 | 23:53:25 UTC

I've managed the BOINC client (installed from the repo) to notice the SWAN_SYNC=1 parameter (in /etc/environment).
(BTW it has to be spelled with all caps to work.)
By default the BOINC client runs under the account called 'boinc', as a daemon (system service).
When it runs this way, it won't regard the environment settings in /etc/environment.
So I stopped the BOINC daemon by opening a terminal and elevating it with

sudo su
then
/etc/init.d/boinc-client stop
then I changed dir to the BOINC data dir with
cd /var/lib/boinc-client
then I've started BOINC with
/usr/bin/boinc
It will run as 'root' and it will have the proper environmental settings, but it can't be managed through the graphical interface.
The acemd will use a full thread, which is 25% CPU on my Core i3 (I've checked it)
The result is amazing: 6.872 sec for a PABLO_2IDP_P01106 task (110.400 credits). It took 8.800 seconds without the SWAN_SYNC.
My GTX 1080Ti got 28% faster by applying SWAN_SYNC=1.
You can check my hosts:Host 391255, Host 482653

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50793 - Posted: 1 Nov 2018 | 0:48:14 UTC

Thanks for the update. I had a hunch the issue with the repository version was because of the account "boinc" permissions preventing the system wide environment variable not being picked up.

The TBar BOINC versions avoid this issue since both the client and Manager are installed and run under the user account.

kksplace
Send message
Joined: 4 Mar 18
Posts: 53
Credit: 1,398,076,749
RAC: 3,583,485
Level
Met
Scientific publications
wat
Message 50794 - Posted: 1 Nov 2018 | 2:07:51 UTC - in response to Message 50792.

Thank you and trying now.

Is there a way to manage via another manager, like BAM, while in this state?

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50795 - Posted: 1 Nov 2018 | 2:39:49 UTC - in response to Message 50794.

You should still be able to control the client with the boinccmd tool.
Controlling_BOINC_remotely

As far as using another manager, you could install Wine and then install BoincTasks to manage the client.

I found BoincTasks indispensable when I was running Windows, so the first thing I did after moving to Linux was to investigate whether BoincTasks would run under Linux. It does, quite well in fact. So I have Wine installed on my daily driver that I sit in front of 8 hours of the day and I control and monitor my 4 other hosts from that chair without having to get out of it and sit in front of the other hosts to interact with them.

BoincTasks, TThrottle

tullio
Send message
Joined: 8 May 18
Posts: 190
Credit: 104,426,808
RAC: 0
Level
Cys
Scientific publications
wat
Message 50796 - Posted: 1 Nov 2018 | 4:23:15 UTC

I am running SuSE Leap 15.0 on two Linux hosts. BOINC 7.8.3 is given to me by SuSE. One is running QC only, no graphic board, the other both QC and acemd on its GTX 750 Ti, CUDA 80, BLOCKED parameter.
Tullio

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50797 - Posted: 1 Nov 2018 | 5:24:26 UTC - in response to Message 50796.

We think it doesn't matter what the BOINC version is . . . if it comes from a repository and is installed as a daemon with the user 'boinc', it is not going to allow the SWAN_SYNC=1 parameter to be enabled on gpu tasks.

Only the Seti user developed BOINC versions which are installed in /home with the user the owner use the parameter correctly.

Retvari Zoltan has figured out a method around the repository versions but without Manager control of the client.

ALAIN_13013
Avatar
Send message
Joined: 11 Sep 08
Posts: 18
Credit: 1,547,879,462
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50798 - Posted: 1 Nov 2018 | 6:59:17 UTC - in response to Message 50785.

ALAIN's second work unit returned also with BLOCKING still in the stderr. So it is not picking up the SWAN_SYNC=1

Thanks to everyone who is reporting back with their results.

I think it maybe necessary to use the BOINC Manager version compiled by TBar at crunchers anonymous in order to get it to work properly.


http://www.arkayn.us/lunatics/BOINC-7.8.3.7z

The default of the BM is set to connect to Seti@home but you can always add GPUGrid as a project and then delete the Seti@home folder.


SWAN_SYNC variable is setting
but BOINC don't read it
____________

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50800 - Posted: 1 Nov 2018 | 9:25:50 UTC - in response to Message 50792.

My workaround in my previous post is just a temporary solution (it's fine for me for the time being).
To completely solve this issue we should:
1. change the owner of /var/lib/boinc-client to the user (I know how to do that, but it's pointless if I can't do the next steps).
2. change the user account under which the boinc daemon starts. There should be a Linux configuration file containing this setting, which I haven't found yet (it's not enough to change the BOINC_USER in /etc/default/boinc-client and in /etc/init.d/boinc-client)
3. the user should have the right to start a daemon (at least in Windows it works that way).

rod4x4
Send message
Joined: 4 Aug 14
Posts: 266
Credit: 2,219,935,054
RAC: 0
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 50801 - Posted: 1 Nov 2018 | 9:59:39 UTC - in response to Message 50800.
Last modified: 1 Nov 2018 | 10:58:41 UTC

I have successfully automated the start of boinc client.

It does weaken the security of the installed boinc package so use at your own risk!!

in the service control file /lib/systemd/system/boinc-client.service:
Under [service] heading, make this the first line
Environment="SWAN_SYNC=1"

Under [service] heading, change User=boinc to:
User=root (this is the line that weakens security)
EDIT: as per testing by Retvari Zoltan, this line does not need to be changed. It can stay as User=boinc (and hence not compromise security). Thanks Retvari Zoltan

You do not need to add SWAN_SYNC=1 to the /etc/environment file as a service creates it's own environment(my understanding)

reboot. (easiest way to implement)

run top in a terminal (or your favourite GUI task manager) to check acemd process is using 100% of cpu

kksplace
Send message
Joined: 4 Mar 18
Posts: 53
Credit: 1,398,076,749
RAC: 3,583,485
Level
Met
Scientific publications
wat
Message 50802 - Posted: 1 Nov 2018 | 10:38:14 UTC - in response to Message 50795.
Last modified: 1 Nov 2018 | 10:44:37 UTC

Success! I did the steps recommended by Retvari Zoltan and then also installed Wine and BoincTasks (thank you Keith Myers) in order to have a manager (which installed very nicely and automatically 'synced up' with the running Boinc Client).

As implied by Zoltan's follow-up post, I did realize I had to leave the terminal open with the root-level Boinc Client running.

Overnight, my PABLO WUs processed about 8 percent faster.

http://www.gpugrid.net//results.php?hostid=487152

Looks like more ideas flowing here -- I will wait until this evening to set up my second host and see what else is discovered.

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50803 - Posted: 1 Nov 2018 | 10:43:33 UTC - in response to Message 50801.
Last modified: 1 Nov 2018 | 10:44:15 UTC

in the service control file /lib/systemd/system/boinc-client.service:
Under [service] heading, make this the first line
Environment="SWAN_SYNC=1"
This is the configuration file I was looking for. Thanks!

Under [service] heading, change User=boinc to:
User=root (this is the line that weakens the security)
I've tested it without changing the user, and it's working, so you don't have to compromise the security of the system.

rod4x4
Send message
Joined: 4 Aug 14
Posts: 266
Credit: 2,219,935,054
RAC: 0
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 50805 - Posted: 1 Nov 2018 | 10:53:09 UTC - in response to Message 50803.

Thanks for the feedback, will reset
User=boinc as you suggested!

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 511
Credit: 4,617,042,755
RAC: 0
Level
Arg
Scientific publications
watwatwatwatwatwatwatwat
Message 50806 - Posted: 1 Nov 2018 | 10:57:43 UTC

Zoltan, now that you've tried all of the potential solutions, what do you think is the easiest, safest, and most fool proof?

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50808 - Posted: 1 Nov 2018 | 11:10:44 UTC - in response to Message 50806.

Zoltan, now that you've tried all of the potential solutions, what do you think is the easiest, safest, and most fool proof?
Rod4x4's solution is the perfect one.
I've also made a post in the "SWAN_SYNC in Linux client" thread based on his solution.

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50809 - Posted: 1 Nov 2018 | 11:15:13 UTC - in response to Message 50805.

Thanks for the feedback, will reset
User=boinc as you suggested!

My tasks that are started as 'root' won't continue after I've started the BOINC manager as 'boinc', so don't be surprised if you have to manage this (aborting the task, or deleting the files from the slot as 'root').

tullio
Send message
Joined: 8 May 18
Posts: 190
Credit: 104,426,808
RAC: 0
Level
Cys
Scientific publications
wat
Message 50810 - Posted: 1 Nov 2018 | 11:20:47 UTC
Last modified: 1 Nov 2018 | 11:22:21 UTC

I am running both python and acemd as user tullio, so I have almost the whole 1 TB disk available to BOINC 7.8.3.
Tullio

mmonnin
Send message
Joined: 2 Jul 16
Posts: 332
Credit: 3,772,896,065
RAC: 4,765,302
Level
Arg
Scientific publications
watwatwatwatwat
Message 50811 - Posted: 1 Nov 2018 | 11:43:50 UTC - in response to Message 50802.

Success! I did the steps recommended by Retvari Zoltan and then also installed Wine and BoincTasks (thank you Keith Myers) in order to have a manager (which installed very nicely and automatically 'synced up' with the running Boinc Client).

As implied by Zoltan's follow-up post, I did realize I had to leave the terminal open with the root-level Boinc Client running.

Overnight, my PABLO WUs processed about 8 percent faster.

http://www.gpugrid.net//results.php?hostid=487152

Looks like more ideas flowing here -- I will wait until this evening to set up my second host and see what else is discovered.


Add --daemon to the boinc client startup command to allow the client to run w/o an active terminal. That's how I start multiple clients on a PC.

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50817 - Posted: 2 Nov 2018 | 9:05:01 UTC - in response to Message 50809.

kksplace wrote:
Retvari Zoltan wrote:
"My tasks that are started as 'root' won't continue after I've started the BOINC manager as 'boinc', so don't be surprised if you have to manage this (aborting the task, or deleting the files from the slot as 'root')."


First, thank you for pursuing the great Linux/Boinc SWAN_SYNC mystery to completion -- I really appreciate it.

Next, on one of my hosts I did the 'run-BOINC-as-root' thing before the final answer. Based on your post quoted above, I might have some difficulty switching back to 'normal' Boinc manager. Do you have a recommendation on how to do it with minimal pain? (I also installed BoincTasks, so I do have a manager running.)

Thank you for any help on this.

You can do it with minimal pain by waiting for the tasks run out on the GPUGrid servers (or you can set the BOINC manager not to request new tasks from the GPUGrid project) and when the GPUGrid tasks run out on your host you can switch back to run BOINC daemon as 'boinc'.
You're welcome!

kksplace
Send message
Joined: 4 Mar 18
Posts: 53
Credit: 1,398,076,749
RAC: 3,583,485
Level
Met
Scientific publications
wat
Message 50818 - Posted: 3 Nov 2018 | 16:47:34 UTC - in response to Message 50817.

You can do it with minimal pain by waiting for the tasks run out on the GPUGrid servers (or you can set the BOINC manager not to request new tasks from the GPUGrid project) and when the GPUGrid tasks run out on your host you can switch back to run BOINC daemon as 'boinc'.


Thank you for the help.

I also discovered that somehow the Boinc Manager and Client became disconnected during the process. I found the following worked to get everything connected again:

https://askubuntu.com/questions/606666/boinc-wont-connect-to-the-client

Now everything back to 'normal' with SWAN_SYNC happily SPINning away.

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50820 - Posted: 3 Nov 2018 | 21:25:56 UTC - in response to Message 50746.

UBUNTU 18.04 LTS
BOINC 7.9.3
NVIDIA driver 390.77
GPU 980 TI

How configure driver and BOINC, to reach 100% GPU load ?

You can't use the SWAN_SYNC environmental variable under Linux to dedicate a CPU thread for the GPUGrid app to increase GPU usage.
(It works only under Windows.)
BTW the GPUGrid app can't reach 100% GPU usage. The maximum is about 95% (under Windows XP with SWAN_SYNC on).

To my surprise I see 99-100% GPU usage under Linux since we've successfully applied SWAN_SYNC=1.
I will convert some of my GTX980Ti hosts to Linux to be able to compare the performance of Windows XP.
The Linux app using more recent CUDA (8.0), so perhaps it helps to increase its performance over the old Windows XP app (CUDA 6.5).

Keith Myers
Send message
Joined: 13 Dec 17
Posts: 1284
Credit: 4,919,281,959
RAC: 6,261,046
Level
Arg
Scientific publications
watwatwatwatwat
Message 50821 - Posted: 3 Nov 2018 | 23:40:56 UTC - in response to Message 50820.

I think it's relatively accepted that the Linux scheduler is smarter than the Windows one. Probably the difference.

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 7,520
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 50822 - Posted: 4 Nov 2018 | 11:27:04 UTC

Don't be surprised if the power dissipation (i.e. the temperature and the fan speed) of your GPU(s) increase (about as much as its speed increases) after you apply the SWAN_SYNC=1.

Post to thread

Message boards : Number crunching : Discussion of Ubuntu 16.04-x64 LTS installation and configuration

//