climateprediction.net home page
New work Discussion

New work Discussion

Message boards : Number crunching : New work Discussion
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 48 · 49 · 50 · 51 · 52 · 53 · 54 . . . 91 · Next

AuthorMessage
Jim1348

Send message
Joined: 15 Jan 06
Posts: 637
Credit: 26,751,529
RAC: 653
Message 62963 - Posted: 17 Nov 2020, 16:15:09 UTC - in response to Message 62962.  

Store at least 0.3 days of work.
Store up to an additional 1 day of work.

Is this ridiculous? Or is it reasonable?

On my Ryzen 3000 series and Ubuntu 18.04, I can usually get work at the default setting of 0.1 + 0.5 days.
Sometimes if I am running other long ones (the Rosetta 24 hour work units for example), I have to increase it to 0.25 + 0.5, or a little more.
It may take a little more on slower machines.

Note that I limit the N216 to running two at a time however, or I take a hit on run times.
ID: 62963 · Report as offensive
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4361
Credit: 16,614,059
RAC: 1,518
Message 62964 - Posted: 17 Nov 2020, 19:30:29 UTC

Store at least 0.3 days of work.
Store up to an additional 1 day of work.

Is this ridiculous? Or is it reasonable?


Seems reasonable to me. I am out of ideas now though.
ID: 62964 · Report as offensive
Profile geophi
Volunteer moderator

Send message
Joined: 7 Aug 04
Posts: 2171
Credit: 64,560,519
RAC: 352
Message 62965 - Posted: 17 Nov 2020, 19:42:55 UTC - in response to Message 62962.  

It may be worth trying increasing the number of days work requested above the default. I know Les didn't find anything in work fetch debug that was helpful.


Do you have a suggestion as to what this should be? I know with different hardware and different OS, the amount will be different, but right now mine is set to

Store at least 0.3 days of work.
Store up to an additional 1 day of work.

Is this ridiculous? Or is it reasonable?

Seems reasonable. Do you run any other projects? If so, try getting work from them. Just wondering if this is a cpdn thing only.

On my CentOS 8 VM, I installed and configured boinc mostly following a couple links

https://boinc.berkeley.edu/forum_thread.php?id=11854
https://boinc.berkeley.edu/wiki/Installing_BOINC_on_Fedora

I also disabled selinux, although I don't know if that really did anything. I am able to get work on World Community Grid. After that task runs, I'll try cpdn.
ID: 62965 · Report as offensive
Profile geophi
Volunteer moderator

Send message
Joined: 7 Aug 04
Posts: 2171
Credit: 64,560,519
RAC: 352
Message 62966 - Posted: 17 Nov 2020, 20:45:55 UTC - in response to Message 62962.  

Jean-David Beyer

The CentOS 8 installation did pick up a cpdn task, so not sure what the difference is between your setup and mine.
ID: 62966 · Report as offensive
Jean-David Beyer

Send message
Joined: 5 Aug 04
Posts: 1082
Credit: 16,655,848
RAC: 6,703
Message 62967 - Posted: 17 Nov 2020, 22:18:32 UTC - in response to Message 62965.  

Seems reasonable. Do you run any other projects? If so, try getting work from them. Just wondering if this is a cpdn thing only.

On my CentOS 8 VM, I installed and configured boinc mostly following a couple links

https://boinc.berkeley.edu/forum_thread.php?id=11854
https://boinc.berkeley.edu/wiki/Installing_BOINC_on_Fedora

I also disabled selinux, although I don't know if that really did anything. I am able to get work on World Community Grid. After that task runs, I'll try cpdn.


Yes, I am running World Community Grid and Rosetta@home and both are working just fine. So it seems to be a cpdn thing only. Am I just being too impatient?

Tue 17 Nov 2020 03:59:34 PM EST | climateprediction.net | Sending scheduler request: To fetch work.
Tue 17 Nov 2020 03:59:34 PM EST | climateprediction.net | Requesting new tasks for CPU
Tue 17 Nov 2020 03:59:36 PM EST | climateprediction.net | Scheduler request completed: got 0 new tasks
Tue 17 Nov 2020 03:59:36 PM EST | climateprediction.net | No tasks sent
Tue 17 Nov 2020 03:59:36 PM EST | climateprediction.net | Project requested delay of 3636 seconds


There seem to be work units available.
UK Met Office HadAM4 at N216 resolution 5297 5277 429.98 (177.44 - 923.93) 25

UK Met Office HadAM4 at N216 resolution
Platform Version Created Average computing
Linux/x86 8.52 7 Jun 2019, 11:07:36 UTC 2,567 GigaFLOPS

ID: 62967 · Report as offensive
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 62968 - Posted: 18 Nov 2020, 0:16:05 UTC

I've long suspected that it's a cpdn thing.
I'll talk to the project.
ID: 62968 · Report as offensive
Profile geophi
Volunteer moderator

Send message
Joined: 7 Aug 04
Posts: 2171
Credit: 64,560,519
RAC: 352
Message 62969 - Posted: 18 Nov 2020, 0:43:36 UTC - in response to Message 62967.  

Yes, I am running World Community Grid and Rosetta@home and both are working just fine. So it seems to be a cpdn thing only. Am I just being too impatient?

No, you're not being too impatient. It should work. I know you installed the recommended 32bit libraries in the instructions. But do this. Download

http://www.cpdn.org/cpdnboinc/applications//hadam4_8.52_i686-pc-linux-gnu
http://www.cpdn.org/cpdnboinc/applications//hadam4_se_8.52_i686-pc-linux-gnu.zip
http://www.cpdn.org/cpdnboinc/applications//hadam4_um_8.52_i686-pc-linux-gnu.zip

into a temp directory under your home directory. Unzip the zipped files then you will likely have to chmod them +x

Then do an

ldd *

within that directory. Just to make sure that there are no unmet dependencies.

If that doesn't show anything problematic, I suggest going through the two links in my last post and see if performing any steps in those linked posts might resolve the problem. I'm wondering if it's a permissions problem of some kind.
ID: 62969 · Report as offensive
KAMasud

Send message
Joined: 6 Oct 06
Posts: 204
Credit: 7,608,986
RAC: 0
Message 62970 - Posted: 18 Nov 2020, 3:04:42 UTC - in response to Message 62954.  

Llinux Mint is like that.
It's based on Ubuntu.

However ...
Don't forget the need for L3 cache. The N216 model type is hungry for that, and slow WAAAAAY down it you cram too many tasks into the workload at once.
4 Megs per model.



Thank you, Les. I am running 9th and 10th gen Intel CPU core i7. Not asking for CPDN tasks. Those I baby along and at most, only allow three tasks to run at any given time. I asked for other projects. Just dump them in VM. Rosetta and Collatz perform like land-grabbers, give an inch, they will take a mile then Boinc will not badger the CPDN server. No badgering the server, no tasks.
N216? Most probably I will get away with running only one task.
ID: 62970 · Report as offensive
KAMasud

Send message
Joined: 6 Oct 06
Posts: 204
Credit: 7,608,986
RAC: 0
Message 62971 - Posted: 18 Nov 2020, 3:17:41 UTC - in response to Message 62962.  

It may be worth trying increasing the number of days work requested above the default. I know Les didn't find anything in work fetch debug that was helpful.


Do you have a suggestion as to what this should be? I know with different hardware and different OS, the amount will be different, but right now mine is set to

Store at least 0.3 days of work.
Store up to an additional 1 day of work.

Is this ridiculous? Or is it reasonable?


Maybe, ridiculous. These WU's take three days and some hours to complete on my laptop. Now if I set a figure of .3days and 1 additional day? "AI" might take ages to figure this one out and good luck to it, I cannot. Why not set it to ten days of work plus ten additional days? May fetch you two WU's.
ID: 62971 · Report as offensive
Mr. P Hucker

Send message
Joined: 9 Oct 20
Posts: 690
Credit: 4,391,754
RAC: 6,918
Message 62972 - Posted: 18 Nov 2020, 4:22:37 UTC - in response to Message 62971.  

It may be worth trying increasing the number of days work requested above the default. I know Les didn't find anything in work fetch debug that was helpful.


Do you have a suggestion as to what this should be? I know with different hardware and different OS, the amount will be different, but right now mine is set to

Store at least 0.3 days of work.
Store up to an additional 1 day of work.

Is this ridiculous? Or is it reasonable?


Maybe, ridiculous. These WU's take three days and some hours to complete on my laptop. Now if I set a figure of .3days and 1 additional day? "AI" might take ages to figure this one out and good luck to it, I cannot. Why not set it to ten days of work plus ten additional days? May fetch you two WU's.


The buffer size should never stop you getting work as it's a minimum. I'm on 0 + 0.13 and it downloaded a CPDN per core. If Boinc wants to get CPDN work, it will ask for 0.13 days of work. That will add up to 1 WU. It won't ask for 0 WUs.
ID: 62972 · Report as offensive
Jean-David Beyer

Send message
Joined: 5 Aug 04
Posts: 1082
Credit: 16,655,848
RAC: 6,703
Message 62973 - Posted: 18 Nov 2020, 4:22:48 UTC - in response to Message 62969.  

No, you're not being too impatient. It should work. I know you installed the recommended 32bit libraries in the instructions. But do this. Download

http://www.cpdn.org/cpdnboinc/applications//hadam4_8.52_i686-pc-linux-gnu
http://www.cpdn.org/cpdnboinc/applications//hadam4_se_8.52_i686-pc-linux-gnu.zip
http://www.cpdn.org/cpdnboinc/applications//hadam4_um_8.52_i686-pc-linux-gnu.zip

into a temp directory under your home directory. Unzip the zipped files then you will likely have to chmod them +x

Then do an

ldd *

within that directory. Just to make sure that there are no unmet dependencies.

If that doesn't show anything problematic, I suggest going through the two links in my last post and see if performing any steps in those linked posts might resolve the problem. I'm wondering if it's a permissions problem of some kind.


$ ldd hadam4_8.52_i686-pc-linux-gnu
ldd: warning: you do not have execution permission for `./hadam4_8.52_i686-pc-linux-gnu'
linux-gate.so.1 (0xf7fb9000)
libpthread.so.0 => /lib/libpthread.so.0 (0xf7f7d000)
libdl.so.2 => /lib/libdl.so.2 (0xf7f78000)
libstdc++.so.6 => /lib/libstdc++.so.6 (0xf7de5000)
libm.so.6 => /lib/libm.so.6 (0xf7d13000)
libgcc_s.so.1 => /lib/libgcc_s.so.1 (0xf7cf6000)
libc.so.6 => /lib/libc.so.6 (0xf7b4f000)
/lib/ld-linux.so.2 (0xf7fbb000)


$ ldd hadam4_se_8.52_i686-pc-linux-gnu.so
linux-gate.so.1 (0xf7f87000)
libnsl.so.1 => /lib/libnsl.so.1 (0xf7e32000)
libstdc++.so.6 => /lib/libstdc++.so.6 (0xf7c9f000)
libm.so.6 => /lib/libm.so.6 (0xf7bcd000)
libgcc_s.so.1 => /lib/libgcc_s.so.1 (0xf7bb0000)
libc.so.6 => /lib/libc.so.6 (0xf7a09000)
/lib/ld-linux.so.2 (0xf7f89000)


$ ldd hadam4_um_8.52_i686-pc-linux-gnu
linux-gate.so.1 (0xf7f76000)
libdl.so.2 => /lib/libdl.so.2 (0xf7f56000)
libm.so.6 => /lib/libm.so.6 (0xf7e84000)
libpthread.so.0 => /lib/libpthread.so.0 (0xf7e63000)
libc.so.6 => /lib/libc.so.6 (0xf7cbc000)
/lib/ld-linux.so.2 (0xf7f78000)

ID: 62973 · Report as offensive
Jean-David Beyer

Send message
Joined: 5 Aug 04
Posts: 1082
Credit: 16,655,848
RAC: 6,703
Message 62974 - Posted: 18 Nov 2020, 4:30:59 UTC - in response to Message 62969.  

If that doesn't show anything problematic, I suggest going through the two links in my last post and see if performing any steps in those linked posts might resolve the problem. I'm wondering if it's a permissions problem of some kind.


Those links seem to have to do with being unable to connect to the server. But I can do that:
Tue 17 Nov 2020 03:59:34 PM EST | climateprediction.net | Sending scheduler request: To fetch work.
Tue 17 Nov 2020 03:59:34 PM EST | climateprediction.net | Requesting new tasks for CPU
Tue 17 Nov 2020 03:59:36 PM EST | climateprediction.net | Scheduler request completed: got 0 new tasks
Tue 17 Nov 2020 03:59:36 PM EST | climateprediction.net | No tasks sent
Tue 17 Nov 2020 03:59:36 PM EST | climateprediction.net | Project requested delay of 3636 seconds

ID: 62974 · Report as offensive
Profile geophi
Volunteer moderator

Send message
Joined: 7 Aug 04
Posts: 2171
Credit: 64,560,519
RAC: 352
Message 62975 - Posted: 18 Nov 2020, 5:16:00 UTC - in response to Message 62974.  


Those links seem to have to do with being unable to connect to the server. But I can do that:
[quote]Tue 17 Nov 2020 03:59:34 PM EST | climateprediction.net | Sending scheduler request: To fetch work.
Tue 17 Nov 2020 03:59:34 PM EST | climateprediction.net | Requesting new tasks for CPU
Tue 17 Nov 2020 03:59:36 PM EST | climateprediction.net | Scheduler request completed: got 0 new tasks
Tue 17 Nov 2020 03:59:36 PM EST | climateprediction.net | No tasks sent
Tue 17 Nov 2020 03:59:36 PM EST | climateprediction.net | Project requested delay of 3636 seconds

You may have already said this, but do you have any other projects active at this time? If so, I suggest suspending the other projects and allowing boinc to request work from cpdn only.
ID: 62975 · Report as offensive
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 62976 - Posted: 18 Nov 2020, 9:04:38 UTC

I'll leave off talking to the project until the results of this testing is finished.
And if real world problems will just get out of my way, I may get a chance to finish rebuilding the machine that won't work, and try again with it.
ID: 62976 · Report as offensive
WB8ILI

Send message
Joined: 1 Sep 04
Posts: 161
Credit: 81,421,805
RAC: 1,225
Message 62979 - Posted: 18 Nov 2020, 16:36:57 UTC

Here is what I do to download work.

1. Set CPDN to the number of tasks you want to run concurrently using the app_config file.
2. Set ALL projects to "No New Tasks"
3. Set the Computing Preferences to 10 days work + 10 days additional days.
4. Set CPDN to request new tasks.
5. Wait for CPDN to request new tasks (up to 1 hour).

It should now download some work unless there is some issue with your computer.

6. Set the Computing Preferences back to "normal" - for example 1 day + 0 additional.
7. Set the remaining projects back to request new work..

Now you should be good for 10 days or more (depending on your computer speed)

My observation is that when I have additional projects (besides CPDN) they keep my queue filled and when a request is sent to CPDN the queue doesn't need any more work. By briefly setting the queue to 20 days (10+10), it will always download CPDN tasks.
ID: 62979 · Report as offensive
Jean-David Beyer

Send message
Joined: 5 Aug 04
Posts: 1082
Credit: 16,655,848
RAC: 6,703
Message 62980 - Posted: 18 Nov 2020, 20:29:02 UTC - in response to Message 62979.  

Here is what I do to download work.

1. Set CPDN to the number of tasks you want to run concurrently using the app_config file.


What do I use for the application? and since there may be more than one, I assume I use multiples in here.

<app_config>
<app>
<name>Application_Name</name>
<max_concurrent>4</max_concurrent>
</app>
</app_config>
ID: 62980 · Report as offensive
Profile Dave Jackson
Volunteer moderator

Send message
Joined: 15 May 09
Posts: 4361
Credit: 16,614,059
RAC: 1,518
Message 62981 - Posted: 18 Nov 2020, 21:16:42 UTC



Here is what I do to download work.

1. Set CPDN to the number of tasks you want to run concurrently using the app_config file.



What do I use for the application? and since there may be more than one, I assume I use multiples in here.


I have never bothered with an app_config.xml
ID: 62981 · Report as offensive
WB8ILI

Send message
Joined: 1 Sep 04
Posts: 161
Credit: 81,421,805
RAC: 1,225
Message 62982 - Posted: 18 Nov 2020, 22:33:23 UTC - in response to Message 62981.  

I am using UBUNTU 20.04

To set the project limit -

1. Navigate to /var/lib/boinc-client/projects/climateprediction.net
2. Create a file named app_config.xml
3. Place the following three lines in the file.

<app_config>
<project_max_concurrent>4</project_max_concurrent>
</app_config>

4. Save it.
5. In BOINC Manager - > Options, select Read Config files
ID: 62982 · Report as offensive
Bryn Mawr

Send message
Joined: 28 Jul 19
Posts: 147
Credit: 12,825,289
RAC: 679
Message 62983 - Posted: 18 Nov 2020, 23:34:40 UTC - in response to Message 62979.  

Here is what I do to download work.

1. Set CPDN to the number of tasks you want to run concurrently using the app_config file.
2. Set ALL projects to "No New Tasks"
3. Set the Computing Preferences to 10 days work + 10 days additional days.
4. Set CPDN to request new tasks.
5. Wait for CPDN to request new tasks (up to 1 hour).

It should now download some work unless there is some issue with your computer.

6. Set the Computing Preferences back to "normal" - for example 1 day + 0 additional.
7. Set the remaining projects back to request new work..

Now you should be good for 10 days or more (depending on your computer speed)

My observation is that when I have additional projects (besides CPDN) they keep my queue filled and when a request is sent to CPDN the queue doesn't need any more work. By briefly setting the queue to 20 days (10+10), it will always download CPDN tasks.


That’s exactly what I’ve been doing and the diagnostics show that I’m requesting 12 WUs and tens of thousands of seconds work but the server allocates no work :-(
ID: 62983 · Report as offensive
WB8ILI

Send message
Joined: 1 Sep 04
Posts: 161
Credit: 81,421,805
RAC: 1,225
Message 62984 - Posted: 19 Nov 2020, 2:18:44 UTC

Bryn Mawr -

Sorry to read you can't get any work. I have 5 or 6 computers and it has worked for me every time.

You have the same OS and AMD processors as me.There must be something that I haven't encountered yet.
ID: 62984 · Report as offensive
Previous · 1 . . . 48 · 49 · 50 · 51 · 52 · 53 · 54 . . . 91 · Next

Message boards : Number crunching : New work Discussion

©2024 climateprediction.net