Panic Mode On (97) Server Problems?

Message boards : Number crunching : Panic Mode On (97) Server Problems?
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 19 · 20 · 21 · 22 · 23 · 24 · 25 . . . 33 · Next

AuthorMessage
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1685119 - Posted: 28 May 2015, 14:04:58 UTC - in response to Message 1685088.  

The Admiral is running Windows, and the plan_class is present:

Windows/x86	7.03 (opencl_ati5_nocal)	30 May 2013, 0:18:19 UTC	7,827 GigaFLOPS
ID: 1685119 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1685128 - Posted: 28 May 2015, 14:25:29 UTC - in response to Message 1685119.  

Apparently it's Not working, or else They would have received tasks.
ID: 1685128 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1685150 - Posted: 28 May 2015, 15:18:13 UTC - in response to Message 1685128.  

Apparently it's Not working, or else They would have received tasks.

We're still waiting to see if there was a driver update between old and new (failed) work fetch requests.
ID: 1685150 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1685156 - Posted: 28 May 2015, 15:37:03 UTC - in response to Message 1685150.  
Last modified: 28 May 2015, 15:42:57 UTC

Apparently it's Not working, or else They would have received tasks.

We're still waiting to see if there was a driver update between old and new (failed) work fetch requests.

While you're waiting you might check WooHoo's results, http://setiathome.berkeley.edu/hosts_user.php?userid=9941230, and note how the NoCALs stopped on the 21st...on all three machines. Then look at the last time Admiral Gloval received a NoCAL....21st, http://setiathome.berkeley.edu/results.php?hostid=7368710&offset=60&show_names=0&state=0&appid=
I haven't been able to download a NoCAL since...the 21st, http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=2243&postid=54222#54222
Don't forget Urs, http://setiweb.ssl.berkeley.edu/beta/forum_thread.php?id=2243&postid=54205#54205

:-0
ID: 1685156 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1685159 - Posted: 28 May 2015, 15:45:09 UTC - in response to Message 1685156.  

Scheduler code update on a Thursday, at both Main and Beta? Seems unlikely, but stranger things have happened at sea. You could ask Eric if David the Tinkerman was in the lab that day.

There haven't been any scheduler (or any server code) changes committed to Git all month.
ID: 1685159 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1685164 - Posted: 28 May 2015, 15:59:29 UTC
Last modified: 28 May 2015, 16:04:31 UTC

I setup a new test host which has a non-CAL GPU(R7 250), Cat 14.4, & is only requesting GPU tasks. So far it looks like it it just sucking air.
Also I set it up to update every 310 seconds.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1685164 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1685172 - Posted: 28 May 2015, 16:10:26 UTC - in response to Message 1685164.  

I setup a new test host which has a non-CAL GPU(R7 250), Cat 14.4, & is only requesting GPU tasks. So far it looks like it it just sucking air.
Also I set it up to update every 310 seconds.

And your driver is saying 'OpenCL 1.2', unlike the OpenCL 2.0 the others have - consciously or not - upgraded to.
ID: 1685172 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1685173 - Posted: 28 May 2015, 16:12:31 UTC - in response to Message 1685164.  

That explains why WooHoo has been burning up all those APs, http://setiweb.ssl.berkeley.edu/beta/top_hosts.php
Of course, it does have it's benefits, http://setiweb.ssl.berkeley.edu/beta/top_teams.php
Release the NoCALs!
ID: 1685173 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1685184 - Posted: 28 May 2015, 16:29:57 UTC

And why is it, with 9 rigs running 24/7, that I have never had any problems with them reporting or tasks, or much else?
Am I just that lucky, or what?
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1685184 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1685189 - Posted: 28 May 2015, 16:36:15 UTC - in response to Message 1685184.  
Last modified: 28 May 2015, 16:39:55 UTC

And why is it, with 9 rigs running 24/7, that I have never had any problems with them reporting or tasks, or much else?
Am I just that lucky, or what?

Your rigs all have NVidia cards, and you haven't updated the drivers recently. The working assumption is that something has changed with the ATI/AMD GPU drivers, and the way SETI responds to work requests when the CAL driver isn't reported. Eric is looking into the Beta server logs.
ID: 1685189 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1685190 - Posted: 28 May 2015, 16:38:54 UTC - in response to Message 1685189.  

And why is it, with 9 rigs running 24/7, that I have never had any problems with them reporting or tasks, or much else?
Am I just that lucky, or what?

Your rigs all have NVidia cards, and you haven't updated the drivers recently. The working assumption is that something has changed with the ATI/AMD GPU drivers, and the way SETI responds to work responds when the CAL driver isn't reported. Eric is looking into the Beta server logs.


Well, there ya have it in a nutshell.
If it ain't broken, I don't fix it.
NV and existing drivers work, and I don't update a thing unless something stops working.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1685190 · Report as offensive
woohoo
Volunteer tester

Send message
Joined: 30 Oct 13
Posts: 972
Credit: 165,671,404
RAC: 5
United States
Message 1685208 - Posted: 28 May 2015, 17:13:13 UTC

I think it might be working now.

I've been running 15.4 beta drivers since last month with no problems until last week. I can't run really old drivers because hawaii and 295x2 support only goes back so far. I try to stay on the latest drivers to see if the problem where the first gpu gets detected as opencl 2.0 and the second gpu gets detected as opencl 1.2 will ever get fixed. I set it to use all gpus to get around that problem.
ID: 1685208 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1685215 - Posted: 28 May 2015, 17:31:47 UTC - in response to Message 1685208.  

I think it might be working now.

I've been running 15.4 beta drivers since last month with no problems until last week. I can't run really old drivers because hawaii and 295x2 support only goes back so far. I try to stay on the latest drivers to see if the problem where the first gpu gets detected as opencl 2.0 and the second gpu gets detected as opencl 1.2 will ever get fixed. I set it to use all gpus to get around that problem.

And Hal's test machine has just got some, too. Eric found a typo in the plan class specification file.
ID: 1685215 · Report as offensive
woohoo
Volunteer tester

Send message
Joined: 30 Oct 13
Posts: 972
Credit: 165,671,404
RAC: 5
United States
Message 1685217 - Posted: 28 May 2015, 17:35:12 UTC

I suppose using lunatics would get around the problem as well. But I like stock so that the apps update automatically. Plus I'm trying to see what level my output peaks at on stock before I try comparing with optimized.
ID: 1685217 · Report as offensive
Profile betreger Project Donor
Avatar

Send message
Joined: 29 Jun 99
Posts: 11361
Credit: 29,581,041
RAC: 66
United States
Message 1685225 - Posted: 28 May 2015, 17:48:22 UTC - in response to Message 1685190.  

I don't update a thing unless something stops working.

Amen
ID: 1685225 · Report as offensive
Profile HAL9000
Volunteer tester
Avatar

Send message
Joined: 11 Sep 99
Posts: 6534
Credit: 196,805,888
RAC: 57
United States
Message 1685227 - Posted: 28 May 2015, 17:54:24 UTC - in response to Message 1685172.  

I setup a new test host which has a non-CAL GPU(R7 250), Cat 14.4, & is only requesting GPU tasks. So far it looks like it it just sucking air.
Also I set it up to update every 310 seconds.

And your driver is saying 'OpenCL 1.2', unlike the OpenCL 2.0 the others have - consciously or not - upgraded to.

OpenCL 2.0 support is only in the GCN 1.1 & 1.2 based cards. Which includes the R7 260s (Bonaire), R9 280s, & R9 290s. I think the driver that added OpenCL 2.0 support was Cat 14.12, 1642.5 (VM).

However it looks like someone found the issue.
SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[
ID: 1685227 · Report as offensive
woohoo
Volunteer tester

Send message
Joined: 30 Oct 13
Posts: 972
Credit: 165,671,404
RAC: 5
United States
Message 1685228 - Posted: 28 May 2015, 17:55:55 UTC

I always hope that an update can increase performance even by 1% but that usually applies to games and not science apps.
ID: 1685228 · Report as offensive
Admiral Gloval
Avatar

Send message
Joined: 31 Mar 13
Posts: 20274
Credit: 5,308,449
RAC: 0
United States
Message 1685289 - Posted: 28 May 2015, 20:01:51 UTC

The 7.03 wu's are back. Yes. Looks like a mod has been made to the app. It now uses even less cpu power (0.0322). Thanks for the return.

ID: 1685289 · Report as offensive
Profile ivan
Volunteer tester
Avatar

Send message
Joined: 5 Mar 01
Posts: 783
Credit: 348,560,338
RAC: 223
United Kingdom
Message 1685308 - Posted: 28 May 2015, 20:44:03 UTC - in response to Message 1685190.  

If it ain't broken, I don't fix it.
Attrib: Ronald Reagan

If we can't fix it, it ain't broke!
Attrib: USMC Engineers
ID: 1685308 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1685395 - Posted: 29 May 2015, 1:47:26 UTC - in response to Message 1685308.  

If it ain't broken, I don't fix it.
Attrib: Ronald Reagan

If we can't fix it, it ain't broke!
Attrib: USMC Engineers

If you are not fixing it, best get a bigger f'ing hammer, kids,,,LOL.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1685395 · Report as offensive
Previous · 1 . . . 19 · 20 · 21 · 22 · 23 · 24 · 25 . . . 33 · Next

Message boards : Number crunching : Panic Mode On (97) Server Problems?


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