Panic Mode On (102) Server Problems?

Message boards : Number crunching : Panic Mode On (102) Server Problems?
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 5 · 6 · 7 · 8 · 9 · 10 · 11 . . . 25 · Next

AuthorMessage
Profile Jeff Buck Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 11 Feb 00
Posts: 1441
Credit: 148,764,870
RAC: 0
United States
Message 1763146 - Posted: 7 Feb 2016, 17:22:40 UTC - in response to Message 1763141.  

VLAR`s starting AR 0.012 not 0.12.

That doesn't seem to be quite right, either. I just checked a recent VLAR on my daily driver, WU 2052285866, 10oc15ab.18223.809244.7.34.117.vlar, and see that it has an AR of 0.021689. So the cutoff must be above that AR, at least.
ID: 1763146 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34253
Credit: 79,922,639
RAC: 80
Germany
Message 1763147 - Posted: 7 Feb 2016, 17:24:27 UTC - in response to Message 1763145.  

Hmmm, here's another;
http://setiathome.berkeley.edu/result.php?resultid=4709513420
Run time: 34 min 3 sec
CPU time: 33 min 36 sec
SETI@home using CUDA accelerated device GeForce GTX 750 Ti
setiathome v8 enhanced x41p_zi, Cuda 6.50 special
Compiled with NVCC 6.5, using 6.5 libraries. Modifications done by petri33.
Detected setiathome_enhanced_v8 task. Autocorrelations enabled, size 128k elements.
Work Unit Info:
...............
WU true angle range is : 0.064373

???


Not sure what you want to say.

VLAR`s starting AR 0.012 not 0.12.

So...All those other people are Wrong?

Let's say this. A normal task on that machine runs for 4.5 minutes and scores 64 credits, http://setiathome.berkeley.edu/result.php?resultid=4709532453
The last one of these "non-VLARs ran for 28 minutes and scored 100 credits, look above.
Now if this one also scores 100, Let's do the math;
34 divided by 4.5 = 7.5 x 64 = 480 credits for a normal task verses 100 for one of these non-VLARS
That's a 380 credit difference...
I WUZ Robbed.
How about that?


I`m not responsible for credit new.

Our in house VLAR task is PG0009, PG0036 is mid range task.
Thats a fact.


With each crime and every kindness we birth our future.
ID: 1763147 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34253
Credit: 79,922,639
RAC: 80
Germany
Message 1763148 - Posted: 7 Feb 2016, 17:28:36 UTC - in response to Message 1763146.  
Last modified: 7 Feb 2016, 17:29:49 UTC

self corrected.


With each crime and every kindness we birth our future.
ID: 1763148 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1763150 - Posted: 7 Feb 2016, 17:34:09 UTC - in response to Message 1763147.  
Last modified: 7 Feb 2016, 17:48:37 UTC

Hmmm, here's another;
http://setiathome.berkeley.edu/result.php?resultid=4709513420
Run time: 34 min 3 sec
CPU time: 33 min 36 sec
SETI@home using CUDA accelerated device GeForce GTX 750 Ti
setiathome v8 enhanced x41p_zi, Cuda 6.50 special
Compiled with NVCC 6.5, using 6.5 libraries. Modifications done by petri33.
Detected setiathome_enhanced_v8 task. Autocorrelations enabled, size 128k elements.
Work Unit Info:
...............
WU true angle range is : 0.064373

???


Not sure what you want to say.

VLAR`s starting AR 0.012 not 0.12.

So...All those other people are Wrong?

Let's say this. A normal task on that machine runs for 4.5 minutes and scores 64 credits, http://setiathome.berkeley.edu/result.php?resultid=4709532453
The last one of these "non-VLARs ran for 28 minutes and scored 100 credits, look above.
Now if this one also scores 100, Let's do the math;
34 divided by 4.5 = 7.5 x 64 = 480 credits for a normal task verses 100 for one of these non-VLARS
That's a 380 credit difference...
I WUZ Robbed.
How about that?


I`m not responsible for credit new.

Our in house VLAR task is PG0009, PG0036 is mid range task.
Thats a fact.

Here's a VLAR, WU true angle range is : 0.017926
http://setiathome.berkeley.edu/result.php?resultid=4707840067
So we Know 0.012 isn't correct. Anyone know the Real number?

Here's another, http://setiathome.berkeley.edu/result.php?resultid=4707367923
15oc15aa.30629.25032.6.33.220.vlar_0
WU true angle range is : 0.021320
ID: 1763150 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1763159 - Posted: 7 Feb 2016, 17:52:00 UTC

Sounds like something got broken in the transition.

1) At what AR should workunits be designated VLAR - i.e. have those letters stuck on the end of the name? We can look in the splitter code (and the message boards) to see what that was before v8, and what it is now.

2) Has anyone seen a task with the designation VLAR getting sent to a NVidia GPU? That would be a scheduler problem, nothing to do with the splitter.
ID: 1763159 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34253
Credit: 79,922,639
RAC: 80
Germany
Message 1763160 - Posted: 7 Feb 2016, 17:55:53 UTC
Last modified: 7 Feb 2016, 17:58:38 UTC

You might be correct TBar.

I found a wu in my database AR 0.059 marked as VLAR.

<workunit>
<workunit_header>
<name>21dc10aa.5028.1875.13.14.86.vlar</name>
<group_info>
<tape_info>
<name>21dc10aa</name>
<start_time>2455551.7086115</start_time>
<last_block_time>2455551.7086115</last_block_time>
<last_block_done>1875</last_block_done>
<missed>0</missed>
<tape_quality>0</tape_quality>
<beam>10</beam>
</tape_info>
<name>21dc10aa.5028.1875.13.14</name>
<data_desc>
<start_ra>6.0913612802179</start_ra>
<start_dec>16.851924049791</start_dec>
<end_ra>6.0939483531098</end_ra>
<end_dec>16.8607412978</end_dec>
<true_angle_range>0.05963402754185</true_angle_range>
<time_recorded>Tue Dec 21 05:00:23 2010</time_recorded>


With each crime and every kindness we birth our future.
ID: 1763160 · Report as offensive
Profile Jeff Buck Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 11 Feb 00
Posts: 1441
Credit: 148,764,870
RAC: 0
United States
Message 1763161 - Posted: 7 Feb 2016, 17:57:09 UTC - in response to Message 1763159.  
Last modified: 7 Feb 2016, 18:10:04 UTC

Whatever the cutoff was under v7, it definitely has changed for v8. I just checked my archives for December and found a VLAR (17ja11af.10230.44838.4.12.50.vlar_0) with an AR of 0.106412. That was the only one above 0.1, although I saw others in the 0.015-0.048 range.

EDIT: And in November, I found 06mr11af.21497.9235.10.12.23.vlar_0 with an AR of 0.115419, so the originally posted cutoff of 0.12 looks like it was probably correct for v7.
ID: 1763161 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1763164 - Posted: 7 Feb 2016, 18:11:16 UTC - in response to Message 1763160.  

You might be correct TBar.

I found a wu in my database AR 0.059 marked as VLAR.

<workunit>
<workunit_header>
<name>21dc10aa.5028.1875.13.14.86.vlar</name>
<group_info>
<tape_info>
<name>21dc10aa</name>
<start_time>2455551.7086115</start_time>
<last_block_time>2455551.7086115</last_block_time>
<last_block_done>1875</last_block_done>
<missed>0</missed>
<tape_quality>0</tape_quality>
<beam>10</beam>
</tape_info>
<name>21dc10aa.5028.1875.13.14</name>
<data_desc>
<start_ra>6.0913612802179</start_ra>
<start_dec>16.851924049791</start_dec>
<end_ra>6.0939483531098</end_ra>
<end_dec>16.8607412978</end_dec>
<true_angle_range>0.05963402754185</true_angle_range>
<time_recorded>Tue Dec 21 05:00:23 2010</time_recorded>

I have no idea what the numbers are for a VLAR, I was just going by what others had said.
But, I'm fairly certain I wuz robbed.
*nods head*
ID: 1763164 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1763167 - Posted: 7 Feb 2016, 18:24:51 UTC

Found several clues.

1) There's a formal definition of what a VLAR is: when the AR is less than the beam width of the telescope being used. That's 0.05 for Arecibo.

2) There's a pragmatic definition, used for 'what not to send to CUDA' That was changed by https://setisvn.ssl.berkeley.edu/trac/changeset/848 (six years ago) to 'anything below 0.12 is called VLAR' (previously, it was 'anything below 0.013' - which clearly doesn't match the formal definition)

3) The splitter code was changed on 21 January 2016 (two weeks ago) to take account of the different beam widths of the different telescopes coming online. We probably need a mathematician (or an astronomer) to take a look at https://setisvn.ssl.berkeley.edu/trac/changeset/3355 - that's probably what went wrong.
ID: 1763167 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34253
Credit: 79,922,639
RAC: 80
Germany
Message 1763168 - Posted: 7 Feb 2016, 18:24:59 UTC

Like Richard said something might have changed in V8 or there is an scheduler issue.

I`m almost certain i had a 0.029 wu a few days ago not marked as VLAR.
OTOH i have no problems with it.


With each crime and every kindness we birth our future.
ID: 1763168 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1763173 - Posted: 7 Feb 2016, 18:33:48 UTC - in response to Message 1763167.  

Um. How about this?

Changeset 3355:

145 145 // check for VLAR workunits
146 if (wugrp.data_desc.true_angle_range < 0.12) {
146 if (wugrp.data_desc.true_angle_range < 1.2*wugrp.receiver_cfg->beam_width) {
147 147 group_is_vlar=true;

If we plug in that old Arecibo value of 0.05 for the beam width, the cutoff point comes out as 0.06, rather than the previous 0.12
ID: 1763173 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13720
Credit: 208,696,464
RAC: 304
Australia
Message 1763175 - Posted: 7 Feb 2016, 18:40:23 UTC - in response to Message 1763168.  

Like Richard said something might have changed in V8 or there is an scheduler issue.

I haven't noticed anything on my systems marked as VLAR going to the GPUs, just the odd longer than estimated run times of some WUs.
And those have been over the last few days, coinciding with the high proportion of VLAR work that was available.
Grant
Darwin NT
ID: 1763175 · Report as offensive
Profile Jeff Buck Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 11 Feb 00
Posts: 1441
Credit: 148,764,870
RAC: 0
United States
Message 1763183 - Posted: 7 Feb 2016, 18:56:41 UTC - in response to Message 1763173.  

Um. How about this?

Changeset 3355:

145 145 // check for VLAR workunits
146 if (wugrp.data_desc.true_angle_range < 0.12) {
146 if (wugrp.data_desc.true_angle_range < 1.2*wugrp.receiver_cfg->beam_width) {
147 147 group_is_vlar=true;

If we plug in that old Arecibo value of 0.05 for the beam width, the cutoff point comes out as 0.06, rather than the previous 0.12

That would make sense. I'm guessing the change didn't get implemented until the outage of 26 Jan, since my archives hold a v8 VLAR from a WU created 26 Jan 2016, 4:11:04 UTC, which has an AR of 0.088912. Since then, the highest AR I can find for a VLAR is 0.036281, whereas I know I've gotten a non-VLAR with an AR of 0.063161, so 0.06 could very well be the cutoff.
ID: 1763183 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1763190 - Posted: 7 Feb 2016, 19:29:02 UTC - in response to Message 1763183.  

I've sent an email to Eric - with a note that it can wait until after the Superbowl weekend is over.
ID: 1763190 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1763266 - Posted: 7 Feb 2016, 22:26:01 UTC - in response to Message 1763190.  

I've sent an email to Eric - with a note that it can wait until after the Superbowl weekend is over.

But is this new definition really a problem? On fairly common Nvidia hardware, the tasks do complete 100%, just take about twice as long as the more common AR's the GPU's normally do. Plus, you get a commensurate increase in credit award though I acknowledge the credit is not linearly scaled. I haven't seen any .VLARs assigned to my GPUs yet by the scheduler. We won't know the full outcome I guess until the scheduler starts sending out Green Bank data.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1763266 · Report as offensive
Profile Jimbocous Project Donor
Volunteer tester
Avatar

Send message
Joined: 1 Apr 13
Posts: 1849
Credit: 268,616,081
RAC: 1,349
United States
Message 1763408 - Posted: 8 Feb 2016, 9:30:08 UTC

Hmmm ... 14 channels in progress off 7 splitters. Here we go again?
ID: 1763408 · Report as offensive
Profile BilBg
Volunteer tester
Avatar

Send message
Joined: 27 May 07
Posts: 3720
Credit: 9,385,827
RAC: 0
Bulgaria
Message 1763482 - Posted: 8 Feb 2016, 17:11:32 UTC - in response to Message 1763190.  

... until after the Superbowl weekend is over.

Forgive my ignorance ... but what is Superbowl? (it must be some kind of sport I guess)
I can Google it of course but ... why not make you a "pleasure" to tell me ;)
 


- ALF - "Find out what you don't do well ..... then don't do it!" :)
 
ID: 1763482 · Report as offensive
JLDun
Volunteer tester
Avatar

Send message
Joined: 21 Apr 06
Posts: 573
Credit: 196,101
RAC: 0
United States
Message 1763487 - Posted: 8 Feb 2016, 17:21:21 UTC - in response to Message 1763266.  
Last modified: 8 Feb 2016, 17:21:39 UTC

But is this new definition really a problem? On fairly common Nvidia hardware, the tasks do complete 100%, just take about twice as long as the more common AR's the GPU's normally do.


Hmm, not so much a problem, as an opportunity to get adjusted to a different percentage of work becoming VLAR?
ID: 1763487 · Report as offensive
Cosmic_Ocean
Avatar

Send message
Joined: 23 Dec 00
Posts: 3027
Credit: 13,516,867
RAC: 13
United States
Message 1763488 - Posted: 8 Feb 2016, 17:22:45 UTC - in response to Message 1763482.  

... until after the Superbowl weekend is over.

Forgive my ignorance ... but what is Superbowl? (it must be some kind of sport I guess)
I can Google it of course but ... why not make you a "pleasure" to tell me ;)

American football. Similar to rugby, but with many more rules. The Super Bowl is the championship game for the end of the season.
Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)
ID: 1763488 · Report as offensive
Profile JaundicedEye
Avatar

Send message
Joined: 14 Mar 12
Posts: 5375
Credit: 30,870,693
RAC: 1
United States
Message 1763520 - Posted: 8 Feb 2016, 19:17:02 UTC
Last modified: 8 Feb 2016, 19:17:27 UTC

The Super Bowl is the championship game for the end of the season.
And the annual launch of every TV commercial for the coming year........";D>

"Sour Grapes make a bitter Whine." <(0)>
ID: 1763520 · Report as offensive
Previous · 1 . . . 5 · 6 · 7 · 8 · 9 · 10 · 11 . . . 25 · Next

Message boards : Number crunching : Panic Mode On (102) 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.