2020 Formula Boinc Sprints

Page 12 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Fardringle

Diamond Member
Oct 23, 2000
9,184
753
126
I think I'm going to pass on this sprint. My GPU resources are rather limited and didn't do well the last time I tried to run Einstein on it.
 

Assimilator1

Elite Member
Nov 4, 1999
24,120
507
126
I keep getting errors on Gravitational Wave search WUs.
Funny you mention that, my 2nd rig seems to have done only 1 WU and that errored out, (the task details shows it errored with someone elses machine to) but my main rig is doing just fine (although they are gamma pulse search WUs), it's currently doing a grav wave search one, will be interesting to see how that goes.....

Anyone else having problems?
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
16,218
3,796
75
Funny you mention that, my 2nd rig seems to have done only 1 WU and that errored out, (the task details shows it errored with someone elses machine to) but my main rig is doing just fine (although they are gamma pulse search WUs), it's currently doing a grav wave search one, will be interesting to see how that goes.....

Anyone else having problems?
It seemed to me that:

1. Doing two grav WUs at a time doesn't work well on a 3GB card. Apparently it needs more than 2GB per WU. :eek:
2. Running some gravs, then some pulses (2 at a time), then some more gravs is what seemed to cause the problem for me in the first place. Edit: A reboot appeared to fix whatever went wrong. Maybe it's my old pre-400 drivers?
3. I took a look at points from each task type. Pulses produce 3,465 points each. Gravs produce only 2,000. Since it takes me longer to run two gravs than two pulses, I'm sticking with pulses now.
 

Assimilator1

Elite Member
Nov 4, 1999
24,120
507
126
I only run 1WU at a time, so far on my 2nd rig all WUs (only 3) have errored out, and to add insult to injury each grav WU on that rig is taking 21,168s!! Bearing in mind the HD 7870XT has much more DP power than the RX 580, 710 GFLOPs vs 386.
I'm going to check the Einstein forums and see if I can update the drivers.
Drivers were 19.9.1 (from last year), trying 20.4.2

It had done about 1/2 of the 1st WU with the new drivers in about 10 mins, so it looks like it's on track now :), although after a reboot all that progress was lost! o_O
 
Last edited:

Assimilator1

Elite Member
Nov 4, 1999
24,120
507
126
Nope, I spoke too soon, and I realised that the progress hadn't been lost, but that the ETA had shot up & the progress bar had been revised.
Remaining time is 19hrs 40mins and climbing fast! So it's going to hit the time limit again.


Outcome: Computation error

Client state: Compute error

Exit status: 197 (0x000000C5) EXIT_TIME_LIMIT_EXCEEDED


So until/if I can fix this, I'm going to shut that rig down as it's a total waste of time & power :(.

Any ideas anyone? (I've asked in the Einstein forum too JFYI).
 

StefanR5R

Elite Member
Dec 10, 2016
5,459
7,717
136
@Assimilator1, disable the bad application in your project preferences. You should still get good work from at least one other of the remaining applications.

----------------

Compared to yesteryear's sprint at Einstein@home, this year the output of teams who are not actively taking part in Formula BOINC should be considerably higher. As far as I have seen, many former SETI@home contributors moved on to Einstein@home.
 

Endgame124

Senior member
Feb 11, 2008
954
669
136
My RX250 is chewing through WUs around every 90 minutes, so far with minimal problems. You guys just need older hardware!
 
  • Haha
Reactions: lane42

Orange Kid

Elite Member
Oct 9, 1999
4,323
2,110
146
Einstein@home day one-ish

1XtremeSystems2560,031,121
2The Scottish Boinc Team1823,081,357
3Gridcoin1519,580,679
4BOINC@AUSTRALIA1216,503,436
5Team China1014,277,402
6UK BOINC Team814,178,933
7Russia Team613,867,617
8SETI.Germany49,882,747
9Rechenkraft.net29,261,376
10TeAm AnandTech18,872,445

From FreeDC

 

Endgame124

Senior member
Feb 11, 2008
954
669
136
Einstein@home day one-ish

1XtremeSystems2560,031,121
2The Scottish Boinc Team1823,081,357
3Gridcoin1519,580,679
4BOINC@AUSTRALIA1216,503,436
5Team China1014,277,402
6UK BOINC Team814,178,933
7Russia Team613,867,617
8SETI.Germany49,882,747
9Rechenkraft.net29,261,376
10TeAm AnandTech18,872,445

From FreeDC

Look at me rocking it with (checks score) 313 points! Something seems not right
 

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
16,218
3,796
75
Look at me rocking it with (checks score) 313 points! Something seems not right
Looks like you loaded up on Gravity Wave WUs. They're all pending a wingman except for the two that took an extra-long time and failed. :(
If you want POINTS, say for a Sprint, one should only be running the Gamma-ray pulsar binary search #1 (GPU) . Gravitational wave points are horrific in comparison.
Your only current results with points are from your ARMs. 💪💪💪 ;)
 

Assimilator1

Elite Member
Nov 4, 1999
24,120
507
126
Looks like with some of them at least you having the same problems as my 2nd rig was :(.

I had to disable that app to get around it for now, I've posted in the Einstein forums about it here.
 

Assimilator1

Elite Member
Nov 4, 1999
24,120
507
126
That would be awesome :cool:

Btw those with grx cards with 3GB of vRAM of less, you will have problems with some of the grav wave WUs, specifically those with a frequency range(gap) of 0.7Hz or more (shown in the Task ID). Previously it errored out stating MEM_OBJECT_ALLOCATION_FAILURE
But it seems now to give an error referring to '... caused breakpoint at location....', other than disabling that specific app or manually (!) filtering out those WUs, I don't yet know if theirs a fix for it.

@Endgame124
Your GPU is also having some errors with the above type WUs, but I don't see why as it states your grx card has 7GB of vRAMo_O, what grx card have you got? [edit] Ah, am I right in thinking that seeing as it's an APU it has an embedded GPU and uses system RAM? Or, at least the vRAM it has is 3GB or less? [edit2] No, it seems you do have a seperate grx card too, I don't get it! lol
 
Last edited:

Endgame124

Senior member
Feb 11, 2008
954
669
136
Endgame
Your GPU is also having some errors with the above type WUs, but I don't see why as it states your grx card has 7GB of vRAMo_O, what grx card have you got?
I have 2 graphics cards in this box.

1) generic Asus R7 250X 1GB model.

2) A10-7870K Integrated (Kaveri) R250 with maximum integrated memory split enabled (should be 8GB). I also used my spare watercooling parts on this host, so the integrated graphcis card is overclocked like 40% - the CPU is overclocked pretty heavily as well, but it is hard capped to 3Ghz when the integrated video is running. The watercooling on this host is total overkill - it hangs out at 28.6c under load when using the integrated GPU.
 

Assimilator1

Elite Member
Nov 4, 1999
24,120
507
126
Cool! Literally :D
Is their a way to tell which GPU the 0.7Hz+ WUs are failing on? I'm guessing it's the 250X, but can we see....... I guess the 'GPU type 2' in the task details says it, is that your R7 250X?
 

Endgame124

Senior member
Feb 11, 2008
954
669
136
Cool! Literally :D
Is their a way to tell which GPU the 0.7Hz+ WUs are failing on? I'm guessing it's the 250X, but can we see....... I guess the 'GPU type 2' in the task details says it, is that your R7 250X?
I have no idea how to tell or how to control the second GPU in boinc. F@H lists the integrated video as GPU0, and the 250X as GPU1, so I would Maybe guess type 2 is the 250X?

My pc patched and started up F@H on the 250X, so maybe I just leave that alone for a while. The Einstein task assigned to that gpu now has an eta 3 days from now.
 

Assimilator1

Elite Member
Nov 4, 1999
24,120
507
126
Yea looks like the 570 is the No2 GPU then. You could keep E@H going on that one, you would just have to disable the grav wave app, as Stefan mentioned.
 
Last edited:

Ken g6

Programming Moderator, Elite Member
Moderator
Dec 11, 1999
16,218
3,796
75
I have no idea how to tell or how to control the second GPU in boinc. F@H lists the integrated video as GPU0, and the 250X as GPU1, so I would Maybe guess type 2 is the 250X?

My pc patched and started up F@H on the 250X, so maybe I just leave that alone for a while. The Einstein task assigned to that gpu now has an eta 3 days from now.
To disable the internal graphics, you should be able to edit your cc_config.xml file in your BOINC directory, and change the <use_all_gpus> setting to 0.
 

Endgame124

Senior member
Feb 11, 2008
954
669
136
To disable the internal graphics, you should be able to edit your cc_config.xml file in your BOINC directory, and change the <use_all_gpus> setting to 0.
The internal graphics are working fine. How do I disable the external 250X?
 

biodoc

Diamond Member
Dec 29, 2005
6,257
2,238
136
The internal graphics are working fine. How do I disable the external 250X?

From this site here, there are a couple of options you can include in "options" section in your cc_config.xml file

<ignore_ati_dev>N</ignore_ati_dev> <----replace "N" with the appropriate device #. In your case "1" is correct I think.

Another possibility is to use <exclude_gpu> which is much more selective. You can restrict it to a specific project and application.

<exclude_gpu>
<url>project_URL</url>
[<device_num>N</device_num>]
[<type>NVIDIA|ATI|intel_gpu</type>]
[<app>appname</app>]
</exclude_gpu>

Here is one of my cc_config.xml files where I included the <ignore_ati_dev>1</ignore_ati_dev> option:

<cc_config>
<log_flags>
<task>1</task>
<file_xfer>1</file_xfer>
<sched_ops>1</sched_ops>
</log_flags>
<options>
<allow_multiple_clients>1</allow_multiple_clients>
<allow_remote_gui_rpc>1</allow_remote_gui_rpc>
<report_results_immediately>1</report_results_immediately>
<no_alt_platform>1</no_alt_platform>
<ignore_ati_dev>1</ignore_ati_dev>
</options>
</cc_config>
 
  • Like
Reactions: Endgame124

Orange Kid

Elite Member
Oct 9, 1999
4,323
2,110
146
Einstein@home day two
I think the real fireworks will be tomorrow;)
Happy 4th to those who celebrate.
Stay safe and try to keep all your fingers and other body parts intact.

1XtremeSystems25101,635,248
2The Scottish Boinc Team1844,213,876
3Gridcoin1536,726,232
4UK BOINC Team1235,121,544
5BOINC@AUSTRALIA1027,989,039
6Team China826,472,040
7Russia Team625,713,458
8TeAm AnandTech421,356,032
9Rechenkraft.net218,542,159
10SETI.Germany117,481,218



From FreeDC

biodoc4,819,8151,732,500
6,552,315​
crashtech2,783,890691,635
3,475,525​
Lane421,233,540852,390
2,085,930​
10esseeTony859,320620,235
1,479,555​
Icecold-Team Anandtech689,315544,205
1,233,520​
Orange Kid557,620388,650
946,270​
Ken_g6 - TeAm Anandtech231,462159,731
391,193​
elawyn144,930187,530
332,460​
phoenicis259,87558,905
318,780​
Howdy218,29576,230
294,525​
[TA]Assimilator1177,47628,930
206,406​
Ingleside48,51062,370
110,880​
Jesper S. Rytke34,00024,000
58,000​
Alan J. Simpson16,75731,878
48,635​
zzuupp42,0000
42,000​
Kiska22,0000
22,000​
Endgame124125313
438​