Gamma-ray pulsar binary search #1 on GPUs

Harri Liljeroos
Harri Liljeroos
Joined: 10 Dec 05
Posts: 4358
Credit: 3215683880
RAC: 2035060

Here's the message log from

Here's the message log from the first failed task:

14-Dec-2016 18:30:35 [Einstein@Home] [sched_op] Starting scheduler request
 14-Dec-2016 18:30:35 [Einstein@Home] Sending scheduler request: To fetch work.
 14-Dec-2016 18:30:35 [Einstein@Home] Requesting new tasks for NVIDIA GPU
 14-Dec-2016 18:30:35 [Einstein@Home] [sched_op] CPU work request: 0.00 seconds; 0.00 devices
 14-Dec-2016 18:30:35 [Einstein@Home] [sched_op] NVIDIA GPU work request: 108418.09 seconds; 2.00 devices
 14-Dec-2016 18:30:35 [Einstein@Home] [sched_op] Intel GPU work request: 0.00 seconds; 0.00 devices
 14-Dec-2016 18:30:38 [Einstein@Home] Scheduler request completed: got 2 new tasks
 14-Dec-2016 18:30:38 [Einstein@Home] [sched_op] Server version 611
 14-Dec-2016 18:30:38 [Einstein@Home] Project requested delay of 60 seconds
 14-Dec-2016 18:30:38 [Einstein@Home] [sched_op] estimated total CPU task duration: 0 seconds
 14-Dec-2016 18:30:38 [Einstein@Home] [sched_op] estimated total NVIDIA GPU task duration: 18516 seconds
 14-Dec-2016 18:30:38 [Einstein@Home] [sched_op] estimated total Intel GPU task duration: 0 seconds
 14-Dec-2016 18:30:38 [Einstein@Home] [sched_op] Deferring communication for 00:01:00
 14-Dec-2016 18:30:38 [Einstein@Home] [sched_op] Reason: requested by project
 14-Dec-2016 18:30:40 [Einstein@Home] Started download of hsgamma_FGRPB1G_1.15_windows_x86_64__FGRPopencl-Beta-nvidia
 14-Dec-2016 18:30:40 [Einstein@Home] Started download of LATeah2003L.dat
 14-Dec-2016 18:30:42 [Einstein@Home] Finished download of LATeah2003L.dat
 14-Dec-2016 18:30:42 [Einstein@Home] Started download of templates_LATeah2003L_0044_9648.dat
 14-Dec-2016 18:30:43 [Einstein@Home] Finished download of hsgamma_FGRPB1G_1.15_windows_x86_64__FGRPopencl-Beta-nvidia
 14-Dec-2016 18:30:43 [Einstein@Home] Finished download of templates_LATeah2003L_0044_9648.dat
 14-Dec-2016 18:30:43 [Einstein@Home] Started download of templates_LATeah2003L_0044_9612.dat
 14-Dec-2016 18:30:44 [Einstein@Home] Finished download of templates_LATeah2003L_0044_9612.dat
 14-Dec-2016 18:31:38 [Einstein@Home] Starting task LATeah2003L_44.0_0_-5e-12_9612_0
 14-Dec-2016 18:31:38 [Einstein@Home] [cpu_sched] Starting task LATeah2003L_44.0_0_-5e-12_9612_0 using hsgamma_FGRPB1G version 115 (FGRPopencl-Beta-nvidia) in slot 0
 14-Dec-2016 18:31:38 [Einstein@Home] [sched_op] Deferring communication for 00:01:31
 14-Dec-2016 18:31:38 [Einstein@Home] [sched_op] Reason: Unrecoverable error for task LATeah2003L_44.0_0_-5e-12_9612_0
 14-Dec-2016 18:31:38 [Einstein@Home] Computation for task LATeah2003L_44.0_0_-5e-12_9612_0 finished
 14-Dec-2016 18:31:38 [Einstein@Home] Output file LATeah2003L_44.0_0_-5e-12_9612_0_0 for task LATeah2003L_44.0_0_-5e-12_9612_0 absent
 14-Dec-2016 18:31:38 [Einstein@Home] Output file LATeah2003L_44.0_0_-5e-12_9612_0_1 for task LATeah2003L_44.0_0_-5e-12_9612_0 absent
 14-Dec-2016 18:31:39 [Einstein@Home] Starting task h1_0625.80_O1C02Cl1In0C__O1MD1CV_CasA1_626.05Hz_21_1
 14-Dec-2016 18:31:39 [Einstein@Home] [cpu_sched] Starting task h1_0625.80_O1C02Cl1In0C__O1MD1CV_CasA1_626.05Hz_21_1 using einstein_O1MD1CV version 100 (AVX) in slot 11
 14-Dec-2016 18:31:39 [Einstein@Home] Process creation failed: Access is denied.

 (0x5) - error code 5 (0x5)
 14-Dec-2016 18:31:39 [Einstein@Home] Process creation failed: Access is denied.

 (0x5) - error code 5 (0x5)
 14-Dec-2016 18:31:40 [Einstein@Home] Process creation failed: Access is denied.

 (0x5) - error code 5 (0x5)
 14-Dec-2016 18:31:40 [Einstein@Home] Process creation failed: Access is denied.

 (0x5) - error code 5 (0x5)
 14-Dec-2016 18:31:41 [Einstein@Home] Process creation failed: Access is denied.

 (0x5) - error code 5 (0x5)
 14-Dec-2016 18:31:42 [Einstein@Home] [sched_op] Deferring communication for 00:03:06
 14-Dec-2016 18:31:42 [Einstein@Home] [sched_op] Reason: Unrecoverable error for task LATeah2003L_44.0_0_-5e-12_9648_0
 14-Dec-2016 18:31:42 [Einstein@Home] Computation for task LATeah2003L_44.0_0_-5e-12_9648_0 finished
 14-Dec-2016 18:31:42 [Einstein@Home] Output file LATeah2003L_44.0_0_-5e-12_9648_0_0 for task LATeah2003L_44.0_0_-5e-12_9648_0 absent
 14-Dec-2016 18:31:42 [Einstein@Home] Output file LATeah2003L_44.0_0_-5e-12_9648_0_1 for task LATeah2003L_44.0_0_-5e-12_9648_0 absent
 14-Dec-2016 18:31:42 [Einstein@Home] [cpu_sched] Preempting h1_0625.80_O1C02Cl1In0C__O1MD1CV_VelaJr1_626.05Hz_14_1 (left in memory)

The application file was downloaded without .exe at the end and that's how it is at the moment in the einstein directory.

EDIT: While I was typing this I got a pop-up window from Avast saying that the hsgamma_FGRPB1G_1.15_windows_x86_64__FGRPopencl-Beta-nvidia
file is clean. I don't know how this was activated now hours after the failed tasks (I didn't ask it to be scanned and Boinc folder should be excluded from scanning) but maybe the access denied error came because Avast had the file locked.

Mad_Max
Mad_Max
Joined: 2 Jan 10
Posts: 154
Credit: 2215048038
RAC: 430914

Yes, "Process creation

Yes, "Process creation failed: Access is denied" error usually caused by OS or antivirus/security software block files and BOINC can not start new process.

I getting such errors with every new app on one of high security host where all new executables forbidden for execution by default until white-listed.

TimeLord04
TimeLord04
Joined: 8 Sep 06
Posts: 1442
Credit: 72378840
RAC: 0

[Update:] 12-14-2016 at 9:50

[Update:]

12-14-2016 at 9:50 PM - PST

The MAC OS X OpenCL Bug noticed by SETI Members and App Developers is now prominent here.  I've just picked up two more Invalids.  (Total of three now.)

Apparently here, (as at SETI), any MAC with "Darwin 15.4.0" or newer can exhibit this OpenCL issue.  TBar confirmed that it is mainly with MAC Laptops; but, also occurs in Hackintosh, (like mine), and some Desktop MACs.  Anyone running MAC OS X El Capitan 10.11.4 or newer should be aware of increased Inconclusives and Invalids running Einstein's new OpenCL WU Apps.

At SETI, it was determined that these MACs should be using newer/newly developed CUDA Apps.  TBar, (as mentioned in a prior Post), created a CUDA42 App, a CUDA75 App, (which I run now when crunching SETI), and a Special CUDA75 App developed by Petri33.  Einstein App Developers here may wish to make contact with TBar and/or Petri33 to possibly help come up with CUDA75 Apps for Einstein on these newer Units.  (If this is even possible.)

I'd hate to see continued rising numbers of Invalids here with these new Units.  As always, I will continue to monitor my MAC/Hackintosh.

 

TL

TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join SETI Refugees

Harri Liljeroos
Harri Liljeroos
Joined: 10 Dec 05
Posts: 4358
Credit: 3215683880
RAC: 2035060

It turned out that Boinc

It turned out that Boinc datafolder was not excluded from Avast scanning. I changed that and during the night more tasks were downloaded and two has already finished calculation. One is now validated and the other is pending. So looks like that is solved.

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4312
Credit: 250609256
RAC: 34665

TimeLord04

TimeLord04 wrote:

[Update:]

12-14-2016 at 9:50 PM - PST

The MAC OS X OpenCL Bug noticed by SETI Members and App Developers is now prominent here.  I've just picked up two more Invalids.  (Total of three now.)

Apparently here, (as at SETI), any MAC with "Darwin 15.4.0" or newer can exhibit this OpenCL issue.

Thanks for the note. However my current priorities are clearly on getting out an efficient Windows version, I'm not sure I have time to address these OSX issues still this year.

 

BM

Kailee71
Kailee71
Joined: 22 Nov 16
Posts: 35
Credit: 42623563
RAC: 0

Hi all, can I just again ask

Hi all,

can I just again ask if anyone is running any GTX 1050Ti? Would like to compare these very lean GPUs with my (quick but very) powerhungry 280x...

Many thanks in advance,

 

Kailee71.

Robert G
Robert G
Joined: 6 Nov 16
Posts: 2
Credit: 11918033
RAC: 0

No problems here

No problems here with hsgamma_FGRPB1G_1.14_x86_64-apple-darwin__FGRPopencl-nvidia-mav and Darwin 16.3.0.

Only with V1.12 i got invalid results / error while computing, and only 1 invalid result with V 1.13.

No crashes with 2 980 TI, and running 2 WUs per card without problems. 

Alexander Favorsky
Alexander Favorsky
Joined: 18 Jun 16
Posts: 36
Credit: 176576035
RAC: 76300

Hi everyone!Started

Hi everyone!

Started receiving FGRPopencl-Beta-nvidia WUs for Windows and they all errored out.

https://einsteinathome.org/host/12298595/tasks/error

15.12.2016 17:39:18 | Einstein@Home | Output file LATeah2003L_44.0_0_-7.5e-11_7848_0_0 for task LATeah2003L_44.0_0_-7.5e-11_7848_0 absent
15.12.2016 17:39:18 | Einstein@Home | Output file LATeah2003L_44.0_0_-7.5e-11_7848_0_1 for task LATeah2003L_44.0_0_-7.5e-11_7848_0 absent

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4312
Credit: 250609256
RAC: 34665

The relevant message is in

The relevant message is in stderr: "Error during OpenCL FFT (error: -36)". You got occasional display driver resets.

BM

Alexander Favorsky
Alexander Favorsky
Joined: 18 Jun 16
Posts: 36
Credit: 176576035
RAC: 76300

I guess there is some problem

I guess there is some problem with these WUs since all other WUs finish without problems.

Anyway, I'll try to update video driver.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.