Computation Errors

Message boards : Number crunching : Computation Errors
Message board moderation

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
entigy

Send message
Joined: 1 Jul 10
Posts: 11
Credit: 419,311,588
RAC: 658,571
Message 1054 - Posted: 19 Sep 2018, 21:59:21 UTC

Started getting errors tonight - WUs are failing after just 6 or 7 seconds.

19/09/2018 22:50:01 | collatz | Starting task collatz_sieve_1cf6ab45-abae-4dea-b0a5-9ee61bb157bc_0
19/09/2018 22:50:07 | collatz | Computation for task collatz_sieve_1cf6ab45-abae-4dea-b0a5-9ee61bb157bc_0 finished
19/09/2018 22:50:07 | collatz | Output file collatz_sieve_1cf6ab45-abae-4dea-b0a5-9ee61bb157bc_0_r575015586_0 for task collatz_sieve_1cf6ab45-abae-4dea-b0a5-9ee61bb157bc_0 absent
19/09/2018 22:52:08 | collatz | Sending scheduler request: To report completed tasks.
19/09/2018 22:52:08 | collatz | Reporting 1 completed tasks

This is with a GTX 1060 6Gb card.
My machine did update the Nvidia drivers to 411.63 tonight - could that be the culprit ???
ID: 1054 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
TimeLord04
Avatar

Send message
Joined: 30 Aug 18
Posts: 271
Credit: 220,927,978
RAC: 2
Message 1055 - Posted: 20 Sep 2018, 3:27:18 UTC
Last modified: 20 Sep 2018, 3:35:06 UTC

IF Win 10 is your OS, AND IF Win 10 did the Upgrade on the Drivers, then you are now missing OpenCL and CUDA Support for your 1060 card.

It's well established over at SETI@Home Number Crunching Forum that Win 10 KILLS OpenCL and CUDA Support for ALL NVIDIA and ATI cards when IT chooses to perform a Windows Driver Update! INSTEAD, grab the same Driver Version from the Manufacturer of your card, (ie: EVGA, go to EVGA.com and DL the proper Driver), and grab the WHQL Driver listed in the Site's Download Section.

It's also recommended to "force" Win 10 to NOT Upgrade "Drivers" of ANY kind. (Setting is buried, I don't know where...) AND, you will need to periodically check on Win 10 to make sure it is STILL set to NOT Update "Drivers", as during a MAJOR Win 10 Update that your preferred Settings will get overridden back to Win 10 Default Settings which in turn will make Win 10 grab the "latest" Win 10 NVIDIA Driver... Rinse, and Repeat...

O&O Shut Up Win 10 program helps to keep things locked down, but even with that, MAJOR Win 10 Updates can override that program, too, and then ONCE AGAIN Win 10 grabs the "latest" Win 10 NVIDIA Driver - AGAIN killing your OpenCL and CUDA Support...

Good luck.


TL

[EDIT:]

AND, IF you are on Win 10 Home, you may NOT have the ability to "force" Win 10 to NOT Update "Drivers"...

Win 10 Pro gives more control, and can let you "Defer" Updates, too... IF you MUST use Win 10 I recommend AT LEAST "Pro" Flavor for the 'better' control for the User. ($99 Upgrade from Home to Pro.)


TL
Have TARDIS will travel!!!
ID: 1055 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
TimeLord04
Avatar

Send message
Joined: 30 Aug 18
Posts: 271
Credit: 220,927,978
RAC: 2
Message 1057 - Posted: 20 Sep 2018, 5:44:17 UTC

One more thing... When installing the proper WHQL Driver, choose Custom and select "Clean Install".


TL
Have TARDIS will travel!!!
ID: 1057 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
entigy

Send message
Joined: 1 Jul 10
Posts: 11
Credit: 419,311,588
RAC: 658,571
Message 1060 - Posted: 20 Sep 2018, 17:37:12 UTC

GeForce Experience alerted me to the new drivers, and I did a custom installation via that (not a 'clean install' though).

I'm running Win 10 Pro, and have 'driver updates' disabled via Group Policy.
ID: 1060 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
TimeLord04
Avatar

Send message
Joined: 30 Aug 18
Posts: 271
Credit: 220,927,978
RAC: 2
Message 1061 - Posted: 20 Sep 2018, 18:01:28 UTC - in response to Message 1060.  
Last modified: 20 Sep 2018, 18:26:05 UTC

GeForce Experience alerted me to the new drivers, and I did a custom installation via that (not a 'clean install' though).

I'm running Win 10 Pro, and have 'driver updates' disabled via Group Policy.

AWESOME!!! Glad you are on Pro and NOT Home...

Are you familiar with Display Driver Uninstaller, "DDU"? Display Driver Uninstaller 18.0.0.0

This will cleanly remove ALL Video Driver vestiges from your system... Use in Safe Mode, and after running DDU, reboot to "Normal" mode, then perform a Custom and Clean Install of your NVIDIA Drivers.

This should eliminate any Driver conflict issues and restore OpenCL and CUDA and Full Driver Support for BOINC.


Good Luck,


TL

[EDIT:]

Also, it may be possible that GeForce Experience ONLY grabbed "Gaming Drivers"... (I don't think this should be the case, but one never knows 100%...) It wouldn't hurt to go to your Manufacturer's Site and manually Download the Full WHQL Driver Version before running DDU and reinstalling the Drivers.

[EDIT 2:]

NVIDIA Driver Link:

NVIDIA WHQL Drivers.

EVGA NVIDIA Driver Link:

EVGA WHQL Drivers.


TL
Have TARDIS will travel!!!
ID: 1061 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
entigy

Send message
Joined: 1 Jul 10
Posts: 11
Credit: 419,311,588
RAC: 658,571
Message 1062 - Posted: 20 Sep 2018, 18:27:06 UTC

Yeah, I've used DDU many times.....

BOINC seems happy enough that I have a GPU installed:

20/09/2018 19:12:35 | | CUDA: NVIDIA GPU 0: GeForce GTX 1060 6GB (driver version 411.63, CUDA version 10.0, compute capability 6.1, 4096MB, 3562MB available, 4730 GFLOPS peak)
20/09/2018 19:12:35 | | OpenCL: NVIDIA GPU 0: GeForce GTX 1060 6GB (driver version 411.63, device version OpenCL 1.2 CUDA, 6144MB, 3562MB available, 4730 GFLOPS peak)

I started one WU as a test and it failed with the following error:

<core_client_version>7.12.1</core_client_version>
<![CDATA[
<message>
Access is denied.
(0x5) - exit code 5 (0x5)</message>
<stderr_txt>
Collatz Conjecture Sieve 1.30 Windows x86_64 for OpenCL
Written by Slicker (Jon Sonntag) of team SETI.USA
Based on the AMD Brook+ kernels by Gipsel of team Planet 3DNow!
Sieve code and OpenCL optimization provided by Sosiris of team BOINC@Taiwan
Processor Type NVIDIA
Device Vendor NVIDIA Corporation
Name GeForce GTX 1060 6GB
Driver Version 411.63
OpenCL Version OpenCL 1.2 CUDA
Error: GPU steps do not match CPU steps. Workunit processing aborted.
19:13:03 (9228): called boinc_finish(5)

</stderr_txt>
]]>


Incidentally, my other PC (which has a GTX 1070, and which also updated its drivers last night) is merrily chugging through some Seti units at the moment....
ID: 1062 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
TimeLord04
Avatar

Send message
Joined: 30 Aug 18
Posts: 271
Credit: 220,927,978
RAC: 2
Message 1063 - Posted: 20 Sep 2018, 19:00:08 UTC
Last modified: 20 Sep 2018, 19:04:50 UTC

Yep, I see that BOINC clearly states that CUDA and OpenCL are present...

The Line stating - "Access Denied" has me wondering... What Anti-Virus Program do you use on that machine??? MANY AVs are "Aggressive" and will block BOINC, and new Units from running properly... IF it is an AV issue, see if you can "Make an Exception" for the BOINC Data Folder so that the AV DOES NOT Scan, (via whatever Live Scan Tool it uses), that Folder. This "should" eliminate any AV conflicts.


TL

[EDIT:]

AND, as a Last Resort... Reinstall BOINC to clear any "Permission Issues" that MIGHT be BOINC related. (I've seen this issue more on MACs than on PCs, where an "app_config.xml" File gets modified, and then the Mac throws up a Message that Permissions have "Changed"... Reinstalling BOINC clears the issue.)


TL
Have TARDIS will travel!!!
ID: 1063 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
TimeLord04
Avatar

Send message
Joined: 30 Aug 18
Posts: 271
Credit: 220,927,978
RAC: 2
Message 1064 - Posted: 20 Sep 2018, 19:51:18 UTC

Just wanted to Post part of my Event Log for comparison:

9/20/2018 12:43:21 PM | | Starting BOINC client version 7.12.1 for windows_x86_64
9/20/2018 12:43:21 PM | | log flags: file_xfer, sched_ops, task
9/20/2018 12:43:21 PM | | Libraries: libcurl/7.47.1 OpenSSL/1.0.2g zlib/1.2.8
9/20/2018 12:43:21 PM | | Data directory: C:\ProgramData\BOINC
9/20/2018 12:43:21 PM | | Running under account "Redacted"
9/20/2018 12:43:21 PM | | CUDA: NVIDIA GPU 0: GeForce GTX 1050 (driver version 388.13, CUDA version 9.1, compute capability 6.1, 2048MB, 1555MB available, 1862 GFLOPS peak)
9/20/2018 12:43:21 PM | | CUDA: NVIDIA GPU 1: GeForce GTX 1050 (driver version 388.13, CUDA version 9.1, compute capability 6.1, 2048MB, 1929MB available, 1862 GFLOPS peak)
9/20/2018 12:43:21 PM | | OpenCL: NVIDIA GPU 0: GeForce GTX 1050 (driver version 388.13, device version OpenCL 1.2 CUDA, 2048MB, 1555MB available, 1862 GFLOPS peak)
9/20/2018 12:43:21 PM | | OpenCL: NVIDIA GPU 1: GeForce GTX 1050 (driver version 388.13, device version OpenCL 1.2 CUDA, 2048MB, 1929MB available, 1862 GFLOPS peak)


My Device 0 is using almost 500 MB for running the Monitor. Device 1 has NO Monitor and thus is using 119 MB reserved for...???...

I notice on your System that about 2 GB is being used/reserved out of your Total 6 GB... (Is this a Driver Related thing??? idk...) My Driver for Win 7 Pro x64 is Version 388.13. Seems like a bigger chunk of your VRAM is being utilized outside of BOINC... (Just an observation.)


TL
Have TARDIS will travel!!!
ID: 1064 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
entigy

Send message
Joined: 1 Jul 10
Posts: 11
Credit: 419,311,588
RAC: 658,571
Message 1068 - Posted: 21 Sep 2018, 18:03:06 UTC
Last modified: 21 Sep 2018, 18:15:41 UTC

Just removed/clean installed the Nvidia software on my GTX 1070 machine - it's made no difference whatsoever:

<core_client_version>7.12.1</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code 9999 (0x270f)</message>
<stderr_txt>
Collatz Conjecture Sieve 1.30 Windows x86_64 for OpenCL
Written by Slicker (Jon Sonntag) of team SETI.USA
Based on the AMD Brook+ kernels by Gipsel of team Planet 3DNow!
Sieve code and OpenCL optimization provided by Sosiris of team BOINC@Taiwan
Collatz Config Settings:
verbose 1 (yes)
kernels/reduction 48
threads 2^9 (512)
lut_size 17 (1048576 bytes)
sieve_size 2^30 (51085096 bytes)
sleep 1
cache_sieve 1 (yes)
reducecpu 0 (no)
Processor Type NVIDIA
Max Dimensions 3
Max Work Items 1024 1024 64
Max Work Groups 1024
Max Kernel Threads 256
Device Vendor NVIDIA Corporation
Name GeForce GTX 1070
Driver Version 411.63
OpenCL Version OpenCL 1.2 CUDA
Device Vendor NVIDIA Corporation
Name GeForce GTX 1070
Driver Version 411.63
OpenCL Version OpenCL 1.2 CUDA
actual threads 512
Error: GPU steps do not match CPU steps. Workunit processing aborted.
18:56:49 (6924): called boinc_finish(9999)

</stderr_txt>
]]>

Seti and Amicable Numbers also use OpenCL and those workunits are completing with no problem....

Edit
Just tweaked my Collatz config file and now I have some successful results.
Curiouser and curiouser...
ID: 1068 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
TimeLord04
Avatar

Send message
Joined: 30 Aug 18
Posts: 271
Credit: 220,927,978
RAC: 2
Message 1069 - Posted: 21 Sep 2018, 19:07:58 UTC
Last modified: 21 Sep 2018, 19:14:47 UTC

["cc_config.xml"] <<----- (Use if you have more than one GPU, EVEN if the second is identical to the first. Put in "C:\ProgramData\BOINC\".)

<cc_config>
<options>
<use_all_gpus>1</use_all_gpus>
<skip_cpu_benchmarks>1</skip_cpu_benchmarks>
<report_results_immediately>1</report_results_immediately>
</options>
</cc_config>


["app_config.xml"] <<----- (Save in the Collatz Project Folder.)

<app_config>
<app>
<name>collatz_sieve</name>
<max_concurrent>2</max_concurrent> <<----- ("2" Allows one Unit per GPU to be run with two GPUs in the System. ALSO, increase this to account for any CPU Threads you wish to crunch with... ie: 1 GPU Unit and 4 CPU Units, set Value to "5".)
<gpu_versions>
<gpu_usage>1.0</gpu_usage>
<cpu_usage>0.5</cpu_usage>
</gpu_versions>
</app>
</app_config>


[".config"] <<----- (Settings below are Optimized for my two EVGA GTX-1050s. Found that "sieve_size=30" was TOO aggressive for my system, reduced to "27".)

verbose=1
kernels_per_reduction=48
threads=9
lut_size=17
sieve_size=27
reduce_cpu=0 <<----- (Can be removed. Slicker and Mikey say this is redundant and useless, now... I'll be taking it out shortly.)
cache_sieve=1
sleep=0


I hope the above helps. Since you only have the one GTX-1070 in the system, you can skip using the "cc_config.xml" File. The "app_config.xml" may prove useful to you, though.

I'm glad your system is now returning good results. Remember to Save your Config Files as "Save As" and select "All Files", (NOT ".txt"), to eliminate bad extensions... You don't want to end up with "xyz.xml.txt", if this was your issue that's why you were crashing Units... (Just a guess at this point...)


TL
Have TARDIS will travel!!!
ID: 1069 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
entigy

Send message
Joined: 1 Jul 10
Posts: 11
Credit: 419,311,588
RAC: 658,571
Message 1070 - Posted: 21 Sep 2018, 20:04:42 UTC
Last modified: 21 Sep 2018, 20:31:44 UTC

So, my GTX 1070 system is successfully completing WUs in under 8 minutes.
I have a 'tweaked' config file but NO app_config file.

My GTX 1060 system, on the other hand, flatly refuses to run any Collatz WU if there's any sort of optimized config file
This is what it contains at the moment (copied from your GTX 1050 settings):

verbose=1
kernels_per_reduction=48
threads=9
lut_size=17
sieve_size=27
cache_sieve=1
sleep=0

If the config file is blank then the WUs complete successfully - but they take about 20 minutes.
They always used to finish after 11 - 12 minutes, but I can't seem to find any optimized settings that it's happy with now....

Edit

Scratch that..
I've copied over the GTX 1070 config and I've managed to complete a unit in around 12 minutes.
Let's hope that's sorted now....
ID: 1070 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
TimeLord04
Avatar

Send message
Joined: 30 Aug 18
Posts: 271
Credit: 220,927,978
RAC: 2
Message 1071 - Posted: 21 Sep 2018, 21:35:20 UTC - in response to Message 1070.  

So, my GTX 1070 system is successfully completing WUs in under 8 minutes.
I have a 'tweaked' config file but NO app_config file.

My GTX 1060 system, on the other hand, flatly refuses to run any Collatz WU if there's any sort of optimized config file
This is what it contains at the moment (copied from your GTX 1050 settings):

verbose=1
kernels_per_reduction=48
threads=9
lut_size=17
sieve_size=27
cache_sieve=1
sleep=0

If the config file is blank then the WUs complete successfully - but they take about 20 minutes.
They always used to finish after 11 - 12 minutes, but I can't seem to find any optimized settings that it's happy with now....

Edit

Scratch that..
I've copied over the GTX 1070 config and I've managed to complete a unit in around 12 minutes.
Let's hope that's sorted now....

Now that it's running right on the 1060, monitor for a week or two. If this remains stable, you may try reducing "lut_size=17" down to "16". (Your cards are better equipped to handle that setting than mine. However; in a couple more weeks, even I might try "16" for the "lut_size=x" setting. The "lut_size=x" has a direct effect on crunching times. The SMALLER the number, the more aggressive the setting, and the faster a card CAN go... The 9x0 and 10x0 Series cards have been reported to do well with "lut_size=16". I'm taking things SLOWER now, with my lowly 1050s after having a bad experience with "sieve_size=30" parameter.

Good luck. :-)


TL
Have TARDIS will travel!!!
ID: 1071 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
TimeLord04
Avatar

Send message
Joined: 30 Aug 18
Posts: 271
Credit: 220,927,978
RAC: 2
Message 1075 - Posted: 23 Sep 2018, 0:10:06 UTC

[EVGA GTX-1050 2GB GDDR5 VRAM Optimized Settings - NEW SWEET SPOT:]

[".config"]

verbose=1
kernels_per_reduction=48
threads=10
lut_size=16
sieve_size=27
cache_sieve=1
sleep=0

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

I've been watching my System over the past 2 weeks, and I've noticed that Crunching Times INCREASED a couple of minutes without any other modifications made by me... The LONGEST Unit crunched by Dev 0 went to approx. 33 Min from my last reported Times on the 8th. Estimated Times also re-adjusted and went UP to over 56 Minutes.

Made the changes Noted in Blue, above, today. Watched a couple of Units crunch... Crunching Times backed down to approx. 30 Min. and 30 Seconds. Estimated Times DROPPED back to just below 55 Min. (54 Min and 52 Seconds.) I think I've hit Peak Performance...


TL


NOTE: Remember, I have TWO EVGA GTX-1050 GPUs.
Have TARDIS will travel!!!
ID: 1075 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
TimeLord04
Avatar

Send message
Joined: 30 Aug 18
Posts: 271
Credit: 220,927,978
RAC: 2
Message 1076 - Posted: 23 Sep 2018, 2:48:57 UTC

[Update:]

Task 9493370

Name collatz_sieve_d859b5a2-47c8-4afa-93e2-8d6c29a3250e_0
Workunit 8386727
Created 22 Sep 2018, 14:28:33 UTC
Sent 22 Sep 2018, 14:41:16 UTC
Report deadline 6 Oct 2018, 14:41:16 UTC
Received 23 Sep 2018, 2:33:07 UTC
Server state Over
Outcome Success
Client state Done
Exit status 0 (0x00000000)
Computer ID 838493
Run time 30 min 36 sec
CPU time 7 sec
Validate state Valid
Credit 30,518.04
Device peak FLOPS 1,867.67 GFLOPS
Application version Collatz Sieve v1.30 (opencl_nvidia_gpu)
windows_x86_64
Peak working set size 97.97 MB
Peak swap size 200.25 MB
Peak disk usage 6.73 MB


Stderr output

<core_client_version>7.12.1</core_client_version>
<![CDATA[
<stderr_txt>
Collatz Conjecture Sieve 1.30 Windows x86_64 for OpenCL
Written by Slicker (Jon Sonntag) of team SETI.USA
Based on the AMD Brook+ kernels by Gipsel of team Planet 3DNow!
Sieve code and OpenCL optimization provided by Sosiris of team BOINC@Taiwan
Collatz Config Settings:
verbose 1 (yes)
kernels/reduction 48
threads 2^10 (1024)
lut_size 16 (524288 bytes)
sieve_size 2^27 (7049172 bytes)
sleep 1
cache_sieve 1 (yes)
reducecpu 0 (no)
Processor Type NVIDIA
Max Dimensions 3
Max Work Items 1024 1024 64
Max Work Groups 1024
Max Kernel Threads 1024
Device Vendor NVIDIA Corporation
Name GeForce GTX 1050
Driver Version 388.13
OpenCL Version OpenCL 1.2 CUDA
Device Vendor NVIDIA Corporation
Name GeForce GTX 1050
Driver Version 388.13
OpenCL Version OpenCL 1.2 CUDA
actual threads 1024
Start 4630059543895795040256
Stop 4630059596672353173504
Best 4630059559472360246767
Highest steps 1967
Total steps 394368128804969
Average steps 569
CPU time 7.76563 seconds
Elapsed time 00:30:33
19:32:38 (4996): called boinc_finish(0)

</stderr_txt>
]]>

-----------------------------------------------------------------------------
19:32:38 - Pacific Time, Reported
-----------------------------------------------------------------------------

Everything seems to be working as it should. Just can't get below 30 Min... I guess that's to be expected with a 2GB GTX-1050 Setup...

I hope your 1060 and 1070 continue to fare better and run faster.


Good Luck,


TL
Have TARDIS will travel!!!
ID: 1076 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
dduggan47

Send message
Joined: 6 Aug 09
Posts: 3
Credit: 42,493,549
RAC: 39,568
Message 1163 - Posted: 15 Oct 2018, 16:14:09 UTC - in response to Message 1076.  

I started having this problem a few days ago on several projects. Today I updated my NVIDA driver (I'm running Win 10 Pro) and all seems well. I haven't finished a Coallatz GPU task since I did it but I've finished a Moo successfully. I also have Collatz and Asteroids tasks that have run well beyond the 30 second failure threshhold.

The driver I've got now is 24.21.13.9836 for the GTX 745.
ID: 1163 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
TimeLord04
Avatar

Send message
Joined: 30 Aug 18
Posts: 271
Credit: 220,927,978
RAC: 2
Message 1167 - Posted: 16 Oct 2018, 13:41:33 UTC - in response to Message 1163.  

I started having this problem a few days ago on several projects. Today I updated my NVIDA driver (I'm running Win 10 Pro) and all seems well. I haven't finished a Coallatz GPU task since I did it but I've finished a Moo successfully. I also have Collatz and Asteroids tasks that have run well beyond the 30 second failure threshhold.

The driver I've got now is 24.21.13.9836 for the GTX 745.

Taking the Bold&Blue Set from your Driver ID, (above), the actual WHQL Driver Ver is: 398.36

I'm still on Driver Ver: 388.13 WHQL. I'm having intermittent Driver Crashing Issues that I CANNOT seem to be able to resolve... I may be Upgrading to a newer Driver to see if this helps.

Also, I'm on Win 7 Pro x64, BUT am on a Gigabyte GA-Z270-HD3 MOBO with an Intel i7 7700K CPU. Windows Update complains that "CPU Too New for Win 7...", yet everything IS working. I've installed the Intel Chipset Driver, the USB 3.x Driver, and "3TB+ Unlock" Driver ALL from Gigabyte, AND ALL Win 7 Pro x64 Supported.


TL
Have TARDIS will travel!!!
ID: 1167 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
TimeLord04
Avatar

Send message
Joined: 30 Aug 18
Posts: 271
Credit: 220,927,978
RAC: 2
Message 1168 - Posted: 16 Oct 2018, 13:54:16 UTC
Last modified: 16 Oct 2018, 13:59:57 UTC

Just encountered this:


Task 11563505

Name collatz_sieve_93262585-fe53-44dd-88e2-be66086dbc20_0
Workunit 10322524
Created 16 Oct 2018, 8:38:47 UTC
Sent 16 Oct 2018, 8:47:00 UTC
Report deadline 30 Oct 2018, 8:47:00 UTC
Received 16 Oct 2018, 13:30:18 UTC
Server state Over
Outcome Computation error
Client state Compute error
Exit status 5 (0x00000005) Unknown error code
Computer ID 838493
Run time 30 min 47 sec
CPU time
Validate state Invalid
Credit 0.00
Device peak FLOPS 1,867.67 GFLOPS
Application version Collatz Sieve v1.30 (opencl_nvidia_gpu)
windows_x86_64
Peak working set size 97.97 MB
Peak swap size 200.20 MB
Peak disk usage 6.73 MB


Stderr output

<core_client_version>7.12.1</core_client_version>
<![CDATA[
<message>
Access is denied.
(0x5) - exit code 5 (0x5)</message>
<stderr_txt>
Collatz Conjecture Sieve 1.30 Windows x86_64 for OpenCL
Written by Slicker (Jon Sonntag) of team SETI.USA
Based on the AMD Brook+ kernels by Gipsel of team Planet 3DNow!
Sieve code and OpenCL optimization provided by Sosiris of team BOINC@Taiwan
Collatz Config Settings:
verbose 1 (yes)
kernels/reduction 48
threads 2^10 (1024)
lut_size 16 (524288 bytes)
sieve_size 2^27 (7049172 bytes)
sleep 1
cache_sieve 1 (yes)
reducecpu 0 (no)
Processor Type NVIDIA
Max Dimensions 3
Max Work Items 1024 1024 64
Max Work Groups 1024
Max Kernel Threads 1024
Device Vendor NVIDIA Corporation
Name GeForce GTX 1050
Driver Version 388.13
OpenCL Version OpenCL 1.2 CUDA
Device Vendor NVIDIA Corporation
Name GeForce GTX 1050
Driver Version 388.13
OpenCL Version OpenCL 1.2 CUDA
actual threads 1024
Error: GPU steps do not match CPU steps. Workunit processing aborted.
06:28:18 (8464): called boinc_finish(5)

</stderr_txt>
]]>

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

NVIDIA Driver Crashed 30 Seconds before being able to view this Report! I'm positive the Driver Crash caused the Computation Error!

I've been having Intermittent Driver Crashes in Win 7 Pro x64 for the past month. This is the FIRST Driver Crash DURING Crunching Hours. (6 PM to 9 AM Pacific.) The Unit "Finished" at "6:28:18 AM - Pacific".

Is this strictly a bad Driver Issue, or a Win 7 Issue??? Any help GREATLY appreciated.


TL
Have TARDIS will travel!!!
ID: 1168 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Gero-T

Send message
Joined: 9 Oct 16
Posts: 9
Credit: 57,231,987,399
RAC: 0
Message 1178 - Posted: 19 Oct 2018, 5:36:26 UTC

i think that are new drivers from Nvidia above 4xx.xx, after installation of 3xx.xx everything works fine for me.
ID: 1178 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Neriyah

Send message
Joined: 25 Mar 18
Posts: 4
Credit: 7,777,312,214
RAC: 3,497,429
Message 1182 - Posted: 19 Oct 2018, 17:35:09 UTC

i think that are new drivers from Nvidia above 4xx.xx, after installation of 3xx.xx everything works fine for me.

This !!!
Excatly the same for me on multiple pcs...
ID: 1182 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Gero-T

Send message
Joined: 9 Oct 16
Posts: 9
Credit: 57,231,987,399
RAC: 0
Message 1225 - Posted: 7 Nov 2018, 14:01:58 UTC - in response to Message 1182.  

i think that are new drivers from Nvidia above 4xx.xx, after installation of 3xx.xx everything works fine for me.

This !!!
Excatly the same for me on multiple pcs...


For me it helped to
remove : sleep= 0 or 1
ID: 1225 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
1 · 2 · Next

Message boards : Number crunching : Computation Errors


©2019 Jon Sonntag; All rights reserved