Consistent "clBuildProgram() failed with error (-11)" after upgrading iMac to 10.15.7 (Catalina)

Questions and Answers : Macintosh : Consistent "clBuildProgram() failed with error (-11)" after upgrading iMac to 10.15.7 (Catalina)
Message board moderation

To post messages, you must log in.

AuthorMessage
Jon Fox

Send message
Joined: 6 Sep 09
Posts: 9
Credit: 554,714,573
RAC: 0
Message 3361 - Posted: 14 Jun 2021, 16:28:09 UTC
Last modified: 14 Jun 2021, 16:30:50 UTC

The following error is occurring for all GPU tasks. If you would also like to see the BOINC log data, it is available. Thanks in advance for any assistance you can provide.

Jon Fox

    <core_client_version>7.16.11</core_client_version>
    <![CDATA[
    <message>
    process exited with code 245 (0xf5, -11)</message>
    <stderr_txt>
    Collatz Conjecture Sieve 1.40 OS X 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 64
    threads 2^8 (256)
    lut_size 20 (8388608 bytes)
    sieve_size 2^30 (51085096 bytes)
    sleep 1
    cache_sieve 1 (yes)
    reducecpu 0 (no)
    Processor Type NVIDIA
    BUILD LOG

    clBuildProgram() failed with error (-11)

    Error -11. Processing Aborted.
    18:18:24 (1222): called boinc_finish

    </stderr_txt>
    ]]>

ID: 3361 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile mikey
Avatar

Send message
Joined: 11 Aug 09
Posts: 963
Credit: 24,557,133,931
RAC: 83,763
Message 3362 - Posted: 15 Jun 2021, 10:34:48 UTC - in response to Message 3361.  

The following error is occurring for all GPU tasks. If you would also like to see the BOINC log data, it is available. Thanks in advance for any assistance you can provide.

Jon Fox

    <core_client_version>7.16.11</core_client_version>
    <![CDATA[
    <message>
    process exited with code 245 (0xf5, -11)</message>
    <stderr_txt>
    Collatz Conjecture Sieve 1.40 OS X 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 64
    threads 2^8 (256)
    lut_size 20 (8388608 bytes)
    sieve_size 2^30 (51085096 bytes)
    sleep 1
    cache_sieve 1 (yes)
    reducecpu 0 (no)
    Processor Type NVIDIA
    BUILD LOG

    clBuildProgram() failed with error (-11)

    Error -11. Processing Aborted.
    18:18:24 (1222): called boinc_finish

    </stderr_txt>
    ]]>



I would first remove the optimization codes and see if it works okay then, it could just be your gpu doesn't like the codes or they need to be scaled back a bit but until you can run a unit successfully you won't know that.
ID: 3362 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Jon Fox

Send message
Joined: 6 Sep 09
Posts: 9
Credit: 554,714,573
RAC: 0
Message 3363 - Posted: 18 Jun 2021, 19:00:25 UTC - in response to Message 3362.  

Thank you for replying. I've been running with these optimization parms for several years but you're right, I'll set them aside and run a totally default setting environment to confirm that they are not the culprit.

Thanks again.
Jon
ID: 3363 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Jon Fox

Send message
Joined: 6 Sep 09
Posts: 9
Credit: 554,714,573
RAC: 0
Message 3364 - Posted: 18 Jun 2021, 22:51:11 UTC - in response to Message 3363.  

Results from one of several runs with no optimization parms:


    <core_client_version>7.16.14</core_client_version>
    <![CDATA[
    <message>
    process exited with code 245 (0xf5, -11)</message>
    <stderr_txt>
    Collatz Conjecture Sieve 1.40 OS X 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
    Read config error: Cannot locate ../../projects/boinc.thesonntags.com_collatz/collatz_sieve_1.40_x86_64-apple-darwin__nvidia_opencl.config
    Processor Type NVIDIA
    BUILD LOG

    clBuildProgram() failed with error (-11)

    Error -11. Processing Aborted.
    15:08:44 (8844): called boinc_finish

    </stderr_txt>
    ]]>



Thanks again in advance for any help with this issue.
Jon Fox

ID: 3364 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile mikey
Avatar

Send message
Joined: 11 Aug 09
Posts: 963
Credit: 24,557,133,931
RAC: 83,763
Message 3365 - Posted: 19 Jun 2021, 3:02:18 UTC - in response to Message 3364.  
Last modified: 19 Jun 2021, 3:02:46 UTC

Results from one of several runs with no optimization parms:


    <core_client_version>7.16.14</core_client_version>
    <![CDATA[
    <message>
    process exited with code 245 (0xf5, -11)</message>
    <stderr_txt>
    Collatz Conjecture Sieve 1.40 OS X 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
    Read config error: Cannot locate ../../projects/boinc.thesonntags.com_collatz/collatz_sieve_1.40_x86_64-apple-darwin__nvidia_opencl.config
    Processor Type NVIDIA
    BUILD LOG

    clBuildProgram() failed with error (-11)

    Error -11. Processing Aborted.
    15:08:44 (8844): called boinc_finish

    </stderr_txt>
    ]]>



Thanks again in advance for any help with this issue.
Jon Fox



When I did a search of Boinc Error codes I found this:

Exit code -11
Exit code -11 is a floating point rounding operation error. Part of the outcome of the calculation was found to be faulty, so calculations were halted. Possible causes: heat, corrupt driver, broken hardware.
https://boinc.mundayweb.com/wiki/index.php?title=Exit_code_-11
ID: 3365 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Jon Fox

Send message
Joined: 6 Sep 09
Posts: 9
Credit: 554,714,573
RAC: 0
Message 3366 - Posted: 19 Jun 2021, 11:44:26 UTC - in response to Message 3365.  

Yes, I came across the same item in one of my searches. With the timing "coincidence" of the error showing up immediately after my upgrade I posted here to see if anyone else in the community had seen a similar error. I suspect that 10.15.7 corrupted my NVIDIA driver.

I'll leave the post up in case someone does run across it and has something to add.


Thanks again.
Jon Fox
ID: 3366 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Tigers_Dave
Avatar

Send message
Joined: 23 Sep 12
Posts: 201
Credit: 90,408,088,852
RAC: 97,504,398
Message 3405 - Posted: 1 Sep 2021, 20:31:00 UTC - in response to Message 3366.  

Jon:

Thanks so much for posting your query. You are running 10.15.7 on a machine that has an NVIDIA GPU? Could you please provide us with more details on your hardware? I am not aware of a Mac that supports both 10.15.7 and an NVIDIA GPU under an Apple-sanctioned configuration.

Regards,

Dave
ID: 3405 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Slicker
Project administrator

Send message
Joined: 11 Jun 09
Posts: 79
Credit: 943,644,517
RAC: 0
Message 3407 - Posted: 2 Sep 2021, 0:35:18 UTC - in response to Message 3365.  

Error code 11 is actually the error from Apple's OpenCL compiler, not BOINC. It occurs while OpenCL is compiling the OpenCL kernels. This is done before any calculations are done. Also, Collatz doesn't use floating point arithmetic. It only uses bit shifts and addition. Apple has had "glitches" with their OpenCL compiler in the past. I even submitted code to both Apple and nVidia years ago and the same error keeps popping up. If I disable optimization, it works fine but takes way longer to complete a work unit. Apple has also had issues where warnings were considered errors and some of the OpenCL optimizations are dependent on the version of OpenCL that is bring run. If the compiler identifies itself as OpenCL v2 vs OpenCL v1 then the kernel uses the v2 optimizations. If the compiler is only "kind of" v2 (like Microsoft's browsers were "kind of" HTML 4 or HTML 5 compliant) and an optimization isn't fully supported, it will fail to compile. On Windows and Linux, the display drive contains the OpenCL compiler so upgrading the nVidia driver installs the latest OpenCL compiler as well. But, Apple includes OpenCL as part of the operating system, so you really can't change it. You are stuck until a new version of OS X is released.
ID: 3407 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Tigers_Dave
Avatar

Send message
Joined: 23 Sep 12
Posts: 201
Credit: 90,408,088,852
RAC: 97,504,398
Message 3408 - Posted: 8 Sep 2021, 16:49:15 UTC - in response to Message 3366.  

Hi Jon:

Sorry that you are having problems with 10.15.7 and Collatz. I agree with you that the 10.15.7 update may be giving you problems. I also agree with Slicker that waiting for another update may be the only solution. Have you tried installing the most recent update? A Catalina update was released in late July. https://support.apple.com/en-us/HT201222

Regards,

Dave
ID: 3408 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Jon Fox

Send message
Joined: 6 Sep 09
Posts: 9
Credit: 554,714,573
RAC: 0
Message 3518 - Posted: 17 Oct 2021, 23:26:55 UTC - in response to Message 3408.  

Good Evening Dave,

Yes, I'm current on the maintenance side. In fact, the original error that I was seeing in July eventually disappeared (with no changes on my side) in the September timeframe and reappeared this week after I applied the latest Catalina update (still listed as 10.15.7). The difference this time is that all GPU-related tasks have failed (e.g., Milky Way GPU tasks), not just Collatz.

I guess I'll just wait it out and hope for a "magic healing" in the vein of the July update <lol>. I'll also be in the market for a new device in the not to distant future. Maybe something from tomorrow's Apple event will entice me to move that change-out even closer.


Thanks,
Jon
ID: 3518 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Tigers_Dave
Avatar

Send message
Joined: 23 Sep 12
Posts: 201
Credit: 90,408,088,852
RAC: 97,504,398
Message 3560 - Posted: 21 Dec 2021, 14:02:02 UTC - in response to Message 3518.  

Hi Jon:

So, did you decide to buy a new Apple device? As for me, I am delighted that Apple has come out with new 14" and 16" laptops. Nonetheless, I think my current MacBook Pro (16" i9) has at least another year as my go-to mobile machine. I am a bit disappointed that Apple has yet to offer an upgraded version of the M1 MacMini - the most expensive MacMini remains the 6-core i5. More importantly, I am disappointed that Apple has yet to figure out how to connect eGPUs to M1 Macs. I do a lot of GPU-based scientific computing and the desktop M1 Macs don't have the GPU computing power that I need. And, of course, the Xeon-based MacPros are way too expensive for my budget.

In any event, happy holidays and happy crunching!

Dave
ID: 3560 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile mikey
Avatar

Send message
Joined: 11 Aug 09
Posts: 963
Credit: 24,557,133,931
RAC: 83,763
Message 3562 - Posted: 22 Dec 2021, 11:20:44 UTC - in response to Message 3362.  

[quote]The following error is occurring for all GPU tasks. If you would also like to see the BOINC log data, it is available. Thanks in advance for any assistance you can provide.

Jon Fox


When I look at the computers I have here at Collatz I see this under the graphics card:
NVIDIA NVIDIA GeForce GTX 1050 Ti (4033MB) driver: 470.86 OpenCL: 3.0

BUT when I look at yours I see this:
NVIDIA GeForce GT 750M (1024MB) OpenCL: 1.2

So to me that means you have not loaded the Nvidia drivers from Nvidia itself to make your card do it's thing, the drivers are required for both Boinc and gaming but since that's card only has 1gb on onboard ram on it it's not a good gaming gpu. The OS only loads drivers that make the display come on the screen not the higher level drivers needed for Boinc or gaming. I googled where to get drivers for a Mac and came up with this but didn't actually go there:
[url]What drivers do I need to run NVIDIA CUDA on a Mac?

"In order to run Mac OS X Applications that leverage the CUDA architecture of certain NVIDIA graphics cards, users will need to download and install the 7.0.64 driver for Mac located here. Contains performance improvements and bug fixes for a wide range of applications."

NVIDIA DRIVERS Quadro & GeForce Mac OS X Driver Release ...
www.nvidia.com/download/driverResults.aspx/88655/en-us[/url]
ID: 3562 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Jon Fox

Send message
Joined: 6 Sep 09
Posts: 9
Credit: 554,714,573
RAC: 0
Message 3564 - Posted: 27 Dec 2021, 8:59:00 UTC - in response to Message 3562.  

Thanks for researching and sharing what you found on this issue. I did attempt to download the latest CUDA drivers but the installation failed due to kernel-level "hardware extension" protection. This lead me to the "Security and Privacy" panel in System Preferences which is now throwing error when I try it open it. Net-net, I'm going to have to do some deep diving on the System Preferences error repair and it looks like it's leading me to restoring OS X and then reapplying the latest updates. Now I have to make decision if I want to invest the time in this 2013 iMac or pull the trigger on replacing the rig (which I am considering now that the M1 iMacs are available).

Again, thank you very much for the pointers on CUDA support. I'll post back here when I've got a working rig again (new or repaired).

Jon
ID: 3564 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Jon Fox

Send message
Joined: 6 Sep 09
Posts: 9
Credit: 554,714,573
RAC: 0
Message 3565 - Posted: 27 Dec 2021, 9:08:10 UTC - in response to Message 3560.  

Hi Dave,

I am still in a "wait state" on the replacement as I am interested in a higher end M1 MacMini which rumor has it will be announced in early-to-mid 2022. I'm also reading about higher end iMac (27/32") which would also be on my short list. The problem with that approach is that if it is mid-year and there's any delay in delivery (all too possible with the supply chain issues we've seen in 2021), my hand may be forced by the untimely demise of my current rig. I'll come back to this thread when I wind my way out of this mess (one way or the other).

Hope you having a safe and happy holiday season.
Jon
ID: 3565 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Tigers_Dave
Avatar

Send message
Joined: 23 Sep 12
Posts: 201
Credit: 90,408,088,852
RAC: 97,504,398
Message 3597 - Posted: 13 Jan 2022, 23:47:53 UTC - in response to Message 3565.  

Hi Jon,

I understand the frustration of not having a suitable replacement available for a dead/dying rig. Recently I had to purchase a used MacMini 8,1 (2018 model) on eBay to replace the one that is dying in my home office. I realize this is a first-world complaint, but I continue to be bummed that Apple won't offer a relatively low-cost yet expandable computer for professionals and hard-core hobbyists.

I hope you had a happy holiday season, too. Best wishes for a wonderful 2022.

Dave
ID: 3597 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile mikey
Avatar

Send message
Joined: 11 Aug 09
Posts: 963
Credit: 24,557,133,931
RAC: 83,763
Message 3601 - Posted: 14 Jan 2022, 10:24:19 UTC - in response to Message 3565.  

Hi Dave,

I am still in a "wait state" on the replacement as I am interested in a higher end M1 MacMini which rumor has it will be announced in early-to-mid 2022. I'm also reading about higher end iMac (27/32") which would also be on my short list. The problem with that approach is that if it is mid-year and there's any delay in delivery (all too possible with the supply chain issues we've seen in 2021), my hand may be forced by the untimely demise of my current rig. I'll come back to this thread when I wind my way out of this mess (one way or the other).

Hope you having a safe and happy holiday season.
Jon


The key to buying technology is....as soon as you walk out the door with your brand new purchase a newer version hits the market!! They like it that way but for us consumers it royally sucks.
mikey
ID: 3601 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Tigers_Dave
Avatar

Send message
Joined: 23 Sep 12
Posts: 201
Credit: 90,408,088,852
RAC: 97,504,398
Message 3618 - Posted: 24 Jan 2022, 15:07:04 UTC - in response to Message 3601.  

Hi Dave,

I am still in a "wait state" on the replacement as I am interested in a higher end M1 MacMini which rumor has it will be announced in early-to-mid 2022. I'm also reading about higher end iMac (27/32") which would also be on my short list. The problem with that approach is that if it is mid-year and there's any delay in delivery (all too possible with the supply chain issues we've seen in 2021), my hand may be forced by the untimely demise of my current rig. I'll come back to this thread when I wind my way out of this mess (one way or the other).

Hope you having a safe and happy holiday season.
Jon


The key to buying technology is....as soon as you walk out the door with your brand new purchase a newer version hits the market!! They like it that way but for us consumers it royally sucks.
mikey


So true, Mikey. It's even worse when the new model is markedly better than the predecessor that you just bought. My biggest recent (relatively speaking) regret is a 16" i9 MacBookPro, which has been supplanted (speed, heat management, battery life, ...) by the Apple Silicon laptops.

Dave
ID: 3618 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile mikey
Avatar

Send message
Joined: 11 Aug 09
Posts: 963
Credit: 24,557,133,931
RAC: 83,763
Message 3619 - Posted: 25 Jan 2022, 11:55:40 UTC - in response to Message 3618.  

Hi Dave,

I am still in a "wait state" on the replacement as I am interested in a higher end M1 MacMini which rumor has it will be announced in early-to-mid 2022. I'm also reading about higher end iMac (27/32") which would also be on my short list. The problem with that approach is that if it is mid-year and there's any delay in delivery (all too possible with the supply chain issues we've seen in 2021), my hand may be forced by the untimely demise of my current rig. I'll come back to this thread when I wind my way out of this mess (one way or the other).

Hope you having a safe and happy holiday season.
Jon


The key to buying technology is....as soon as you walk out the door with your brand new purchase a newer version hits the market!! They like it that way but for us consumers it royally sucks.
mikey


So true, Mikey. It's even worse when the new model is markedly better than the predecessor that you just bought. My biggest recent (relatively speaking) regret is a 16" i9 MacBookPro, which has been supplanted (speed, heat management, battery life, ...) by the Apple Silicon laptops.

Dave


There are places you could sell it but you will take a loss on it, I usually just keep stuff until it just can't do what I want it to do anymore and that includes crunching
ID: 3619 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Questions and Answers : Macintosh : Consistent "clBuildProgram() failed with error (-11)" after upgrading iMac to 10.15.7 (Catalina)


©2022 Jon Sonntag; All rights reserved