New Macbook Pro GPU

Richard
Richard
Joined: 29 Apr 10
Posts: 39
Credit: 39269698
RAC: 0
Topic 197252

I just purchased a new Macbook Pro with an integrated Intel GPU, and an Nvidia GT 750 M.

My BOINC log shows me that it is requesting work for the GPU(s), but not getting any. Does Einstein@home support either of these GPUs?

I have seen various reports of software that I may need to DL, but before I go hunting I was hoping someone here might know what I need to do.

Thanks in advance for your help.

Jord
Joined: 26 Jan 05
Posts: 2952
Credit: 5893653
RAC: 62

New Macbook Pro GPU

Could you please post the BOINC start-up messages? The first 40 or so lines will do.

Yes, your Mac asks for work for the Nvidia GPU (not for the Intel GPU), but it won't get any as there seems to be a problem with your CUDA version.

From the contact log: [version] CUDA compute capability required min: 100, supplied: 0

When you post the messages, we can see what BOINC says about what it found.
Although, I see your version of Operating System is shown as Darwin 13.0.2, which as far as I know is OS X 10.9, or Mavericks. There is no real compatible Nvidia CUDA driver available for Mavericks yet. So if you have Mavericks, you'll have to wait.

In any case, please post your messages, that'll clear up a lot.

Richard
Richard
Joined: 29 Apr 10
Posts: 39
Credit: 39269698
RAC: 0

Thanks. This is OS X

Thanks. This is OS X Mavericks, so I guess I will just have to wait. Here is the start of my log:

Mon Nov 4 10:28:00 2013 | | No config file found - using defaults
Mon Nov 4 10:28:00 2013 | | Starting BOINC client version 7.0.65 for x86_64-apple-darwin
Mon Nov 4 10:28:00 2013 | | log flags: file_xfer, sched_ops, task
Mon Nov 4 10:28:00 2013 | | Libraries: libcurl/7.26.0 OpenSSL/1.0.1c zlib/1.2.5 c-ares/1.9.1
Mon Nov 4 10:28:00 2013 | | Data directory: /Library/Application Support/BOINC Data
Mon Nov 4 10:28:00 2013 | | Processor: 8 GenuineIntel Intel(R) Core(TM) i7-4850HQ CPU @ 2.30GHz [x86 Family 6 Model 70 Stepping 1]
Mon Nov 4 10:28:00 2013 | | Processor features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM PBE SSE3 PCLMULQDQ DTES64 MON DSCPL VMX SMX EST TM2 SSSE3 FMA CX16 TPR PDCM SSE4.1 SSE4.2 x2APIC MOVBE POPCNT AES PCID XSAVE OSXSAVE SEGLIM64 TSCTMR AVX1.0 RDRAND F16C
Mon Nov 4 10:28:00 2013 | | OS: Mac OS X 10.9.0 (Darwin 13.0.2)
Mon Nov 4 10:28:00 2013 | | Memory: 16.00 GB physical, 407.64 GB virtual
Mon Nov 4 10:28:00 2013 | | Disk: 465.12 GB total, 407.40 GB free
Mon Nov 4 10:28:00 2013 | | Local time is UTC -5 hours
Mon Nov 4 10:28:00 2013 | | OpenCL: NVIDIA GPU 0: GeForce GT 750M (driver version 8.20.15 310.40.15f03, device version OpenCL 1.2, 2048MB, 2048MB available, 30 GFLOPS peak)
Mon Nov 4 10:28:00 2013 | | OpenCL: Intel GPU 0: Iris Pro (driver version 1.2(Sep 29 2013 19:45:54), device version OpenCL 1.2, 1024MB, 1024MB available, 2688 GFLOPS peak)
Mon Nov 4 10:28:00 2013 | Einstein@Home | URL http://einstein.phys.uwm.edu/; Computer ID 9284410; resource share 100
Mon Nov 4 10:28:00 2013 | Einstein@Home | General prefs: from Einstein@Home (last modified 18-Feb-2011 09:03:17)
Mon Nov 4 10:28:00 2013 | Einstein@Home | Host location: none
Mon Nov 4 10:28:00 2013 | Einstein@Home | General prefs: using your defaults
Mon Nov 4 10:28:00 2013 | | Preferences:
Mon Nov 4 10:28:00 2013 | | max memory usage when active: 8192.00MB
Mon Nov 4 10:28:00 2013 | | max memory usage when idle: 14745.60MB
Mon Nov 4 10:28:00 2013 | | max disk usage: 100.00GB
Mon Nov 4 10:28:00 2013 | | don't compute while active
Mon Nov 4 10:28:00 2013 | | suspend work if non-BOINC CPU load exceeds 25 %
Mon Nov 4 10:28:00 2013 | | (to change preferences, visit a project web site or select Preferences in the Manager)
Mon Nov 4 10:28:00 2013 | | Not using a proxy
Mon Nov 4 10:28:01 2013 | | Suspending computation - computer is in use
Mon Nov 4 10:32:37 2013 | Einstein@Home | Restarting task h1_0688.50_S6Directed__S6CasAf40a_688.85Hz_478_1 using einstein_S6CasA version 105 in slot 0
Mon Nov 4 10:32:37 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_688.85Hz_421_0 using einstein_S6CasA version 105 in slot 1
Mon Nov 4 10:32:37 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_688.85Hz_422_0 using einstein_S6CasA version 105 in slot 3
Mon Nov 4 10:32:37 2013 | Einstein@Home | Restarting task h1_0688.50_S6Directed__S6CasAf40a_688.85Hz_477_0 using einstein_S6CasA version 105 in slot 4
Mon Nov 4 10:32:37 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_688.85Hz_420_0 using einstein_S6CasA version 105 in slot 5
Mon Nov 4 10:32:37 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_688.85Hz_419_0 using einstein_S6CasA version 105 in slot 7
Mon Nov 4 10:32:37 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_688.85Hz_418_0 using einstein_S6CasA version 105 in slot 6
Mon Nov 4 10:32:37 2013 | Einstein@Home | Restarting task LATeah0045U_1136.0_680460_0.0_1 using hsgamma_FGRP2 version 112 in slot 2
Mon Nov 4 10:33:02 2013 | Einstein@Home | Sending scheduler request: To fetch work.
Mon Nov 4 10:33:02 2013 | Einstein@Home | Requesting new tasks for NVIDIA and intel_gpu
Mon Nov 4 10:33:04 2013 | Einstein@Home | Scheduler request completed: got 0 new tasks
Mon Nov 4 10:33:04 2013 | Einstein@Home | No work sent
Mon Nov 4 10:33:04 2013 | Einstein@Home | see scheduler log messages on http://einstein.phys.uwm.edu//host_sched_logs/9284/9284410
Mon Nov 4 10:45:07 2013 | Einstein@Home | Sending scheduler request: To fetch work.
Mon Nov 4 10:45:07 2013 | Einstein@Home | Requesting new tasks for intel_gpu
Mon Nov 4 10:45:10 2013 | Einstein@Home | Scheduler request completed: got 0 new tasks
Mon Nov 4 10:45:10 2013 | Einstein@Home | No work sent
Mon Nov 4 10:45:10 2013 | Einstein@Home | see scheduler log messages on http://einstein.phys.uwm.edu//host_sched_logs/9284/9284410
Mon Nov 4 10:47:10 2013 | Einstein@Home | Sending scheduler request: To fetch work.
Mon Nov 4 10:47:10 2013 | Einstein@Home | Requesting new tasks for NVIDIA
Mon Nov 4 10:47:13 2013 | Einstein@Home | Scheduler request completed: got 0 new tasks
Mon Nov 4 10:47:13 2013 | Einstein@Home | No work sent
Mon Nov 4 10:47:13 2013 | Einstein@Home | see scheduler log messages on http://einstein.phys.uwm.edu//host_sched_logs/9284/9284410
Mon Nov 4 10:55:19 2013 | | Suspending computation - computer is in use
Mon Nov 4 10:59:55 2013 | | Resuming computation
Mon Nov 4 11:03:36 2013 | Einstein@Home | Sending scheduler request: To fetch work.
Mon Nov 4 11:03:36 2013 | Einstein@Home | Requesting new tasks for intel_gpu
Mon Nov 4 11:03:38 2013 | Einstein@Home | Scheduler request completed: got 0 new tasks

Jord
Joined: 26 Jan 05
Posts: 2952
Credit: 5893653
RAC: 62

RE: Mon Nov 4 10:28:00

Quote:
Mon Nov 4 10:28:00 2013 | | OpenCL: NVIDIA GPU 0: GeForce GT 750M (driver version 8.20.15 310.40.15f03, device version OpenCL 1.2, 2048MB, 2048MB available, 30 GFLOPS peak)
Mon Nov 4 10:28:00 2013 | | OpenCL: Intel GPU 0: Iris Pro (driver version 1.2(Sep 29 2013 19:45:54), device version OpenCL 1.2, 1024MB, 1024MB available, 2688 GFLOPS peak)


Yup, as you can see, only the OpenCL part of the Nvidia GPU is recognized. Kind of a weird driver version, but that aside.

This project does have an Intel GPU OpenCL application, but I'm afraid it's for Windows only.

Oliver Behnke
Oliver Behnke
Moderator
Administrator
Joined: 4 Sep 07
Posts: 984
Credit: 25171438
RAC: 28

@Richard: Can you please

@Richard:

Can you please give me a link to the CUDA driver you installed, if any? As Jord already noted, the version number seems to indicate that you use a driver that's newer than the latest official release by NVIDIA. Maybe it's the one provided by Apple/OSX itself?

Background: we don't support OpenCL on NVIDIA GPUs, so you'd have to install the latest CUDA driver which would be 5.5.28 at the time of writing. We also don't yet support OpenCL on Intel GPUs on OSX, but we plan to so shortly. Your client will fetch work for the Iris GPU automatically as soon as we deployed the application.

Thanks,
Oliver

Einstein@Home Project

Claggy
Claggy
Joined: 29 Dec 06
Posts: 560
Credit: 2699403
RAC: 0

From the Boinc Dev

From the Boinc Dev Forums:

GPU problem on Mac OS X with 7.0.65

Quote:

BOINC currently does not support CUDA on dual-GPU MacBook Pros. This is because the way the CUDA drivers work, testing for the GPU's presence at BOINC startup causes the NVIDIA GPU to remain selected even when the computer is running on batteries, rapidly running down the battery. This happens even if BOINC suspends operation when running on battery power.

Testing for OpenCL does not have this problem, so we still support GPU computing using OpenCL on these computers. You will notice that your Event Log says:

Fri Apr 26 11:18:19 2013 | | BOINC does not support CUDA on this computer.
Fri Apr 26 11:18:19 2013 | | OpenCL: NVIDIA GPU 0: GeForce GT 650M (driver version CLH 1.0, device version OpenCL 1.1, 1024MB, 1024MB available, 29 GFLOPS peak)

I filed a bug report with NVIDIA on October 23, 2012. requesting a way to release the NVIDIA GPU so that it can switch back to the power-efficient integrated Intel GPU. Despite repeatedly contacting them, I am still waiting for them to do something about this. Because we have had so many complaints from users about BOINC running down their batteries, we have no choice but to disable CUDA support on MacBook Pros until NVIDIA provides a solution.

Cheers,
--Charlie

Claggy

Oliver Behnke
Oliver Behnke
Moderator
Administrator
Joined: 4 Sep 07
Posts: 984
Credit: 25171438
RAC: 28

Fair enough, but we aren't at

Fair enough, but we aren't at the point yet. Right now the GPU isn't even detected as a CUDA device, hence the missing message "BOINC does not support CUDA on this computer"...

FYI: I verified this on my own dual NVIDIA GPU MacBook Pro where CUDA 5.5.28 also doesn't work (expected because GPUs aren't > sm_1x architecture).

Oliver

Einstein@Home Project

Jord
Joined: 26 Jan 05
Posts: 2952
Credit: 5893653
RAC: 62

RE: Right now the GPU isn't

Quote:
Right now the GPU isn't even detected as a CUDA device, hence the missing message "BOINC does not support CUDA on this computer"...


The missing message was a apparently temporary thing for the dual GPU Macbook Pro only. The helpful messages are now all hidden under the flag. {shrug}

Oliver Behnke
Oliver Behnke
Moderator
Administrator
Joined: 4 Sep 07
Posts: 984
Credit: 25171438
RAC: 28

@Richard: you need to install

@Richard: you need to install the latest CUDA driver from NVIDIA. The driver shipped by Apple isn't CUDA-enabled which is why BOINC can't use CUDA on your device.

Please install the driver, reboot just to be sure and start BOINC again. You should be good to go then.

HTH,
Oliver

Einstein@Home Project

Richard
Richard
Joined: 29 Apr 10
Posts: 39
Credit: 39269698
RAC: 0

Thanks Oliver. I

Thanks Oliver.

I downloaded the driver you suggested. I still don't get any tasks. It looks like it doesn't see the CUDA driver. But I did install and reboot as you suggested. The scheduler log still shows:

2013-11-18 13:21:46.1186 [PID=21233] [version] CUDA compute capability required min: 100, supplied: 0

Mon Nov 18 08:21:38 2013 | | No config file found - using defaults
Mon Nov 18 08:21:38 2013 | | Starting BOINC client version 7.0.65 for x86_64-apple-darwin
Mon Nov 18 08:21:38 2013 | | log flags: file_xfer, sched_ops, task
Mon Nov 18 08:21:38 2013 | | Libraries: libcurl/7.26.0 OpenSSL/1.0.1c zlib/1.2.5 c-ares/1.9.1
Mon Nov 18 08:21:38 2013 | | Data directory: /Library/Application Support/BOINC Data
Mon Nov 18 08:21:38 2013 | | Processor: 8 GenuineIntel Intel(R) Core(TM) i7-4850HQ CPU @ 2.30GHz [x86 Family 6 Model 70 Stepping 1]
Mon Nov 18 08:21:38 2013 | | Processor features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM PBE SSE3 PCLMULQDQ DTES64 MON DSCPL VMX SMX EST TM2 SSSE3 FMA CX16 TPR PDCM SSE4.1 SSE4.2 x2APIC MOVBE POPCNT AES PCID XSAVE OSXSAVE SEGLIM64 TSCTMR AVX1.0 RDRAND F16C
Mon Nov 18 08:21:38 2013 | | OS: Mac OS X 10.9.0 (Darwin 13.0.2)
Mon Nov 18 08:21:38 2013 | | Memory: 16.00 GB physical, 378.27 GB virtual
Mon Nov 18 08:21:38 2013 | | Disk: 465.12 GB total, 378.02 GB free
Mon Nov 18 08:21:38 2013 | | Local time is UTC -5 hours
Mon Nov 18 08:21:38 2013 | | BOINC does not support CUDA on this computer.
Mon Nov 18 08:21:38 2013 | | OpenCL: NVIDIA GPU 0: GeForce GT 750M (driver version 8.20.15 310.40.15f03, device version OpenCL 1.2, 2048MB, 2048MB available, 30 GFLOPS peak)
Mon Nov 18 08:21:38 2013 | | OpenCL: Intel GPU 0: Iris Pro (driver version 1.2(Sep 29 2013 19:45:54), device version OpenCL 1.2, 1024MB, 1024MB available, 2688 GFLOPS peak)
Mon Nov 18 08:21:38 2013 | Einstein@Home | URL http://einstein.phys.uwm.edu/; Computer ID 9284410; resource share 100
Mon Nov 18 08:21:38 2013 | Einstein@Home | General prefs: from Einstein@Home (last modified 18-Feb-2011 09:03:17)
Mon Nov 18 08:21:38 2013 | Einstein@Home | Host location: none
Mon Nov 18 08:21:38 2013 | Einstein@Home | General prefs: using your defaults
Mon Nov 18 08:21:38 2013 | | Reading preferences override file
Mon Nov 18 08:21:38 2013 | | Preferences:
Mon Nov 18 08:21:38 2013 | | max memory usage when active: 8192.00MB
Mon Nov 18 08:21:38 2013 | | max memory usage when idle: 14745.60MB
Mon Nov 18 08:21:38 2013 | | max disk usage: 100.00GB
Mon Nov 18 08:21:38 2013 | | don't compute while active
Mon Nov 18 08:21:38 2013 | | suspend work if non-BOINC CPU load exceeds 25 %
Mon Nov 18 08:21:38 2013 | | (to change preferences, visit a project web site or select Preferences in the Manager)
Mon Nov 18 08:21:38 2013 | | Not using a proxy
Mon Nov 18 08:21:38 2013 | Einstein@Home | Restarting task h1_0688.50_S6Directed__S6CasAf40a_688.9Hz_412_0 using einstein_S6CasA version 105 in slot 7
Mon Nov 18 08:21:38 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_689.05Hz_270_0 using einstein_S6CasA version 105 in slot 1
Mon Nov 18 08:21:38 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_689.05Hz_269_1 using einstein_S6CasA version 105 in slot 4
Mon Nov 18 08:21:38 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_689.05Hz_268_0 using einstein_S6CasA version 105 in slot 3
Mon Nov 18 08:21:38 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_688.85Hz_413_2 using einstein_S6CasA version 105 in slot 2
Mon Nov 18 08:21:38 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_689.05Hz_267_0 using einstein_S6CasA version 105 in slot 5
Mon Nov 18 08:21:38 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_689.05Hz_266_0 using einstein_S6CasA version 105 in slot 6
Mon Nov 18 08:21:38 2013 | Einstein@Home | Restarting task h1_0688.45_S6Directed__S6CasAf40a_689.05Hz_265_0 using einstein_S6CasA version 105 in slot 0
Mon Nov 18 08:21:38 2013 | Einstein@Home | Sending scheduler request: To fetch work.
Mon Nov 18 08:21:38 2013 | Einstein@Home | Reporting 2 completed tasks
Mon Nov 18 08:21:38 2013 | Einstein@Home | Requesting new tasks for NVIDIA and intel_gpu
Mon Nov 18 08:21:46 2013 | Einstein@Home | Scheduler request completed: got 0 new tasks
Mon Nov 18 08:21:46 2013 | Einstein@Home | No work sent
Mon Nov 18 08:21:46 2013 | Einstein@Home | see scheduler log messages on http://einstein.phys.uwm.edu//host_sched_logs/9284/9284410
Mon Nov 18 08:21:46 2013 | Einstein@Home | Binary Radio Pulsar Search (Arecibo) is not available for your type of computer.

Oliver Behnke
Oliver Behnke
Moderator
Administrator
Joined: 4 Sep 07
Posts: 984
Credit: 25171438
RAC: 28

Hm, that's strange. I'll have

Hm, that's strange. I'll have to take a closer look at our test hardware. If I find a box with a suitable (> sm_1x) GPU in it, I'm going to try to reproduce that issue locally.

Meanwhile, if you're willing to test even further, you could try and install CUDA itself and compile the samples provided. They include a deviceQuery[Drv] tool which would finally determine whether this is an NVIDIA/OSX issue or a problem in BOINC.

Cheers,
Oliver

Einstein@Home Project

Comment viewing options

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