Hello -

I have a MacBook Pro with Radeon Pro 560 + Intel HD graphics 630. During startup Bionic will not use the 560. I set cc_config.xml to use all GPUs but I see it says "0MB, 0MB available" for the 560. No tasks are event attempted to be assigned. I do see tasks asked for the 630 but none assigned.

Anyone know how to fix this problem? Im running OpenCL 1.2 and Bionic 7.8.2


Twitter: @illmaticwes81
[/img]

Can you post the first thirty or forty lines of your BOINC event log please. It is to be found under the "advanced/tools.
Without seeing this extract from the log it would appear that its a shared memory issue, but the log will contain other information that will help diagnose your problem.

Before posting the log and if BOINC 7.6.34 or 7.8.2, please pass by on Options->Event Log options and check coproc_debug->OK. Then exit and restart BOINC, then give the log. This gives an extra layer of information about the GPU.

---

Thanks for the prompt Jord - I forgot I normally run with those on and most people don't.

Thank for for the prompt reply and help. Here is the info you required:

Fri Sep 15 06:38:13 2017 | | Starting BOINC client version 7.8.2 for x86_64-apple-darwin
Fri Sep 15 06:38:13 2017 | | log flags: file_xfer, sched_ops, task, coproc_debug
Fri Sep 15 06:38:13 2017 | | Libraries: libcurl/7.50.2 OpenSSL/1.1.0 zlib/1.2.8 c-ares/1.11.0
Fri Sep 15 06:38:13 2017 | | Data directory: /Library/Application Support/BOINC Data
Fri Sep 15 06:38:13 2017 | | [coproc] launching child process at /Applications/BOINCManager.app/Contents/Resources/boinc
Fri Sep 15 06:38:13 2017 | | [coproc] relative to directory /Library/Application Support/BOINC Data
Fri Sep 15 06:38:13 2017 | | [coproc] with data directory /Library/Application Support/BOINC Data
Fri Sep 15 06:38:13 2017 | | OpenCL: AMD/ATI GPU 0: AMD Radeon Pro 560 Compute Engine (driver version 1.2 (Jun 15 2017 18:41:53), device version OpenCL 1.2, 4096MB, 4096MB available, 1161 GFLOPS peak)
Fri Sep 15 06:38:13 2017 | | OpenCL: Intel GPU 0: Intel(R) HD Graphics 630 (driver version 1.2(Jun 15 2017 18:45:21), device version OpenCL 1.2, 1536MB, 1536MB available, 211 GFLOPS peak)
Fri Sep 15 06:38:13 2017 | | OpenCL CPU: Intel(R) Core(TM) i7-7820HQ CPU @ 2.90GHz (OpenCL driver vendor: Apple, driver version 1.1, device version OpenCL 1.2)
Fri Sep 15 06:38:13 2017 | | [coproc] NVIDIA: dlopen(/usr/local/cuda/lib/libcuda.dylib, 2): image not found
Fri Sep 15 06:38:13 2017 | | [coproc] Caught SIGSEGV in OpenCL detection
Fri Sep 15 06:38:18 2017 | | Host name: Wesleys-MacBook-Pro.local
Fri Sep 15 06:38:18 2017 | | Processor: 8 GenuineIntel Intel(R) Core(TM) i7-7820HQ CPU @ 2.90GHz [x86 Family 6 Model 158 Stepping 9]
Fri Sep 15 06:38:18 2017 | | 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 pni 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 avx rdrand f16c
Fri Sep 15 06:38:18 2017 | | OS: Mac OS X 10.12.6 (Darwin 16.7.0)
Fri Sep 15 06:38:18 2017 | | Memory: 16.00 GB physical, 270.52 GB virtual
Fri Sep 15 06:38:18 2017 | | Disk: 464.70 GB total, 270.28 GB free
Fri Sep 15 06:38:18 2017 | | Local time is UTC -4 hours
Fri Sep 15 06:38:18 2017 | | Config: GUI RPC allowed from any host
Fri Sep 15 06:38:18 2017 | | Config: use all coprocessors
Fri Sep 15 06:38:18 2017 | World Community Grid | URL http://www.worldcommunitygrid.org/; Computer ID 4064176; resource share 100
Fri Sep 15 06:38:18 2017 | World Community Grid | General prefs: from World Community Grid (last modified 22-Aug-2017 18:08:32)
Fri Sep 15 06:38:18 2017 | World Community Grid | Host location: none
Fri Sep 15 06:38:18 2017 | World Community Grid | General prefs: using your defaults
Fri Sep 15 06:38:18 2017 | | Reading preferences override file
Fri Sep 15 06:38:18 2017 | | Preferences:
Fri Sep 15 06:38:18 2017 | | max memory usage when active: 8192.00 MB
Fri Sep 15 06:38:18 2017 | | max memory usage when idle: 14745.60 MB
Fri Sep 15 06:38:18 2017 | | max disk usage: 270.28 GB
Fri Sep 15 06:38:18 2017 | | (to change preferences, visit a project web site or select Preferences in the Manager)
Fri Sep 15 06:38:23 2017 | World Community Grid | Sending scheduler request: To fetch work.
Fri Sep 15 06:38:23 2017 | World Community Grid | Requesting new tasks for Intel GPU
Fri Sep 15 06:38:30 2017 | World Community Grid | Scheduler request completed: got 0 new tasks


twitter: @illmaticwes81
WCG team: 51163

BTW: As soon as I turned on the debug I see the memory now correct for the Radeon 560 (go figure). I'm also running on battery which is the only change I made here. I still do no see work being requested for the GPU though...

OK, so the "washing machine mechanic" effect still works!

It can take a few cycles for work to start flowing to a newly discovered GPU. Have another look after a couple of tasks have completed and reported and more work has been requested.

And a late thought...
Does the project currently have an application and work for your combination of GPU and OS? (Not all projects have every combination available all the time)

Mind you, I don't think World Community Grid has any GPU apps - either ATI/AMD or Intel: you client has probably gone into indefinite backoff. A single click on the project 'update' button will probably trigger a request, but you won't get any actual work until you attach to a GPU-supporting project.

There is one thing of note which I'll forward to the developer.

Fri Sep 15 06:38:13 2017 | | [coproc] Caught SIGSEGV in OpenCL detection

I'm not sure if this is a benign message or something of more significance. Any SigSegV error is worrisome enough, IMO.

---

Okay, the developer asks that you run CLInfo from a terminal and post the output of it. You can get it from my Dropbox: https://www.dropbox.com/s/uaq75ghl4uj91gj/CLInfo.zip?dl=0

---