MAC OS X El Capitan and NVIDIA Web Driver.

Message boards : Number crunching : MAC OS X El Capitan and NVIDIA Web Driver.
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3

AuthorMessage
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5517
Credit: 528,817,460
RAC: 242
United States
Message 1771028 - Posted: 11 Mar 2016, 20:22:05 UTC - in response to Message 1771026.  
Last modified: 11 Mar 2016, 20:31:34 UTC

Yeah, went back and looked at the app_info.

There is no section for the intel opencl in there. Guess I need to scrounge around and find one for the mac.

Edit..

TBar you got a intel Opencl for the app_info I can use?
ID: 1771028 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1771035 - Posted: 11 Mar 2016, 20:41:22 UTC - in response to Message 1771028.  

Yeah, went back and looked at the app_info.

There is no section for the intel opencl in there. Guess I need to scrounge around and find one for the mac.

Edit..

TBar you got a intel Opencl for the app_info I can use?

Not currently. I removed the one at C.A. due to it's poor record of creating a large number of Inconclusive results. The tasks would verify eventually, however the App wasn't any better than the one available on Main.
Basically, it didn't meet my standards.

I'm considering attempting to build a new iGPU App in Yosemite using Xcode 6.1.1. As far as I know both the App on Main & the previous App were compiled in Mountain Lion. It's possible it may work better compiled in Yosemite. Of course, since I don't have an iGPU, the new App would be completely untested.
Hmmm, it may work in standalone on a nVidia card so I could at least tell if it works at all.
ID: 1771035 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14650
Credit: 200,643,578
RAC: 874
United Kingdom
Message 1771037 - Posted: 11 Mar 2016, 20:49:36 UTC

You don't need a new app. Just writing a definition section in app_info.xml and running the one you had before would do the trick. The easiest way would be to ask someone who's still running a Mac intel_gpu as stock, and get them to post both the app_version entries you can adapt into app_info, and the download urls.

I did much the same for CPU apps in my thread

SETI@Home Version 8 - app_info.xml files ... (all OS)

to ease people through the v8 transition before third-party apps were ready.
ID: 1771037 · Report as offensive
Profile TimeLord04
Volunteer tester
Avatar

Send message
Joined: 9 Mar 06
Posts: 21140
Credit: 33,933,039
RAC: 23
United States
Message 1771134 - Posted: 12 Mar 2016, 5:48:54 UTC
Last modified: 12 Mar 2016, 5:51:57 UTC

Just found an interesting Unit that has completed and been given credit.

750 TI on PC CUDA 50 vs 750 TI MAC OpenCL.

The CUDA 50 WASN'T even on Lunatics... It processed SOOOO MUCH FASTER than MAC OpenCL... I find this fascinating.

Task Info on MAC OS X El Capitan.


TL

[EDIT:]

MAC OS X:

Name 15jl10ac.10715.82997.11.38.115_1
Workunit 2089522873
Created 10 Mar 2016, 12:16:40 UTC
Sent 10 Mar 2016, 18:09:52 UTC
Report deadline 3 May 2016, 1:58:13 UTC
Received 10 Mar 2016, 21:37:46 UTC
Server state Over
Outcome Success
Client state Done
Exit status 0 (0x0)
Computer ID 7952666
Run time 23 min 11 sec
CPU time 6 min 54 sec
Validate state Valid
Credit 75.43
Device peak FLOPS 805.39 GFLOPS
Application version SETI@home v8 v8.00 (opencl_nvidia_mac)
x86_64-apple-darwin
Peak working set size 75.72 MB
Peak swap size 3,017.63 MB
Peak disk usage 0.05 MB




Stderr output

<core_client_version>7.6.22</core_client_version>
<![CDATA[
<stderr_txt>
OpenCL platform detected: Apple
Number of OpenCL devices found : 1
BOINC assigns slot on device #0.
Info: BOINC provided OpenCL device ID used
DOUBLE_FP supported.
cl_khr_fp64 supported.
cl_APPLE_fp64_basic_ops supported.
FERMI : true

Build features: SETI8 Non-graphics OpenCL USE_OPENCL_NV OCL_ZERO_COPY OCL_CHIRP3 ASYNC_SPIKE FFTW SSE3 64bit
System: Darwin x86_64 Kernel: 15.3.0
CPU : Intel(R) Core(TM)2 Extreme CPU X9650 @ 3.00GHz
GenuineIntel x86, Family 6 Model 23 Stepping 6
Features : FPU TSC PAE APIC MTRR MMX SSE SSE2 HT SSE3 SSSE3 SSE4.1

OpenCL-kernels filename : MultiBeam_Kernels_r3321.cl
ar=0.426068 NumCfft=196139 NumGauss=1108478748 NumPulse=226384665776 NumTriplet=452772060038
Currently allocated 209 MB for GPU buffers
In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768
OS X optimized setiathome_v8 application
Version info: SSE3x (Intel, Core 2-optimized v8-nographics) V5.13 by Alex Kan
SSE3x OS X 64bit Build 3321 , Ported by : Raistmer, JDWhale, Urs Echternacht


OpenCL version by Raistmer, r3321

Number of OpenCL platforms: 1


OpenCL Platform Name: Apple
Number of devices: 1
Max compute units: 5
Max work group size: 1024
Max clock frequency: 1254Mhz
Max memory allocation: 536870912
Cache type: None
Cache line size: 0
Cache size: 0
Global memory size: 2147483648
Constant buffer size: 65536
Max number of constant args: 9
Local memory type: Scratchpad
Local memory size: 49152
Queue properties:
Out-of-Order: No
Name: GeForce GTX 750 Ti
Vendor: NVIDIA
Driver version: 10.9.14 346.03.05f02
Version: OpenCL 1.2
Extensions: cl_APPLE_SetMemObjectDestructor cl_APPLE_ContextLoggingFunctions cl_APPLE_clut cl_APPLE_query_kernel_names cl_APPLE_gl_sharing cl_khr_gl_event cl_khr_byte_addressable_store cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_APPLE_fp64_basic_ops cl_khr_fp64 cl_khr_3d_image_writes cl_khr_depth_images cl_khr_gl_depth_images cl_khr_gl_msaa_sharing cl_khr_image2d_from_buffer cl_APPLE_ycbcr_422 cl_APPLE_rgb_422


Work Unit Info:
...............
Credit multiplier is : 2.85
WU true angle range is : 0.426068
Used GPU device parameters are:
Number of compute units: 5
Single buffer allocation size: 128MB
Total device global memory: 2048MB
max WG size: 1024
local mem type: Real
FERMI path used: yes
LotOfMem path: no
period_iterations_num=50
Spike: peak=24.08791, time=32.72, d_freq=1421127402.52, chirp=-7.2046, fft_len=16k
Spike: peak=24.19478, time=32.72, d_freq=1421127402.48, chirp=-7.224, fft_len=16k
Spike: peak=24.2678, time=32.72, d_freq=1421127402.52, chirp=-7.2591, fft_len=16k
Spike: peak=24.02498, time=32.72, d_freq=1421127402.45, chirp=-7.2795, fft_len=16k
Spike: peak=24.20604, time=32.72, d_freq=1421127402.53, chirp=-7.2952, fft_len=16k
Spike: peak=24.5584, time=32.72, d_freq=1421127402.5, chirp=-7.3146, fft_len=16k
Spike: peak=24.22543, time=32.72, d_freq=1421127402.46, chirp=-7.334, fft_len=16k
Spike: peak=24.25604, time=32.72, d_freq=1421127402.54, chirp=-7.3497, fft_len=16k
Spike: peak=24.46971, time=32.72, d_freq=1421127402.47, chirp=-7.3701, fft_len=16k
Spike: peak=24.03967, time=32.72, d_freq=1421127402.55, chirp=-7.3858, fft_len=16k
Spike: peak=24.63006, time=32.72, d_freq=1421127402.51, chirp=-7.4052, fft_len=16k
Spike: peak=24.53827, time=32.72, d_freq=1421127402.47, chirp=-7.4246, fft_len=16k
Spike: peak=24.51567, time=32.72, d_freq=1421127402.53, chirp=-7.4412, fft_len=16k
Spike: peak=24.62417, time=32.72, d_freq=1421127402.49, chirp=-7.4606, fft_len=16k
Spike: peak=24.05137, time=32.72, d_freq=1421127402.45, chirp=-7.48, fft_len=16k
Spike: peak=24.41826, time=32.72, d_freq=1421127402.53, chirp=-7.4958, fft_len=16k
Spike: peak=24.23978, time=32.72, d_freq=1421127402.46, chirp=-7.5161, fft_len=16k
Spike: peak=24.15542, time=32.72, d_freq=1421127402.54, chirp=-7.5318, fft_len=16k
Spike: peak=24.49376, time=32.72, d_freq=1421127402.5, chirp=-7.5512, fft_len=16k
Spike: peak=24.15768, time=32.72, d_freq=1421127402.46, chirp=-7.5706, fft_len=16k
Spike: peak=24.19212, time=32.72, d_freq=1421127402.48, chirp=-7.6067, fft_len=16k
Spike: peak=24.09331, time=32.72, d_freq=1421127402.52, chirp=-7.6418, fft_len=16k
Autocorr: peak=18.06716, time=33.55, delay=5.7593, d_freq=1421123608.8, chirp=16.747, fft_len=128k
Spike: peak=24.49231, time=87.24, d_freq=1421121090.41, chirp=28.035, fft_len=128k
Spike: peak=24.08869, time=87.24, d_freq=1421121090.42, chirp=28.036, fft_len=128k
Pulse: peak=1.950637, time=75.51, period=0.485, d_freq=1421120676.87, score=1.035, chirp=44.897, fft_len=256
Pulse: peak=1.999583, time=75.51, period=0.485, d_freq=1421120676.83, score=1.061, chirp=45.402, fft_len=256
Pulse: peak=2.024932, time=75.51, period=0.485, d_freq=1421120676.79, score=1.075, chirp=45.906, fft_len=256
Pulse: peak=2.024149, time=75.51, period=0.485, d_freq=1421120676.68, score=1.075, chirp=46.41, fft_len=256
Pulse: peak=1.993112, time=75.51, period=0.485, d_freq=1421120676.64, score=1.058, chirp=46.915, fft_len=256
OpenCL queue synchronized
SETI@Home Informational message -9 result_overflow
NOTE: The number of results detected equals the storage space allocated.

Best spike: peak=24.63006, time=32.72, d_freq=1421127402.51, chirp=-7.4052, fft_len=16k
Best autocorr: peak=18.06716, time=33.55, delay=5.7593, d_freq=1421123608.8, chirp=16.747, fft_len=128k
Best gaussian: peak=3.18197, mean=0.5231707, ChiSq=1.334423, time=67.95, d_freq=1421118179.57,
score=-0.7257614, null_hyp=2.164555, chirp=11.378, fft_len=16k
Best pulse: peak=2.024932, time=75.51, period=0.485, d_freq=1421120676.79, score=1.075, chirp=45.906, fft_len=256
Best triplet: peak=0, time=-2.121e+11, period=0, d_freq=0, chirp=0, fft_len=0


Flopcounter: 15205069976953.570312

Spike count: 24
Autocorr count: 1
Pulse count: 5
Triplet count: 0
Gaussian count: 0
Time cpu in use since last restart: 414.8 seconds
GPU device sync requested... ...GPU device synched
13:30:19 (1954): called boinc_finish(0)

</stderr_txt>
]]>
TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos
ID: 1771134 · Report as offensive
Chris Adamek
Volunteer tester

Send message
Joined: 15 May 99
Posts: 251
Credit: 434,772,072
RAC: 236
United States
Message 1771180 - Posted: 12 Mar 2016, 15:40:20 UTC - in response to Message 1771134.  

Which is why we were suggesting the Cuda app Tbar put together. Further optimization that is in work will make the newer version eventually be 2-3x faster than even the cuda 5/6.5 version he has now. You can take a look at his Mac to see how fast his 750ti's run if you want to see what the future looks like.:)

Chris
ID: 1771180 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1771274 - Posted: 12 Mar 2016, 23:30:21 UTC

There are some Mac's around and some are running stock.
Some have more processing units and some are faster.
Some have compiled their own and some are still waiting for.

http://setiathome.berkeley.edu/workunit.php?wuid=2091505693
To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1771274 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1771282 - Posted: 13 Mar 2016, 0:08:26 UTC - in response to Message 1771274.  

I don't see any SIGBUS Errors there. I've tried just about everything possible in Yosemite and I'm still getting these SIGBUS Errors. I suppose it's time to move up to El Capitan and try it there.
ID: 1771282 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 1771295 - Posted: 13 Mar 2016, 1:16:39 UTC - in response to Message 1771282.  
Last modified: 13 Mar 2016, 1:24:40 UTC

Flat Makefile build process under El-Capitan went pretty smoothly here (whether stock or pasting Petri's optimisations over the top), can't say the same for the traditional gnutools route (though still used ./configure to generate the sah_config for now).

My *guess* with the SIGBUS errors in yosemite is something to do with the boinc libraries, just because signals and busses sound like the stuff that boincapi+lib handle, signal trapping and IO of file and shared memory type. I used the prescribed Xcode project method for building those libraries, since the gnutools path seems unmaintained for Darwin in newer Boinc revisions.

Good to see some others getting their hands dirty to figure things out. Glad I put the Alpha code there so others can play while I keep trying to bolt down generic stock.

[Edit:] disparate buildsystems on Mac Vs Linux Vs Windows is the main sticking point, with each having a ridiculous number of quirks and breakages on top of Cuda version compatibility concerns. Have been testing out what will be needed to go to Gradle automation for the CPU portions, which will make future builds cross platform much more uniform (and eventually automatically regression tested and deployed to website on command).

Now I just need more monitors :/
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 1771295 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1771323 - Posted: 13 Mar 2016, 4:10:47 UTC - in response to Message 1771295.  

Seems to be slightly better in El Capitan, however, I'm still seeing SIGBUS Errors with the App compiled in Yosemite. I did see a bonic-master Error in Yosemite dealing with the graphics section;
/System/Library/Frameworks/Security.framework/Headers/SecAccessControl.h:43:1: error: expected function body after function declarator __OSX_AVAILABLE_STARTING(__MAC_10_10, __IPHONE_8_0);
/System/Library/Frameworks/Security.framework/Headers/SecItem.h:421:39: error: expected ';' after top level declarator
extern CFTypeRef kSecAttrAccessControl
/System/Library/Frameworks/Foundation.framework/Headers/NSOperation.h:6:9: fatal error: 'sys/qos.h' file not found #import <sys/qos.h>
libtool: error: 'libboinc_graphics2_la-mac_icon.lo' is not a valid libtool object

This stops libboinc_graphics2_la from building, but doesn't seem to stop the two BOINC libraries needed by the compiler.
This is probably why I was never able to build an app from the seti-boinc folder, it kept failing with the graphics error.
Fortunately, you don't have that problem if you disable graphics.

The graphics don't even try to build in El Capitan, unfortunately, I'm back looking at the EXACT same ASM problems as before in Xcode 7.2.1. This time there is a new Error though;
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.11.sdk/System/Library/Frameworks/CoreFoundation.framework/Headers/CFLocale.h(172): error: extra text after expected end of number

I never saw that Error in Yosemite. I think I'm about to trash another 10.8 GBs of Xcode 7.2.1 and try an older version in El Capitan.
ID: 1771323 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 1771330 - Posted: 13 Mar 2016, 5:02:01 UTC - in response to Message 1771323.  
Last modified: 13 Mar 2016, 5:05:10 UTC

yeah can disable graphics and ignore those crappy OpenGL graphics libs.

When I enable graphics for XBranch, It'll be with Unreal Engine, and so Vulkan + DX12, and totally different non-screensaver purpose. Just because I want to cruise around in VR checking out the nanofarm...
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 1771330 · Report as offensive
Previous · 1 · 2 · 3

Message boards : Number crunching : MAC OS X El Capitan and NVIDIA Web Driver.


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.