I have a machine that has a Nvidia and an Intel GPU. Coprocessors:NVIDIA GeForce GTX 1080 Ti (4095MB) driver: 430.64 INTEL Intel(R) UHD Graphics 630 (3216MB)
If I select both GPU's in my preferences then I get work for both GPU's and they are returned normally. Examples are these tasks:
https://einsteinathome.org/task/852027168 and https://einsteinathome.org/task/852194731
If I only click the Intel GPU in my preferences and don't select Nvidia then I get this error:
2019-05-11 08:57:37.8980 [PID=5466] Request: [USER#xxxxx] [HOST#12777891] [IP xxx.xxx.xxx.163] client 7.14.2 2019-05-11 08:57:37.9908 [PID=5466 ] [debug] have_master:1 have_working: 1 have_db: 1 2019-05-11 08:57:37.9909 [PID=5466 ] [debug] using working prefs 2019-05-11 08:57:37.9909 [PID=5466 ] [debug] have db 1; dbmod 1548993257.000000; global mod 1548993257.000000 2019-05-11 08:57:37.9918 [PID=5466 ] [handle] [HOST#12777891] [RESULT#853302262] [WU#404170877] got result (DB: server_state=4 outcome=0 client_state=0 validate_state=0 delete_state=0) 2019-05-11 08:57:37.9918 [PID=5466 ] [handle] cpu time 1278.469000 credit/sec 0.013713, claimed credit 17.532164 2019-05-11 08:57:37.9919 [PID=5466 ] [handle] [RESULT#853302262] [WU#404170877]: setting outcome SUCCESS 2019-05-11 08:57:37.9919 [PID=5466 ] [handle] [HOST#12777891] [RESULT#853302260] [WU#404170876] got result (DB: server_state=4 outcome=0 client_state=0 validate_state=0 delete_state=0) 2019-05-11 08:57:37.9919 [PID=5466 ] [handle] cpu time 1274.125000 credit/sec 0.013713, claimed credit 17.472593 2019-05-11 08:57:37.9919 [PID=5466 ] [handle] [RESULT#853302260] [WU#404170876]: setting outcome SUCCESS 2019-05-11 08:57:38.0059 [PID=5466 ] [send] effective_ncpus 12 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999 2019-05-11 08:57:38.0059 [PID=5466 ] [send] effective_ngpus 2 max_jobs_on_host_gpu 999999 2019-05-11 08:57:38.0059 [PID=5466 ] [send] Not using matchmaker scheduling; Not using EDF sim 2019-05-11 08:57:38.0059 [PID=5466 ] [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00 2019-05-11 08:57:38.0059 [PID=5466 ] [send] CUDA: req 0.00 sec, 0.00 instances; est delay 0.00 2019-05-11 08:57:38.0059 [PID=5466 ] [send] Intel GPU: req 180.00 sec, 1.00 instances; est delay 0.00 2019-05-11 08:57:38.0059 [PID=5466 ] [send] work_req_seconds: 0.00 secs 2019-05-11 08:57:38.0059 [PID=5466 ] [send] available disk 99.87 GB, work_buf_min 0 2019-05-11 08:57:38.0060 [PID=5466 ] [send] active_frac 0.999974 on_frac 0.991527 DCF 0.880018 2019-05-11 08:57:38.0068 [PID=5466 ] [mixed] sending non-locality work first (0.9109) 2019-05-11 08:57:38.0294 [PID=5466 ] [version] Checking plan class 'FGRPopencl-ati' 2019-05-11 08:57:38.0333 [PID=5466 ] [version] reading plan classes from file '/BOINC/projects/EinsteinAtHome/plan_class_spec.xml' 2019-05-11 08:57:38.0333 [PID=5466 ] [version] parsed project prefs setting 'gpu_util_fgrp': 1.000000 2019-05-11 08:57:38.0333 [PID=5466 ] [version] No ATI devices found 2019-05-11 08:57:38.0333 [PID=5466 ] [version] Checking plan class 'FGRPopencl-nvidia' 2019-05-11 08:57:38.0333 [PID=5466 ] [version] parsed project prefs setting 'gpu_util_fgrp': 1.000000 2019-05-11 08:57:38.0333 [PID=5466 ] [version] NVidia compute capability: 601 2019-05-11 08:57:38.0334 [PID=5466 ] [version] Peak flops supplied: 7.71008e+11 2019-05-11 08:57:38.0334 [PID=5466 ] [version] plan class ok 2019-05-11 08:57:38.0334 [PID=5466 ] [version] Don't need CUDA jobs, skipping version 117 for hsgamma_FGRPB1G (FGRPopencl-nvidia) 2019-05-11 08:57:38.0334 [PID=5466 ] [version] Checking plan class 'FGRPopencl1K-ati' 2019-05-11 08:57:38.0334 [PID=5466 ] [version] parsed project prefs setting 'gpu_util_fgrp': 1.000000 2019-05-11 08:57:38.0334 [PID=5466 ] [version] No ATI devices found 2019-05-11 08:57:38.0334 [PID=5466 ] [version] Checking plan class 'FGRPopencl1K-nvidia' 2019-05-11 08:57:38.0334 [PID=5466 ] [version] parsed project prefs setting 'gpu_util_fgrp': 1.000000 2019-05-11 08:57:38.0334 [PID=5466 ] [2019-05-11 08:57:37.9614 [PID=5477 ] SCHEDULER_REQUEST::parse(): unrecognized: <allow_multiple_clients>0</allow_multiple_clients> 2019-05-11 08:57:38.0334 [PID=5466 ] [version] Peak flops supplied: 7.71008e+11 2019-05-11 08:57:38.0334 [PID=5466 ] [version] plan class ok 2019-05-11 08:57:38.0334 [PID=5466 ] [version] Don't need CUDA jobs, skipping version 122 for hsgamma_FGRPB1G (FGRPopencl1K-nvidia) 2019-05-11 08:57:38.0335 [PID=5466 ] [version] Checking plan class 'FGRPopenclTV-nvidia' 2019-05-11 08:57:38.0335 [PID=5466 ] [version] parsed project prefs setting 'gpu_util_fgrp': 1.000000 2019-05-11 08:57:38.0335 [PID=5466 ] [version] NVidia compute capability: 601 2019-05-11 08:57:38.0335 [PID=5466 ] [version] CUDA compute capability required min: 700, supplied: 601 2019-05-11 08:57:38.0335 [PID=5466 ] [version] no app version available: APP#40 (hsgamma_FGRPB1G) PLATFORM#9 (windows_x86_64) min_version 0 2019-05-11 08:57:38.0335 [PID=5466 ] [version] no app version available: APP#40 (hsgamma_FGRPB1G) PLATFORM#2 (windows_intelx86) min_version 0 2019-05-11 08:57:38.0335 [PID=5466 ] [version] Checking plan class 'opencl-intel_gpu' 2019-05-11 08:57:38.0336 [PID=5466 ] [version] parsed project prefs setting 'gpu_util_brp': 1.000000 2019-05-11 08:57:38.0336 [PID=5466 ] [version] Intel GPU device name: 'Intel(R) UHD Graphics 630' doesn't match 'HD Graphics [123]|HD Graphics 40' 2019-05-11 08:57:38.0336 [PID=5466 ] [version] Checking plan class 'opencl-intel_gpu-Beta' 2019-05-11 08:57:38.0336 [PID=5466 ] [version] beta test app versions not allowed in project prefs. 2019-05-11 08:57:38.0336 [PID=5466 ] [version] Checking plan class 'opencl-intel_gpu-new' 2019-05-11 08:57:38.0336 [PID=5466 ] [version] parsed project prefs setting 'gpu_util_brp': 1.000000 2019-05-11 08:57:38.0336 [PID=5466 ] [version] [HOST#12777891] device name: 'Intel(R) UHD Graphics 630'; OpenCL driver version: 25.20.100.6374; platform version: OpenCL 2.1; device version: OpenCL 2.1 NEO 2019-05-11 08:57:38.0336 [PID=5466 ] [version] driver version 2521006374, min: 0, max: 1018103906 2019-05-11 08:57:38.0336 [PID=5466 ] [version] driver version required max: 1018103906, supplied: 2521006374 2019-05-11 08:57:38.0336 [PID=5466 ] [version] Checking plan class 'opencl-intel_gpu' 2019-05-11 08:57:38.0337 [PID=5466 ] [version] parsed project prefs setting 'gpu_util_brp': 1.000000 2019-05-11 08:57:38.0337 [PID=5466 ] [version] Intel GPU device name: 'Intel(R) UHD Graphics 630' doesn't match 'HD Graphics [123]|HD Graphics 40' 2019-05-11 08:57:38.0337 [PID=5466 ] [version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#9 (windows_x86_64) min_version 0 2019-05-11 08:57:38.0337 [PID=5466 ] [version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#2 (windows_intelx86) min_version 0 2019-05-11 08:57:38.0346 [PID=5466 ] [version] Checking plan class 'FGRPSSE' 2019-05-11 08:57:38.0346 [PID=5466 ] [version] plan class ok 2019-05-11 08:57:38.0346 [PID=5466 ] [version] Don't need CPU jobs, skipping version 108 for hsgamma_FGRP5 (FGRPSSE) 2019-05-11 08:57:38.0346 [PID=5466 ] [version] no app version available: APP#46 (hsgamma_FGRP5) PLATFORM#9 (windows_x86_64) min_version 0 2019-05-11 08:57:38.0346 [PID=5466 ] [version] no app version available: APP#46 (hsgamma_FGRP5) PLATFORM#2 (windows_intelx86) min_version 0 2019-05-11 08:57:38.0400 [PID=5466 ] [mixed] sending locality work second 2019-05-11 08:57:38.0428 [PID=5466 ] [debug] [HOST#12777891] MSG(high) No work sent 2019-05-11 08:57:38.0428 [PID=5466 ] [debug] [HOST#12777891] MSG(high) see scheduler log messages on https://einsteinathome.org/host/12777891/log 2019-05-11 08:57:38.0428 [PID=5466 ] Sending reply to [HOST#12777891]: 0 results, delay req 60.00 2019-05-11 08:57:38.0429 [PID=5466 ] Scheduler ran 0.149 seconds
BOINC finds both GPU's
11/05/2019 5:48:50 PM | | CUDA: NVIDIA GPU 0: GeForce GTX 1080 Ti (driver version 430.64, CUDA version 10.2, compute capability 6.1, 4096MB, 3546MB available, 12336 GFLOPS peak)
11/05/2019 5:48:50 PM | | OpenCL: NVIDIA GPU 0: GeForce GTX 1080 Ti (driver version 430.64, device version OpenCL 1.2 CUDA, 11264MB, 3546MB available, 12336 GFLOPS peak)
11/05/2019 5:48:50 PM | | OpenCL: Intel GPU 0: Intel(R) UHD Graphics 630 (driver version 25.20.100.6374, device version OpenCL 2.1 NEO, 3217MB, 3217MB available, 230 GFLOPS peak)
11/05/2019 5:48:50 PM | | OpenCL CPU: Intel(R) Core(TM) i7-8700 CPU @ 3.20GHz (OpenCL driver vendor: Intel(R) Corporation, driver version 7.6.0.716, device version OpenCL 2.1 (Build 716))
Proud Founder of
Have a look at my WebCam<
Copyright © 2024 Einstein@Home. All rights reserved.
In your successful task (with
)
In your successful task (with both GPUs available), it reports the application used as
Binary Radio Pulsar Search (Arecibo) v1.34 (opencl-intel_gpu-Beta)
but in the extended server log, there's a line
2019-05-11 08:57:38.0336 [PID=5466 ] [version] beta test app versions not allowed in project prefs.
It seems as if you de-selected Beta apps as well when you turned off the NVidia card. The later Intel GPUs require the Beta app.
Thanks I didn't see that but
)
Thanks I didn't see that but it now gives a different error:
https://einsteinathome.org/host/12777891/log
Proud Founder of
Have a look at my WebCam<
I still can't run Intel GPU
)
I still can't run Intel GPU on it's own without AMD or NVIDIA GPU turned on.
Proud Founder of
Have a look at my WebCam<
Quote:Only one Beta app
)
isn't an error message - it's just to let you know that there's a protective measure put in place by the project to prevent rogue apps producing invalid science without being checked.
Because so many of us are new using the Beta app, it can be quite difficult to get your first allocation of work - BOINC delays a long time after getting a succession of "no work supplied" responses. But once you have work, boinc will go on fetching replacements as each task finishes, bypassing the delays.
Your best bet is to find a time when you can babysit the machine, and keep clicking the update button every couple of minutes until one of them strikes lucky: once you've got a few tasks cached, it should run by itself from there on.