Through all the trees I cannot see the wood anymore. I am reading and reading. But what drivers do I need for the Intel GPU to run WU's under Windows10 and where are those to find.
WU's run smooth on my CPU-GPU under Windows7. However I was forced to change to Windows 10 on this PC and with the drivers installed that I used under Windows 7 it is not working.
I'm not sure if you are addressing your concerns directly at me or perhaps more generally :-).
I don't run Windows of any flavour and I have never tried to use an Intel iGPU. When they first appeared, Intel didn't supply any support for Linux. I don't know if that has changed or not. Anything I do know (most of which I've probably forgotten) is as a result of reading threads like this one.
If I were in your shoes, I would try a couple of approaches. As you had a working driver for Windows 7, can you go back to wherever you sourced the driver and see if you can find versions for Windows 10? If there are multiple versions available, try a few different ones.
If that doesn't work, I would browse through any threads here that discuss Intel iGPUs. For any of the people who have posted about using one, take a look at the list of computers associated with each name (most people have not hidden their computers) and see if you can find any hosts using Windows 10 and crunching iGPU tasks. Shouldn't be too hard to find some. Send any such users a polite PM asking if they would mind sharing any tricks/tips about how they got things working.
If neither of the above give you a solution, you could try googling something like, "windows 10 driver for <your model> intel gpu". You could also try adding extra terms like 'boinc' and/or 'distributed computing' or 'crunching' into the mix.
You could also try searching on the BOINC website. I only browse there very occasionally so don't have a good feel for what you might find but I'd be surprised if there aren't some posts about Windows 10 and iGPUs and suitable drivers that you will be able to find there.
I've been crunching with Intel GPU till two days ago but now not receiving any job without doing anything in the config neither in the web and I get this feedback from the web:
I also have re-started the project twice and made a double check of the site preferences and it's accepted test app, so I do not understand why in below log says NO
2018-09-03 10:58:31.1785 [PID=2386] Request: [USER#xxxxx] [HOST#12568941] [IP xxx.xxx.xxx.93] client 7.12.1
2018-09-03 10:58:31.2492 [PID=2386 ] [debug] have_master:1 have_working: 1 have_db: 1
2018-09-03 10:58:31.2492 [PID=2386 ] [debug] using working prefs
2018-09-03 10:58:31.2492 [PID=2386 ] [debug] have db 1; dbmod 1475423097.000000; global mod 1475423097.000000
2018-09-03 10:58:31.2492 [PID=2386 ] [send] effective_ncpus 8 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
2018-09-03 10:58:31.2492 [PID=2386 ] [send] effective_ngpus 2 max_jobs_on_host_gpu 999999
2018-09-03 10:58:31.2492 [PID=2386 ] [send] Not using matchmaker scheduling; Not using EDF sim
2018-09-03 10:58:31.2492 [PID=2386 ] [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
2018-09-03 10:58:31.2492 [PID=2386 ] [send] CUDA: req 0.00 sec, 0.00 instances; est delay 0.00
2018-09-03 10:58:31.2492 [PID=2386 ] [send] Intel GPU: req 249772.99 sec, 1.00 instances; est delay 0.00
2018-09-03 10:58:31.2492 [PID=2386 ] [send] work_req_seconds: 0.00 secs
2018-09-03 10:58:31.2493 [PID=2386 ] [send] available disk 25.64 GB, work_buf_min 259200
2018-09-03 10:58:31.2493 [PID=2386 ] [send] active_frac 0.999838 on_frac 0.974200 DCF 1.000000
2018-09-03 10:58:31.2502 [PID=2386 ] [resend] [HOST#12568941] found lost [RESULT#779722267]: p2030.20161124.G202.61+00.54.S.b6s0g0.00000_262_1
2018-09-03 10:58:31.2513 [PID=2386 ] [version] Checking plan class 'opencl-intel_gpu'
2018-09-03 10:58:31.2554 [PID=2386 ] [version] reading plan classes from file '/BOINC/projects/EinsteinAtHome/plan_class_spec.xml'
2018-09-03 10:58:31.2554 [PID=2386 ] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
2018-09-03 10:58:31.2555 [PID=2386 ] [version] Intel GPU device name: 'Intel(R) HD Graphics 4600' doesn't match 'HD Graphics [123]|HD Graphics 40'
2018-09-03 10:58:31.2555 [PID=2386 ] [version] Checking plan class 'opencl-intel_gpu-Beta'
2018-09-03 10:58:31.2555 [PID=2386 ] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
2018-09-03 10:58:31.2555 [PID=2386 ] [version] [HOST#12568941] device name: 'Intel(R) HD Graphics 4600'; OpenCL driver version: 20.19.15.4531; platform version: OpenCL 1.2; device version: OpenCL 1.2
2018-09-03 10:58:31.2555 [PID=2386 ] [version] Peak flops supplied: 1.92e+11
2018-09-03 10:58:31.2555 [PID=2386 ] [version] plan class ok
2018-09-03 10:58:31.2555 [PID=2386 ] [version] beta test app versions not allowed in project prefs.
2018-09-03 10:58:31.2555 [PID=2386 ] [version] Checking plan class 'opencl-intel_gpu-new'
2018-09-03 10:58:31.2555 [PID=2386 ] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
2018-09-03 10:58:31.2555 [PID=2386 ] [version] [HOST#12568941] device name: 'Intel(R) HD Graphics 4600'; OpenCL driver version: 20.19.15.4531; platform version: OpenCL 1.2; device version: OpenCL 1.2
2018-09-03 10:58:31.2555 [PID=2386 ] [version] driver version 2019154531, min: 0, max: 1018103906
2018-09-03 10:58:31.2555 [PID=2386 ] [version] driver version required max: 1018103906, supplied: 2019154531
2018-09-03 10:58:31.2556 [PID=2386 ] [version] Checking plan class 'opencl-intel_gpu'
2018-09-03 10:58:31.2556 [PID=2386 ] [version] parsed project prefs setting 'gpu_util_brp': 1.000000
2018-09-03 10:58:31.2556 [PID=2386 ] [version] Intel GPU device name: 'Intel(R) HD Graphics 4600' doesn't match 'HD Graphics [123]|HD Graphics 40'
2018-09-03 10:58:31.2556 [PID=2386 ] [version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#9 (windows_x86_64) min_version 0
2018-09-03 10:58:31.2556 [PID=2386 ] [version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#2 (windows_intelx86) min_version 0
2018-09-03 10:58:31.2556 [PID=2386 ] [CRITICAL] [HOST#12568941] can't resend [RESULT#779722267]: no app version for einsteinbinary_BRP4
2018-09-03 10:58:31.2556 [PID=2386 ] [resend] [HOST#12568941] 1 lost results, resent 0
2018-09-03 10:58:31.2557 [PID=2386 ] [mixed] sending locality work first (0.0065)
2018-09-03 10:58:31.2557 [PID=2386 ] [mixed] sending non-locality work second
2018-09-03 10:58:31.2734 [PID=2386 ] [send] [HOST#12568941] will accept beta work. Scanning for beta work.
2018-09-03 10:58:31.2841 [PID=2386 ] [version] Checking plan class 'FGRPopencl-ati'
2018-09-03 10:58:31.2841 [PID=2386 ] [version] parsed project prefs setting 'gpu_util_fgrp': 1.000000
2018-09-03 10:58:31.2841 [PID=2386 ] [version] No ATI devices found
2018-09-03 10:58:31.2842 [PID=2386 ] [version] Checking plan class 'FGRPopencl-nvidia'
2018-09-03 10:58:31.2842 [PID=2386 ] [version] parsed project prefs setting 'gpu_util_fgrp': 1.000000
2018-09-03 10:58:31.2842 [PID=2386 ] [version] Peak flops supplied: 1.40928e+11
2018-09-03 10:58:31.2842 [PID=2386 ] [version] plan class ok
2018-09-03 10:58:31.2842 [PID=2386 ] [version] Don't need CUDA jobs, skipping version 117 for hsgamma_FGRPB1G (FGRPopencl-nvidia)
2018-09-03 10:58:31.2842 [PID=2386 ] [version] Checking plan class 'FGRPopencl1K-ati'
2018-09-03 10:58:31.2842 [PID=2386 ] [version] parsed project prefs setting 'gpu_util_fgrp': 1.000000
2018-09-03 10:58:31.2842 [PID=2386 ] [version] No ATI devices found
2018-09-03 10:58:31.2842 [PID=2386 ] [version] Checking plan class 'FGRPopencl1K-nvidia'
2018-09-03 10:58:31.2842 [PID=2386 ] [version] parsed project prefs setting 'gpu_util_fgrp': 1.000000
2018-09-03 10:58:31.2842 [PID=2386 ] [version] Peak flops supplied: 1.40928e+11
2018-09-03 10:58:31.2842 [PID=2386 ] [version] plan class ok
2018-09-03 10:58:31.2842 [PID=2386 ] [version] Don't need CUDA jobs, skipping version 120 for hsgamma_FGRPB1G (FGRPopencl1K-nvidia)
2018-09-03 10:58:31.2842 [PID=2386 ] [version] no app version available: APP#40 (hsgamma_FGRPB1G) PLATFORM#9 (windows_x86_64) min_version 0
2018-09-03 10:58:31.2842 [PID=2386 ] [version] no app version available: APP#40 (hsgamma_FGRPB1G) PLATFORM#2 (windows_intelx86) min_version 0
2018-09-03 10:58:31.2853 [PID=2386 ] [version] Checking plan class 'FGRPSSE'
2018-09-03 10:58:31.2853 [PID=2386 ] [version] plan class ok
2018-09-03 10:58:31.2853 [PID=2386 ] [version] Don't need CPU jobs, skipping version 108 for hsgamma_FGRP5 (FGRPSSE)
2018-09-03 10:58:31.2853 [PID=2386 ] [version] no app version available: APP#46 (hsgamma_FGRP5) PLATFORM#9 (windows_x86_64) min_version 0
2018-09-03 10:58:31.2853 [PID=2386 ] [version] no app version available: APP#46 (hsgamma_FGRP5) PLATFORM#2 (windows_intelx86) min_version 0
2018-09-03 10:58:31.2906 [PID=2386 ] [send] [HOST#12568941] is looking for work from a non-preferred application
2018-09-03 10:58:31.2992 [PID=2386 ] [debug] [HOST#12568941] MSG(high) No work sent
2018-09-03 10:58:31.2992 [PID=2386 ] [debug] [HOST#12568941] MSG(high) xxxxx
2018-09-03 10:58:31.2993 [PID=2386 ] Sending reply to [HOST#12568941]: 0 results, delay req 60.00
2018-09-03 10:58:31.2994 [PID=2386 ] Scheduler ran 0.125 seconds
I'll preface these comments by noting that I know nothing about using Intel GPUs for crunching so I could easily be missing the blindingly obvious.
I've looked through the server log you supplied and I've also looked at what was sent back with one of your previously completed successful results that is still visible in the online database. The completed result showed that the app used was "Binary Radio Pulsar Search (Arecibo) v1.34 (opencl-intel_gpu-Beta)". I could be wrong but my impression is that to be using an app labeled 'beta', you have to have the setting for test apps enabled.
With that in mind, I'm a bit puzzled why the log snip says, at one point (I've truncated time & date, etc):-
.... Checking plan class 'opencl-intel_gpu-Beta'
.... parsed project prefs setting 'gpu_util_brp': 1.000000
.... [HOST#12568941] device name: 'Intel(R) HD Graphics 4600'; OpenCL driver version: 20.19.15.4531;
platform version: OpenCL 1.2; device version: OpenCL 1.2
.... Peak flops supplied: 1.92e+11
.... plan class ok
.... beta test app versions not allowed in project prefs.
That last line seems to suggest that you don't have test apps enabled???
So the first thing I'd suggest is to check that you really do have the setting for test apps enabled for the location (venue) that the host is assigned to (if you're using more than just the default location). If you really think you have, I'd try changing it to the opposite setting, making sure you save the change, and then change it back and save the 2nd change as well. Then see if the server will send you some new work.
There seems to be another 'plan class' that the server checked in order to see if you were eligible for work. Again, here is the relevant bit:-
.... Checking plan class 'opencl-intel_gpu-new'
.... parsed project prefs setting 'gpu_util_brp': 1.000000
.... [HOST#12568941] device name: 'Intel(R) HD Graphics 4600'; OpenCL driver version: 20.19.15.4531;
platform version: OpenCL 1.2; device version: OpenCL 1.2
.... driver version 2019154531, min: 0, max: 1018103906
.... driver version required max: 1018103906, supplied: 2019154531
This test was originally implemented to allow only known good driver versions to get work. As you can see, your driver version is higher than the allowed maximum. Unless there has been a recent driver update from a version that was "1018103906" or lower, I can't see that this would be anything to do with being refused work. However, are you sure there have been no recent updates from Microsoft that may have had anything to do with the iGPU drivers?
- i7-4771 CPU @ 3.50GHz, Number of processors:8
- NVIDIA GeForce GTX 1050 Ti (4095MB) driver: 399.07
- INTEL Intel(R) HD Graphics 4600 (1629MB)
- Microsoft Windows 10 Professional x64 Edition, (10.00.17134.00)
I had a real big problem with computer b4, and at last it burst last August so backep up BOINC wharehouse folder and start from the scratch again in computer, fdisk, format and clean install of win 10 x64.
I started to restore the wharehouse folder and installed BOINC x64 again, and the computer was cruching perfectly even Einstein projects with iGPU too untill 3 days ago that started the problem of not receiving any unit with my previous log in the server.
the driver installed was the one with the o.s. 20.19.15.4531.
today after reading this post I went to device administrator and update the driver to 20.19.15.4624 I don't know if I'll receive new units yet??
I'm using home seeting set-up as:
RESOURCE SETTINGS
Resource share:
Use CPU: NO
Use NVIDIA GPU: NO
Use ATI GPU: YES
Use INTEL GPU: YES
this is the new server's log with the updated iGPU driver:
[debug] have_master:1 have_working: 1 have_db: 1
[debug] using working prefs
[debug] have db 1; dbmod 1535995396.000000; global mod 0.000000
[send] effective_ncpus 8 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
[send] effective_ngpus 2 max_jobs_on_host_gpu 999999
[send] Not using matchmaker scheduling; Not using EDF sim
[send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
[send] CUDA: req 0.00 sec, 0.00 instances; est delay 0.00
[send] Intel GPU: req 87009.85 sec, 1.00 instances; est delay 0.00
[send] work_req_seconds: 0.00 secs
[send] available disk 25.68 GB, work_buf_min 259200
[send] active_frac 0.999860 on_frac 0.975401 DCF 1.000000
[resend] [HOST#xxxx] found lost [RESULT#779722267]: p2030.20161124.G202.61+00.54.S.b6s0g0.00000_262_1
[version] Checking plan class 'opencl-intel_gpu'
[version] reading plan classes from file '/BOINC/projects/EinsteinAtHome/plan_class_spec.xml'
[version] parsed project prefs setting 'gpu_util_brp': 1.000000
[version] Intel GPU device name: 'Intel(R) HD Graphics 4600' doesn't match 'HD Graphics [123]|HD Graphics 40'
[version] Checking plan class 'opencl-intel_gpu-Beta'
[version] parsed project prefs setting 'gpu_util_brp': 1.000000
[version] [HOST#xxxx] device name: 'Intel(R) HD Graphics 4600'; OpenCL driver version: 20.19.15.4624; platform version: OpenCL 1.2; device version: OpenCL 1.2
[version] Peak flops supplied: 1.92e+11
[version] plan class ok
[version] beta test app versions not allowed in project prefs.
[version] Checking plan class 'opencl-intel_gpu-new'
[version] parsed project prefs setting 'gpu_util_brp': 1.000000
[version] [HOST#xxxx] device name: 'Intel(R) HD Graphics 4600'; OpenCL driver version: 20.19.15.4624; platform version: OpenCL 1.2; device version: OpenCL 1.2
[version] driver version 2019154624, min: 0, max: 1018103906
[version] driver version required max: 1018103906, supplied: 2019154624
[version] Checking plan class 'opencl-intel_gpu'
[version] parsed project prefs setting 'gpu_util_brp': 1.000000
[version] Intel GPU device name: 'Intel(R) HD Graphics 4600' doesn't match 'HD Graphics [123]|HD Graphics 40'
[version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#9 (windows_x86_64) min_version 0
[version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#2 (windows_intelx86) min_version 0
[CRITICAL] [HOST#xxxx] can't resend [RESULT#779722267]: no app version for einsteinbinary_BRP4
[resend] [HOST#xxxx] 1 lost results, resent 0
[mixed] sending locality work first (0.1881)
[mixed] sending non-locality work second
[send] [HOST#xxxx] will accept beta work. Scanning for beta work.
[version] Checking plan class 'FGRPSSE'
[version] plan class ok
[version] Don't need CPU jobs, skipping version 108 for hsgamma_FGRP5 (FGRPSSE)
[version] no app version available: APP#46 (hsgamma_FGRP5) PLATFORM#9 (windows_x86_64) min_version 0
[version] no app version available: APP#46 (hsgamma_FGRP5) PLATFORM#2 (windows_intelx86) min_version 0
[send] [HOST#xxxx] is looking for work from a non-preferred application
[debug] [HOST#xxxx] MSG(high) No work sent
[debug] [HOST#xxxx] MSG(high) see scheduler log messages on https://einsteinathome.org/host/12568941/log
Sending reply to [HOST#xxxx]: 0 results, delay req 60.00
Scheduler ran 0.118 seconds
[version] beta test app versions not allowed in project prefs.
This line is still present in the server log.
If you go to your account and then click on the computer in question, what's show in the box next to "Location" close to the bottom of the page?
And then if you go to a comparison view of the project preferences, do you have the right preferences in the column matching the location for the computer?
Do you understand what the server log is telling you? It is still saying:-
[version] beta test app versions not allowed in project prefs.
Your successfully returned results in the online database show that you were using the beta app previously. If you want to crunch as before, you have to get your preferences to say that test apps are allowed. I know you say that you do have test apps allowed, but, for some reason, the server doesn't see that. I guess it could be some sort of new bug with the server code.
Have you tried what I suggested last time about changing the preference to 'NO', clicking 'SAVE', then changing back to 'YES', clicking 'SAVE' again? There is a 'save changes' button right down at the bottom of the page.
On that project preferences page where you check the setting for beta/test apps, there is a drop-down box at the top left which shows the 'location' (generic/home/school/work) whose preferences you are actually changing. Have you checked that whatever location shows in that box is the same as the location that is listed in the BOINC Manager event log? Check the startup messages when you first launch BOINC to see what location your host is set to.
If you can't get the server to accept that you wish to use test apps, your only other option is to downgrade your driver, NOT upgrade it. Unless you can use test apps, the server is restricted (as shown in the log) and can't send tasks if your driver is more recent than 10.18.10.3906.
EDIT: Holmis' reply hadn't been posted when I started composing the above. I didn't see it until after I posted. We are basically giving you the same advice.
Hello, Through all the
)
Hello,
Through all the trees I cannot see the wood anymore. I am reading and reading. But what drivers do I need for the Intel GPU to run WU's under Windows10 and where are those to find.
WU's run smooth on my CPU-GPU under Windows7. However I was forced to change to Windows 10 on this PC and with the drivers installed that I used under Windows 7 it is not working.
Thank you for your guiding/help.
Greetings from
TJ
TJ_14 wrote:Thank you for
)
I'm not sure if you are addressing your concerns directly at me or perhaps more generally :-).
I don't run Windows of any flavour and I have never tried to use an Intel iGPU. When they first appeared, Intel didn't supply any support for Linux. I don't know if that has changed or not. Anything I do know (most of which I've probably forgotten) is as a result of reading threads like this one.
If I were in your shoes, I would try a couple of approaches. As you had a working driver for Windows 7, can you go back to wherever you sourced the driver and see if you can find versions for Windows 10? If there are multiple versions available, try a few different ones.
If that doesn't work, I would browse through any threads here that discuss Intel iGPUs. For any of the people who have posted about using one, take a look at the list of computers associated with each name (most people have not hidden their computers) and see if you can find any hosts using Windows 10 and crunching iGPU tasks. Shouldn't be too hard to find some. Send any such users a polite PM asking if they would mind sharing any tricks/tips about how they got things working.
If neither of the above give you a solution, you could try googling something like, "windows 10 driver for <your model> intel gpu". You could also try adding extra terms like 'boinc' and/or 'distributed computing' or 'crunching' into the mix.
You could also try searching on the BOINC website. I only browse there very occasionally so don't have a good feel for what you might find but I'd be surprised if there aren't some posts about Windows 10 and iGPUs and suitable drivers that you will be able to find there.
Good luck with your research :-).
Cheers,
Gary.
Thank you for your advise
)
Thank you for your advise Gary.
Indeed it was a general message. I did search here in several discussions and help desk, but could not find a helpful link hence my question.
Greetings from
TJ
Hi, I've been crunching
)
Hi,
I've been crunching with Intel GPU till two days ago but now not receiving any job without doing anything in the config neither in the web and I get this feedback from the web:
I also have re-started the project twice and made a double check of the site preferences and it's accepted test app, so I do not understand why in below log says NO
Roses don't bloom on the sailor's grave
I'll preface these comments
)
I'll preface these comments by noting that I know nothing about using Intel GPUs for crunching so I could easily be missing the blindingly obvious.
I've looked through the server log you supplied and I've also looked at what was sent back with one of your previously completed successful results that is still visible in the online database. The completed result showed that the app used was "Binary Radio Pulsar Search (Arecibo) v1.34 (opencl-intel_gpu-Beta)". I could be wrong but my impression is that to be using an app labeled 'beta', you have to have the setting for test apps enabled.
With that in mind, I'm a bit puzzled why the log snip says, at one point (I've truncated time & date, etc):-
That last line seems to suggest that you don't have test apps enabled???
So the first thing I'd suggest is to check that you really do have the setting for test apps enabled for the location (venue) that the host is assigned to (if you're using more than just the default location). If you really think you have, I'd try changing it to the opposite setting, making sure you save the change, and then change it back and save the 2nd change as well. Then see if the server will send you some new work.
There seems to be another 'plan class' that the server checked in order to see if you were eligible for work. Again, here is the relevant bit:-
This test was originally implemented to allow only known good driver versions to get work. As you can see, your driver version is higher than the allowed maximum. Unless there has been a recent driver update from a version that was "1018103906" or lower, I can't see that this would be anything to do with being refused work. However, are you sure there have been no recent updates from Microsoft that may have had anything to do with the iGPU drivers?
Cheers,
Gary.
Yes, a go back drivers seems
)
Yes, a go back drivers seems necessary.
I use 10.18.10.4425 under Windows Seven with no problem
I have: - i7-4771 CPU @
)
I have:
- i7-4771 CPU @ 3.50GHz, Number of processors:8
- NVIDIA GeForce GTX 1050 Ti (4095MB) driver: 399.07
- INTEL Intel(R) HD Graphics 4600 (1629MB)
- Microsoft Windows 10 Professional x64 Edition, (10.00.17134.00)
I had a real big problem with computer b4, and at last it burst last August so backep up BOINC wharehouse folder and start from the scratch again in computer, fdisk, format and clean install of win 10 x64.
I started to restore the wharehouse folder and installed BOINC x64 again, and the computer was cruching perfectly even Einstein projects with iGPU too untill 3 days ago that started the problem of not receiving any unit with my previous log in the server.
the driver installed was the one with the o.s. 20.19.15.4531.
today after reading this post I went to device administrator and update the driver to 20.19.15.4624 I don't know if I'll receive new units yet??
I'm using home seeting set-up as:
RESOURCE SETTINGS
Resource share:
Use CPU: NO
Use NVIDIA GPU: NO
Use ATI GPU: YES
Use INTEL GPU: YES
BETA SETTINGS
Run test applications?: YES
FTB I do not see what should be the problem?
Roses don't bloom on the sailor's grave
this is the new server's log
)
[debug] have_master:1 have_working: 1 have_db: 1
[debug] using working prefs
[debug] have db 1; dbmod 1535995396.000000; global mod 0.000000
[send] effective_ncpus 8 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
[send] effective_ngpus 2 max_jobs_on_host_gpu 999999
[send] Not using matchmaker scheduling; Not using EDF sim
[send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
[send] CUDA: req 0.00 sec, 0.00 instances; est delay 0.00
[send] Intel GPU: req 87009.85 sec, 1.00 instances; est delay 0.00
[send] work_req_seconds: 0.00 secs
[send] available disk 25.68 GB, work_buf_min 259200
[send] active_frac 0.999860 on_frac 0.975401 DCF 1.000000
[resend] [HOST#xxxx] found lost [RESULT#779722267]: p2030.20161124.G202.61+00.54.S.b6s0g0.00000_262_1
[version] Checking plan class 'opencl-intel_gpu'
[version] reading plan classes from file '/BOINC/projects/EinsteinAtHome/plan_class_spec.xml'
[version] parsed project prefs setting 'gpu_util_brp': 1.000000
[version] Intel GPU device name: 'Intel(R) HD Graphics 4600' doesn't match 'HD Graphics [123]|HD Graphics 40'
[version] Checking plan class 'opencl-intel_gpu-Beta'
[version] parsed project prefs setting 'gpu_util_brp': 1.000000
[version] [HOST#xxxx] device name: 'Intel(R) HD Graphics 4600'; OpenCL driver version: 20.19.15.4624; platform version: OpenCL 1.2; device version: OpenCL 1.2
[version] Peak flops supplied: 1.92e+11
[version] plan class ok
[version] beta test app versions not allowed in project prefs.
[version] Checking plan class 'opencl-intel_gpu-new'
[version] parsed project prefs setting 'gpu_util_brp': 1.000000
[version] [HOST#xxxx] device name: 'Intel(R) HD Graphics 4600'; OpenCL driver version: 20.19.15.4624; platform version: OpenCL 1.2; device version: OpenCL 1.2
[version] driver version 2019154624, min: 0, max: 1018103906
[version] driver version required max: 1018103906, supplied: 2019154624
[version] Checking plan class 'opencl-intel_gpu'
[version] parsed project prefs setting 'gpu_util_brp': 1.000000
[version] Intel GPU device name: 'Intel(R) HD Graphics 4600' doesn't match 'HD Graphics [123]|HD Graphics 40'
[version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#9 (windows_x86_64) min_version 0
[version] no app version available: APP#19 (einsteinbinary_BRP4) PLATFORM#2 (windows_intelx86) min_version 0
[CRITICAL] [HOST#xxxx] can't resend [RESULT#779722267]: no app version for einsteinbinary_BRP4
[resend] [HOST#xxxx] 1 lost results, resent 0
[mixed] sending locality work first (0.1881)
[mixed] sending non-locality work second
[send] [HOST#xxxx] will accept beta work. Scanning for beta work.
[version] Checking plan class 'FGRPSSE'
[version] plan class ok
[version] Don't need CPU jobs, skipping version 108 for hsgamma_FGRP5 (FGRPSSE)
[version] no app version available: APP#46 (hsgamma_FGRP5) PLATFORM#9 (windows_x86_64) min_version 0
[version] no app version available: APP#46 (hsgamma_FGRP5) PLATFORM#2 (windows_intelx86) min_version 0
[send] [HOST#xxxx] is looking for work from a non-preferred application
[debug] [HOST#xxxx] MSG(high) No work sent
[debug] [HOST#xxxx] MSG(high) see scheduler log messages on https://einsteinathome.org/host/12568941/log
Sending reply to [HOST#xxxx]: 0 results, delay req 60.00
Scheduler ran 0.118 seconds
Roses don't bloom on the sailor's grave
Corsair_2 wrote:[version]
)
This line is still present in the server log.
If you go to your account and then click on the computer in question, what's show in the box next to "Location" close to the bottom of the page?
And then if you go to a comparison view of the project preferences, do you have the right preferences in the column matching the location for the computer?
Do you understand what the
)
Do you understand what the server log is telling you? It is still saying:-
[version] beta test app versions not allowed in project prefs.Your successfully returned results in the online database show that you were using the beta app previously. If you want to crunch as before, you have to get your preferences to say that test apps are allowed. I know you say that you do have test apps allowed, but, for some reason, the server doesn't see that. I guess it could be some sort of new bug with the server code.
Have you tried what I suggested last time about changing the preference to 'NO', clicking 'SAVE', then changing back to 'YES', clicking 'SAVE' again? There is a 'save changes' button right down at the bottom of the page.
On that project preferences page where you check the setting for beta/test apps, there is a drop-down box at the top left which shows the 'location' (generic/home/school/work) whose preferences you are actually changing. Have you checked that whatever location shows in that box is the same as the location that is listed in the BOINC Manager event log? Check the startup messages when you first launch BOINC to see what location your host is set to.
If you can't get the server to accept that you wish to use test apps, your only other option is to downgrade your driver, NOT upgrade it. Unless you can use test apps, the server is restricted (as shown in the log) and can't send tasks if your driver is more recent than 10.18.10.3906.
EDIT: Holmis' reply hadn't been posted when I started composing the above. I didn't see it until after I posted. We are basically giving you the same advice.
Cheers,
Gary.