Ok, just had my first AP v7.05 error out due to Exceeded Time Limit. Anyway to correct the estimate that comes with the APs to give them more time so that the APs finish and we can get a correct estimate after 11 validated work units? As it stands now, I don't think anyone will finish.
Thanks


Zalster

Just as before. Suspend the task(s) in the BOINC Manager. Stop BOINC, Open client_state.xml and search for Suspended. Copy the task name and search for the <workunit> entry for that task. Move the decimal point one space to the right in the entries;
<rsc_fpops_est>xxx</rsc_fpops_est>
<rsc_fpops_bound>xxx</rsc_fpops_bound>
Repeat for all required tasks. Save client_state.xml and start BOINC.

Now I'm getting an Error with ap_7.01r2559_sse3_OSX64;

<![CDATA[
<message>
too many exit(0)s
</message>
<stderr_txt>
Process creation (../../projects/setiathome.berkeley.edu/ap_7.01r2559_sse3_OSX64) failed: errno=13
Process creation (../../projects/setiathome.berkeley.edu/ap_7.01r2559_sse3_OSX64) failed: errno=13
Process creation (../../projects/setiathome.berkeley.edu/ap_7.01r2559_sse3_OSX64) failed: errno=13
....

This is a Permissions Issue. Someone didn't set the Permissions correctly on the App before posting it for Download.....arkayn!

I would have no way of setting permissions, I no longer have a Mac.

Flimsy excuse at best. Well I know where I got the last Mac App that did this, and I know how he told me to fix it ;-)

The same fix works on this App;
CD to the directory and set the Permissions with;
TomsMacPro:~ Tom$ cd /Volumes/Mov1/BOINC/Maverick/BOINC\ Data/projects/setiathome.berkeley.edu
TomsMacPro:setiathome.berkeley.edu Tom$ sudo chmod g+x ap_7.01r2559_sse3_OSX64
TomsMacPro:setiathome.berkeley.edu Tom$ sudo chmod g+w ap_7.01r2559_sse3_OSX64

The first line seems to be the important one. Now it works.

Ok, just had my first AP v7.05 error out due to Exceeded Time Limit. Anyway to correct the estimate that comes with the APs to give them more time so that the APs finish and we can get a correct estimate after 11 validated work units? As it stands now, I don't think anyone will finish.
Thanks

Zalster

If anyone else encounters this problem before their 11 validations are complete, they can:

Stop BOINC
Open client_state.xml for editing in a plain-text editor such as notepad.
Locate the <workunit> definition sections for the under-estimated AP tasks.
Increase the value for the <rsc_fpops_bound> tag by a substantial amount - say an extre zero or two.
Save file
Restart BOINC

- all with extreme care to avoid damaging any other part or the structure of the file.

All AP workunits should have the same <rsc_fpops_bound> value, so once you've found the first one, you should be able to fix them all with a global 'replace all'.

<rsc_fpops_bound> is the one that causes the 'Exceeded Time Limit' error: you can also change <rsc_fpops_est> - that will change the displayed runtime estimate (cosmetically) if you want, and prevent cache over-filling, but not prevent the error on its own.

Do we have a problem with the following? Intel iGPU is explicitly stated in global preferences as NO.

Invalid OpenCL GPU index: 1
WARNING: boinc_get_opencl_ids failed with code -33
OpenCL platform detected: Intel(R) Corporation
OpenCL platform detected: NVIDIA Corporation
WARNING: BOINC supplied wrong platform!

http://setiathome.berkeley.edu/result.php?resultid=3777429148

State of this result: Состояние проверки Правильный (Correct).
Sounds as error?

Looking at a wingman's computer

http://setiathome.berkeley.edu/show_host_detail.php?hostid=6869357

Zalster

Thanks for reporting, question about this BOINC error code rised on BOINC alpha list.

Is that related to the sticky at the top of this forum?

It's related to AstroPulse version 7.x release.

Do we have a problem with the following? Intel iGPU is explicitly stated in global preferences as NO.

The Global (Computing) Preferences don't have any preferences to set Intel GPUs to No.

http://setiathome.berkeley.edu/prefs.php?subset=global

Project preferences do have preferences to set Intel GPUs to No, but since they only apply to a single project, they are not global preferences.

http://setiathome.berkeley.edu/prefs.php?subset=project

Claggy

Do we have a problem with the following? Intel iGPU is explicitly stated in global preferences as NO.

The Global (Computing) Preferences don't have any preferences to set Intel GPUs to No.

http://setiathome.berkeley.edu/prefs.php?subset=global

Project preferences do have preferences to set Intel GPUs to No, but since they only apply to a single project, they are not global preferences.

http://setiathome.berkeley.edu/prefs.php?subset=project

Claggy

I agree I used the wrong choice of terminology, but the question still remains - Is this an actual error when the project preferences is set to NO for Intel iGPU and should it be a concern? If so, what needs to be done to rectify? Did not see this problem/condition on BETA.

Invalid OpenCL GPU index: 1
WARNING: boinc_get_opencl_ids failed with code -33
OpenCL platform detected: Intel(R) Corporation
OpenCL platform detected: NVIDIA Corporation
WARNING: BOINC supplied wrong platform!

---

I don't buy computers, I build them!!

I agree I used the wrong choice of terminology, but the question still remains - Is this an actual error when the project preferences is set to NO for Intel iGPU and should it be a concern? If so, what needs to be done to rectify? Did not see this problem/condition on BETA.

Invalid OpenCL GPU index: 1
WARNING: boinc_get_opencl_ids failed with code -33
OpenCL platform detected: Intel(R) Corporation
OpenCL platform detected: NVIDIA Corporation
WARNING: BOINC supplied wrong platform!

Nothing to do with the project preferences, The server supplied it to Application version AstroPulse v7 Anonymous platform (NVIDIA GPU),

But boinc_get_opencl_ids got a bit confused, and Raistmer's fallback code sent to the right device,
and was computed on the GTX750Ti, it built kernels for it:

INFO: can't open binary kernel file: D:\BOINC/projects/setiathome.berkeley.edu\AstroPulse_Kernels_r2721.cl_GeForceGTX750Ti.bin_V7_TWIN_FFA_34411, continue with recompile...
Info : Building Program (binary, clBuildProgram):main kernels: OK code 0
INFO: binary kernel file created
WARNING: can't open binary kernel file for oclFFT plan: D:\BOINC/projects/setiathome.berkeley.edu\AP_clFFTplan_GeForceGTX750Ti_32768_gr64_lr8_wg256_tw0_r2721.bin_34411, continue with recompile...
GPU device sync requested... ...GPU device synched

There already has been a fix, (over two weeks ago) But no apps have been deployed with this fix yet:

http://boinc.berkeley.edu/gitweb/?p=boinc-v2.git;a=commit;h=328d87be2625ce68a04b0d1caf29e3826eab25f3

API: Fix a bug in boinc_get_opencl_ids() which incorrectly returned CL_INVALID_DEVICE on some systems. This would happen on hosts with multiple GPUs if:
* the host has GPUs from two different vendors with different OpenCL platforms (e.g., NVIDIA and AMD), and
* the second platform queried has more GPUs than the first one, and
* the requested GPU is a higher numbed one in the second platform.

Claggy

I agree I used the wrong choice of terminology, but the question still remains - Is this an actual error when the project preferences is set to NO for Intel iGPU and should it be a concern? If so, what needs to be done to rectify? Did not see this problem/condition on BETA.

Invalid OpenCL GPU index: 1
WARNING: boinc_get_opencl_ids failed with code -33
OpenCL platform detected: Intel(R) Corporation
OpenCL platform detected: NVIDIA Corporation
WARNING: BOINC supplied wrong platform!

Nothing to do with the project preferences, The server supplied it to Application version AstroPulse v7 Anonymous platform (NVIDIA GPU),

But boinc_get_opencl_ids got a bit confused, and Raistmer's fallback code sent to the right device,
and was computed on the GTX750Ti, it built kernels for it:


There already has been a fix, (over two weeks ago) But no apps have been

INFO: can't open binary kernel file: D:\BOINC/projects/setiathome.berkeley.edu\AstroPulse_Kernels_r2721.cl_GeForceGTX750Ti.bin_V7_TWIN_FFA_34411, continue with recompile...
Info : Building Program (binary, clBuildProgram):main kernels: OK code 0
INFO: binary kernel file created
WARNING: can't open binary kernel file for oclFFT plan: D:\BOINC/projects/setiathome.berkeley.edu\AP_clFFTplan_GeForceGTX750Ti_32768_gr64_lr8_wg256_tw0_r2721.bin_34411, continue with recompile...
GPU device sync requested... ...GPU device synched

deployed with this fix yet:

http://boinc.berkeley.edu/gitweb/?p=boinc-v2.git;a=commit;h=328d87be2625ce68a04b0d1caf29e3826eab25f3

API: Fix a bug in boinc_get_opencl_ids() which incorrectly returned CL_INVALID_DEVICE on some systems. This would happen on hosts with multiple GPUs if:
* the host has GPUs from two different vendors with different OpenCL platforms (e.g., NVIDIA and AMD), and
* the second platform queried has more GPUs than the first one, and
* the requested GPU is a higher numbed one in the second platform.

Claggy

Thanks Claggy, I guess that the error is a bit confusing as I did not see this on BETA. It is only affecting a small number of tasks, but it is sporadic. Is there an eta on when the fix will be deployed to main? I did expect to see the other though as it was the first time the app ran on main.

INFO: can't open binary kernel file: D:\BOINC/projects/setiathome.berkeley.edu\AstroPulse_Kernels_r2721.cl_GeForceGTX750Ti.bin_V7_TWIN_FFA_34411, continue with recompile...
Info : Building Program (binary, clBuildProgram):main kernels: OK code 0
INFO: binary kernel file created
WARNING: can't open binary kernel file for oclFFT plan: D:\BOINC/projects/setiathome.berkeley.edu\AP_clFFTplan_GeForceGTX750Ti_32768_gr64_lr8_wg256_tw0_r2721.bin_34411, continue with recompile...
GPU device sync requested... ...GPU device synched

---

I don't buy computers, I build them!!

Is there an eta on when the fix will be deployed to main?

I'm guessing Lunatics Installer v0.43 will be released sometime within the next 48 hours (touch wood) - I've just fixed the last notified bug, and I'm planning to to test a release candidate later this evening after I've tidied up some documentation.

Tidying up the OpenCL detection will have to wait until Installer v0.43a - Raistmer doesn't want to delay things by deploying and testing the OpenCL fix that Claggy posted. It's cosmetic only as far as SETI is concerned, because the fallback internal detection corrects the mistake - but it's important that we do test it as soon at the AP v7 release pressure has died down, for the benefit of other BOINC projects that possibly don't have that internal fallback code.

Is there an eta on when the fix will be deployed to main?

I'm guessing Lunatics Installer v0.43 will be released sometime within the next 48 hours (touch wood) - I've just fixed the last notified bug, and I'm planning to to test a release candidate later this evening after I've tidied up some documentation.

Tidying up the OpenCL detection will have to wait until Installer v0.43a - Raistmer doesn't want to delay things by deploying and testing the OpenCL fix that Claggy posted. It's cosmetic only as far as SETI is concerned, because the fallback internal detection corrects the mistake - but it's important that we do test it as soon at the AP v7 release pressure has died down, for the benefit of other BOINC projects that possibly don't have that internal fallback code.

Thanks Richard.

---

I don't buy computers, I build them!!

Was just looking at an Inconclusive that showed up this morning, mainly because it was the first one under AP v7. It's WU 1613138028. What caught my eye was not just that my NVIDIA GPU identified 4 repeating pulses, while my wingman only found 1 with his ATI GPU. It's that the characteristics of his 1 repeating pulse seem to be very different from any of the 4 that my machine found (at least according to the Stderr). That makes me wonder if there's an NVIDIA vs. ATI issue here.

My Stderr shows:

    single pulses: 1
repetitive pulses: 4
  percent blanked: 0.00
Single pulse: peak_power=215.756 dm=-5293 fft_num=21659648 peak_bin=21666944 scale=7
Rep. pulse: num_std_devs=6.919 peak_power=3778.954 dm=-5552 peak_bin=3840 scale=4 ffa_scale=0 period=309.3236
Rep. pulse: num_std_devs=6.995 peak_power=2794.698 dm=9440 peak_bin=6320 scale=4 ffa_scale=0 period=427.5104
Rep. pulse: num_std_devs=6.964 peak_power=3832.457 dm=10064 peak_bin=4048 scale=4 ffa_scale=0 period=305.6033
Rep. pulse: num_std_devs=6.823 peak_power=6413.703 dm=-12800 peak_bin=1888 scale=4 ffa_scale=1 period=356.0778

Wingman's Stderr shows:

    single pulses: 1
repetitive pulses: 1
  percent blanked: 0.00
Single pulse: peak_power=215.8 dm=-5293 fft_num=21659648 peak_bin=21666944 scale=7
Rep. pulse: num_std_devs=7.3 peak_power=2586 dm=6128 peak_bin=2992 scale=4 ffa_scale=0 period=467.5833

Both of us are running stock, and the tiebreaker task has been sent to a stock ATI Mac, so it should be interesting to see how that turns out. In the meantime, any thoughts?

Was just looking at an Inconclusive that showed up this morning, mainly because it was the first one under AP v7. It's WU 1613138028. What caught my eye was not just that my NVIDIA GPU identified 4 repeating pulses, while my wingman only found 1 with his ATI GPU. It's that the characteristics of his 1 repeating pulse seem to be very different from any of the 4 that my machine found (at least according to the Stderr). That makes me wonder if there's an NVIDIA vs. ATI issue here.

My Stderr shows:

    single pulses: 1
repetitive pulses: 4
  percent blanked: 0.00
Single pulse: peak_power=215.756 dm=-5293 fft_num=21659648 peak_bin=21666944 scale=7
Rep. pulse: num_std_devs=6.919 peak_power=3778.954 dm=-5552 peak_bin=3840 scale=4 ffa_scale=0 period=309.3236
Rep. pulse: num_std_devs=6.995 peak_power=2794.698 dm=9440 peak_bin=6320 scale=4 ffa_scale=0 period=427.5104
Rep. pulse: num_std_devs=6.964 peak_power=3832.457 dm=10064 peak_bin=4048 scale=4 ffa_scale=0 period=305.6033
Rep. pulse: num_std_devs=6.823 peak_power=6413.703 dm=-12800 peak_bin=1888 scale=4 ffa_scale=1 period=356.0778

Wingman's Stderr shows:

    single pulses: 1
repetitive pulses: 1
  percent blanked: 0.00
Single pulse: peak_power=215.8 dm=-5293 fft_num=21659648 peak_bin=21666944 scale=7
Rep. pulse: num_std_devs=7.3 peak_power=2586 dm=6128 peak_bin=2992 scale=4 ffa_scale=0 period=467.5833

Both of us are running stock, and the tiebreaker task has been sent to a stock ATI Mac, so it should be interesting to see how that turns out. In the meantime, any thoughts?

I dont think so.

Your wingman has 15 invalids so far.
First check looks like an driver issue on his host.

---

With each crime and every kindness we birth our future.

Both of us are running stock, and the tiebreaker task has been sent to a stock ATI Mac, so it should be interesting to see how that turns out. In the meantime, any thoughts?

Host is running CAL 1.4.1546 which is from Cat 11.9 and uses SDK 2.5, and APP runtime 923.1 which is from Cat 12.4 and uses SDK 2.7, and has masses of inconclusives, Basically it's got mismatched CAL and APP drivers:

Validation inconclusive tasks for computer 6656238

Computer 6656238

Claggy

Okay, thanks guys! I did see that that he had one invalid AP v7, but he hasn't really run many of those yet. I just thought I raise a flag early in case there was anything to it. Those ATI driver issues and combinations are way beyond me!

Okay, thanks guys! I did see that that he had one invalid AP v7, but he hasn't really run many of those yet. I just thought I raise a flag early in case there was anything to it. Those ATI driver issues and combinations are way beyond me!

They sometimes seem to beyond ATI as well, so you're in good company.

Okay, thanks guys! I did see that that he had one invalid AP v7, but he hasn't really run many of those yet. I just thought I raise a flag early in case there was anything to it. Those ATI driver issues and combinations are way beyond me!

They sometimes seem to beyond ATI as well, so you're in good company.

Sorry not this one.
For sure user issue.

---

With each crime and every kindness we birth our future.

6) Example of "user error": too old ATi drivers in use:

http://setiathome.berkeley.edu/results.php?hostid=7320889

But now such host will produce clear invalids as computation errors and will not pollute database with "processed" but incorrect results. Also there is zero chance that 2 such results "validate" against each other, chance that always existed before.

<core_client_version>7.2.42</core_client_version>
<![CDATA[
<message>
too many exit(0)s
</message>
<stderr_txt>
Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns device 0
Info: BOINC provided OpenCL device ID used
ERROR: unsupported OpenCL runtime version: OpenCL 1.1 AMD-APP-SDK-v2.5 (775.2). Please update drivers! Exiting...
Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns device 0
Info: BOINC provided OpenCL device ID used
ERROR: unsupported OpenCL runtime version: OpenCL 1.1 AMD-APP-SDK-v2.5 (775.2). Please update drivers! Exiting...
Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns device 0
Info: BOINC provided OpenCL device ID used
ERROR: unsupported OpenCL runtime version: OpenCL 1.1 AMD-APP-SDK-v2.5 (775.2). Please update drivers! Exiting...
Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns device 0
Info: BOINC provided OpenCL device ID used
ERROR: unsupported OpenCL runtime version: OpenCL 1.1 AMD-APP-SDK-v2.5 (775.2). Please update drivers! Exiting...
Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns device 0
Info: BOINC provided OpenCL device ID used
ERROR: unsupported OpenCL runtime version: OpenCL 1.1 AMD-APP-SDK-v2.5 (775.2). Please update drivers! Exiting...
Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns device 0
Info: BOINC provided OpenCL device ID used
ERROR: unsupported OpenCL runtime version: OpenCL 1.1 AMD-APP-SDK-v2.5 (775.2). Please update drivers! Exiting...
Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns device 0
Info: BOINC provided OpenCL device ID used
ERROR: unsupported OpenCL runtime version: OpenCL 1.1 AMD-APP-SDK-v2.5 (775.2). Please update drivers! Exiting...
Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: Advanced Micro Devices, Inc.
BOINC assigns device 0
Info: BOINC provided OpenCL device ID used
ERROR: unsupported OpenCL runtime version: OpenCL 1.1 AMD-APP-SDK-v2.5 (775.2). Please update drivers! Exiting...

</stderr_txt>
]]>

7) can't get list of devices for some reason:
[2] AMD ATI Radeon HD 5x00 series (Redwood) (1024MB) driver: 1.4.1523 OpenCL: 1.1
http://setiathome.berkeley.edu/results.php?hostid=6189952
-32 code is "Invalid platform" error.

<core_client_version>7.0.25</core_client_version>
<![CDATA[
<message>
too many boinc_temporary_exit()s
</message>
<stderr_txt>
Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
WARNING: boinc_get_opencl_ids failed with code -1
Error: Getting Platforms. (clGetPlatformsIDs)
BOINC assigns device 0
WARNING: BOINC failed to provide OpenCL device, using own enumeration abilities
ERROR: OpenCL kernel/call 'clGetDeviceIDs (second call)' call failed (-32) in file ..\..\..\..\src\GPU_lock.cpp near line 812.
Waiting 30 sec before restart...
Running on device number: 1
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
WARNING: boinc_get_opencl_ids failed with code -1
Error: Getting Platforms. (clGetPlatformsIDs)
BOINC assigns device 1
WARNING: BOINC failed to provide OpenCL device, using own enumeration abilities
ERROR: OpenCL kernel/call 'clGetDeviceIDs (second call)' call failed (-32) in file ..\..\..\..\src\GPU_lock.cpp near line 812.
Waiting 30 sec before restart...
...............

7A) same as 7) but for iGPU: INTEL Intel(R) HD Graphics (1752MB) OpenCL: 1.2
http://setiathome.berkeley.edu/results.php?hostid=7344909