Can't seem to DL anything for my windows machine.

Must have been a quick short run.

Must have been a quick short run.

They don't to actual calculations (for that we have to wait for the nontrivial psi4 port), but they do test the distribution system, dependencies, and so on.

Waiting for more windows consistent queue....

Hi, something new? My 88 thread server is hungry for some native Win work. If want to help with testing, can let me know.

I think the team is busy with Windows GPU application issues, see here for more details: https://www.gpugrid.net/forum_thread.php?id=4802

https://www.gpugrid.net/forum_thread.php?id=4802

https://www.gpugrid.net/forum_thread.php?id=4802

Also, several members of staff are on vacation at the moment, so progress will be slow until they return.

Is there any windows CPU WU's, I set everything to be enable but I get no tasks.

I got a few beta but they failed, looks like it was installing Python and minconda while trying to run another WU at the same time.

hopefully next round is good.

I see the anaconda option in the start menu, but non of the task actually worked.

if I click it it does not work as path is wrong it has

C:\%windir%\System32\cmd.exe "/K" C:\ProgramData\BOINC\projects\www.gpugrid.net\miniconda\Scripts\activate.bat C:\ProgramData\BOINC\projects\www.gpugrid.net\miniconda

but %windir% = C:\windows so it actually resolves to c:\c\windows.

I see that it trys to update conda version but it didn't work as the reported version is still 4.5.4

I see these 2 errors:

Traceback (most recent call last):
File "C:\ProgramData\BOINC\slots\3\qmml3\lib\site-packages\psi4\__init__.py", line 55, in <module>
from . import core
ImportError: DLL load failed: The specified module could not be found.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "run.py", line 13, in <module>
import psi4
File "C:\ProgramData\BOINC\slots\3\qmml3\lib\site-packages\psi4\__init__.py", line 60, in <module>
raise ImportError("{0}".format(err))

I can see the deepdiff, jsonpickle, psi4 in the pkgs folder.

this was most common causes of error. one was

C:\ProgramData\BOINC\projects\www.gpugrid.net\miniconda\Scripts\conda install -m -y -p C:\ProgramData\BOINC\slots\1\qmml3 --override-channels -c gpugridbeta -c defaults --file requirements.txt
) 9>C:\ProgramData\BOINC\projects\www.gpugrid.net\miniconda.lock
The process cannot access the file because it is being used by another process.
17:08:33 (8180): install_miniconda.bat exited; CPU time 0.046875
17:08:33 (8180): app exit status: 0x1
17:08:33 (8180): called boinc_finish(195)

Seems like they are working now. I see 100% cpu load, running 4x4cores.

Didn't notice any :)

3 UT validated on i7-3770 W10 host ! :-)

After few minutes, CPU load is OK: see the picture !

Just out of curiosity, my Linux PC has been receiving Quantum Chemistry, beta test v3.32 (mt). PC number 258077. Is that supposed to happen? I was under the impression that the beta test tasks were for windows only.

Just out of curiosity, my Linux PC has been receiving Quantum Chemistry, beta test v3.32 (mt). PC number 258077. Is that supposed to happen? I was under the impression that the beta test tasks were for windows only.

Is there any drivers in the package? I had some stability issues with my computer, which I never saw before. I had one bug check that indicated python as culprit, I doubt it but just checking.

I just picked up some of the beta test 3.33 tasks on two different Windows machines. After 10+ minutes, it looks like the tasks are only using one thread.

After 15+ minutes, one of the machines got this error message:

Debug Error!
Program: C:\ProgramData\BOINC\slots\0\qmml3\pythin.exe
abort() has been called
(Press Retry to debug the application)

I think I have the same problem on my hostid 193594

I try to increase disk usage.

Toni asked:

The root error is here

PSIO_ERROR: unit = 97, errval = 10
PSIO_ERROR: 10 (lseek failed)

your machine appears to be the only one with this specific failure. Do you have space in your HDD, or other uncommon setups (e.g. FAT instead of NTFS)?

I see the same on 2 computers, one has 238GB free with boinc allowed to use upto 218GB.


The other has 366GB free with same settings.

I clicked ignore and it seemed to do something, but not full CPU load.

I continue to download wu for linux (QC and Beta). Nothing for windows.
Do i have to change something in my profile?

NOTE: This should be in the Linux section, not Windows, so you can move it, though I expect it would apply to either.

This one failed with "196 (0xc4) EXIT_DISK_LIMIT_EXCEEDED".
http://www.gpugrid.net/result.php?resultid=18977559

The BointTasks message is:

Aborting task 1718_36_33_32_50_14f550fc_n00001-SDOERR_SELE6-0-1-RND8308_2: exceeded disk limit: 61716.49MB > 57220.46MB

any progress with the planned QC app for Windows?

If I look at the Project Status page, it currently shows the following:

Long runs (= GPU tasks) unsent: 1,550; in progress: 3,634; users: 1,002
Quantum Chemistry unsent: 179,513; in progress: 550; users: 71

this shows a vast imbalance, the reason for which is that QC CPU tasks are still available for Linux users only.

this shows a vast imbalance, the reason for which is that QC CPU tasks are still available for Linux users only.

I am running SuSE Leap 42.3 on my main Linux box, and Leap 15.0 on my HP laptop. The kernel of Leap 15.0 is more advanced and is equal to that of the Enterprise version of SuSE Linux, SLES.It is also being updated regularly, while updates of 42.3 seem finished. But I have most of my tools on it, including Kaffeine, and it runs GPUGRID tasks, both CPU and GPU since the box has a GTX 750 Ti GPU board.
Tullio
____________

Do you have a task number to check?

this shows a vast imbalance, the reason for which is that QC CPU tasks are still available for Linux users only.

And not all distro are welcome.
I tried some major distro and i have some problems (for example with Fedora).
No problems with "old" Mint

After installing nVidia driver 411.70 on my Windows 10 PC and a Windows 1809 upgrade, my first GPU task completed and validated on it. So far GPU tasks were validated only on my Linux box with SuSE Leap 42.3 and a GTX 750 Ti.
Tullio
____________

After installing nVidia driver 411.70 on my Windows 10 PC and a Windows 1809 upgrade ...

Thanks. What distro is it? Is it updated (i.e. security patches?). The other problem seems to be that the failure was not seen as such, and thus restarted.

No, I installed it via Geforce. Now GPUGRID tasks run successfully and also Einstein@home GPU tasks.
Tullio

any progress in developing and testing QC tasks for Windows ?

Fedora 28 with security update.
Now i return to Mint....

The same thing happens to me on SELE6 tasks.
Tullio

any progress in developing and testing QC tasks for Windows ?

No sorry, for this week and maybe the next we have other priorities. But we will definitely come back to it soon.

And also this week has gone....

Their priorities are most likely scientific. There is the side we don't see, the biology/chemistry side. The computation is only half of the puzzle.

Maybe now that IBM has bought RedHat there will be more Linux users.
Tullio

Stefan wrote on October 10:

No sorry, for this week and maybe the next we have other priorities. But we will definitely come back to it soon.

Stefan, thanks for the quick information :-)

I would say a good estimate is from January considering we will need to make a sprint to finish everything before holidays.

I would say a good estimate is from January considering we will need to make a sprint to finish everything before holidays.

We are ready! :-P

I received the second batch of Quantum Chemistry, beta test v3.33 (mt) for Windows today, except for 2 errors (so far), they are finishing successfully.

http://www.gpugrid.net/results.php?hostid=263612&offset=0&show_names=0&state=0&appid=35


I also get this message in the Stderr output:

==> WARNING: A newer version of conda exists. <==
current version: 4.5.4
latest version: 4.5.12

http://www.gpugrid.net/result.php?resultid=20150516

Got the following error message for a Windows QC_beta task:

16:34:31 (7484): wrapper: running .\qmml3\python.exe (run.py)
Detected memory leaks!
Dumping objects ->
..\api\boinc_api.cpp(309) : {1788} normal block at 0x0000022AE7E17A40, 8 bytes long.
Data: < * > 00 00 DC E7 2A 02 00 00
..\lib\diagnostics_win.cpp(417) : {555} normal block at 0x0000022AE7E40000, 1080 bytes long.
Data: < > 98 14 00 00 CD CD CD CD E4 01 00 00 00 00 00 00
..\zip\boinc_zip.cpp(122) : {294} normal block at 0x0000022AE7DFF750, 260 bytes long.
Data: < > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
{280} normal block at 0x0000022AE7DF99D0, 55 bytes long.
Data: < o > 01 00 00 00 6F 00 CD CD 00 00 00 00 00 00 00 00
{275} normal block at 0x0000022AE7E1BCE0, 44 bytes long.
Data: < * > 01 00 00 00 00 00 CD CD 01 BD E1 E7 2A 02 00 00
{270} normal block at 0x0000022AE7E1BE30, 43 bytes long.
Data: < Q * > 01 00 00 00 00 00 CD CD 51 BE E1 E7 2A 02 00 00
{265} normal block at 0x0000022AE7E1BC00, 44 bytes long.
Data: < ! * > 01 00 00 00 00 00 CD CD 21 BC E1 E7 2A 02 00 00
{260} normal block at 0x0000022AE7E1C1B0, 44 bytes long.
Data: < * > 01 00 00 00 00 00 CD CD D1 C1 E1 E7 2A 02 00 00
Object dump complete.

</stderr_txt>
]]>

I received the second batch of Quantum Chemistry, beta test v3.33 (mt) for Windows today, except for 2 errors (so far), they are finishing successfully.

http://www.gpugrid.net/results.php?hostid=263612&offset=0&show_names=0&state=0&appid=35


I also get this message in the Stderr output:

==> WARNING: A newer version of conda exists. <==
current version: 4.5.4
latest version: 4.5.12

http://www.gpugrid.net/result.php?resultid=20150516

My first Windows wu:

195 (0xc3) EXIT_CHILD_FAILED

11:18:52 (12032): install_miniconda.bat exited; CPU time 99.796875
11:18:52 (12032): wrapper: running .\qmml3\python.exe (run.py)
PSIO_ERROR: unit = 97, errval = 12
PSIO_ERROR: 12 (error writing to file)
11:44:45 (12032): .\qmml3\python.exe exited; CPU time 4998.312500
11:44:45 (12032): app exit status: 0xc0000409
11:44:45 (12032): called boinc_finish(195)
0 bytes in 0 Free Blocks.
396 bytes in 4 Normal Blocks.
1144 bytes in 1 CRT Blocks.
0 bytes in 0 Ignore Blocks.
0 bytes in 0 Client Blocks.
Largest number used: 0 bytes.
Total allocations: 22092877 bytes.

My first Windows wu:

195 (0xc3) EXIT_CHILD_FAILED

11:18:52 (12032): install_miniconda.bat exited; CPU time 99.796875
11:18:52 (12032): wrapper: running .\qmml3\python.exe (run.py)
PSIO_ERROR: unit = 97, errval = 12
PSIO_ERROR: 12 (error writing to file)
11:44:45 (12032): .\qmml3\python.exe exited; CPU time 4998.312500
11:44:45 (12032): app exit status: 0xc0000409
11:44:45 (12032): called boinc_finish(195)
0 bytes in 0 Free Blocks.
396 bytes in 4 Normal Blocks.
1144 bytes in 1 CRT Blocks.
0 bytes in 0 Ignore Blocks.
0 bytes in 0 Client Blocks.
Largest number used: 0 bytes.
Total allocations: 22092877 bytes.

I am still getting the 196 (0xc4) EXIT_DISK_LIMIT_EXCEEDED errors even with disc usage space for boincs set at 500 gigs.

Just now I tried to download QC tasks and/or QC beta tasks for windows.

However, none are available (according to a notice in the BOINCS Manager).

So all the "unsent" tasks shown on the Server Status Page are still for Linux only?
I thought that there are some available now for Windows.

Take 3 for test.
If I understand correctly, calculation progress for these WUs does not display correctly. How much time in average this WUs consuming?

P.S. Such WUs producing quite heavy load on storage subsystem. When 3 of them simultaneously writing data on disk, this generates a stream of almost 1 Gigabyte! o_O (~ 900 MiB/ps).
Without high performance SSD, it is better not to run more than one such WU : )


Upd.

One of them already died -_-
https://www.gpugrid.net/result.php?resultid=20269815

So all the "unsent" tasks shown on the Server Status Page are still for Linux only?

Use CPU - yes
Run test applications? - yes
Quantum Chemistry (CPU, beta): yes

What makes me wonder though is that after a few minutes the progress percentage in the BOING Manager got stuck at 1.098%, although the time is progressing, and the Windows task manager shows that Python is running (and using about 214MB RAM).
What I just noticed is that whereas under "remaining time", before some 6 hours where shown; now, about 13 minutes later, the value for "remaining time" has gone up to about 18 hours.
Is this okay, or is the tasks faulty?

the task failed with the error
196 (0xc4) EXIT_DISK_LIMIT_EXCEEDED
after 4.461 secs.

http://gpugrid.net/result.php?resultid=20291478

Another taks is running, so let's see.

Is this okay, or is the tasks faulty?

the task failed with the error
196 (0xc4) EXIT_DISK_LIMIT_EXCEEDED

105 errors / failed on 188 calculated WU's
I understand this is a BETA application.
Hope this helps.

so far, all 3 tasks which got downloaded failed with this exit_disk_limit_exceeded error after about 3900 - 4800 seconds.

I do understand that these tasks are beta, so one cannot expect that they succeed.
Maybe I am lucky though and one of these tasks will be successful eventually.

As a result, from 30 tasks taken only 14 successfully completed.
There are no more test tasks, so I continue crunching WCG.

Since I had started with these tasks yesterday, 20 were processed so far, out of which only 2 succeeded.
Hence, I now stopped downloading them, it doesn't make any sense.

What I don't understand is:

13.01.2019 07:46:22 | GPUGRID | Aborting task 194r20-TONI_TST10-0-1-RND3322_3: exceeded disk limit: 41204.15MB > 28610.23MB

so why is there not sufficient disk limit set by the project to begin with ???

Since I had started with these tasks yesterday, 20 were processed so far, out of which only 2 succeeded.
Hence, I now stopped downloading them, it doesn't make any sense.

What I don't understand is:

13.01.2019 07:46:22 | GPUGRID | Aborting task 194r20-TONI_TST10-0-1-RND3322_3: exceeded disk limit: 41204.15MB > 28610.23MB

so why is there not sufficient disk limit set by the project to begin with ???

The trick to this, is to find which file contains the limit and change it to a larger number and this task would have finished successfully.

The trick to this, is to find which file contains the limit and change it to a larger number and this task would have finished successfully.

any advice how to get this accomplished ?

Give it try, if you like.

Something similar to this, was done in the past for oversize output files. It worked.

Give it try, if you like.

Something similar to this, was done in the past for oversize output files. It worked.

I prefer a bugfixed version of the app.

I will raise the limit for the next batch ...

half an hour ago, I got a few tasks downloaded. So let's see whether they will succeed this time, or fail again.

However, what I notice is although my settings in the app_config.xml are for 2-core (which is shown in brackets in the BOINC manager), the Windows task-manager shows that only 1 core is being used by Python. Any explanation for this discrepancy?

the first task from today got finished successfully after some 4.250 secs, which is very nice :-)
http://gpugrid.net/result.php?resultid=20386975

Right thereafter, the next one got started, and again, although it should be 2-core, only one core is used.

Does anyone else make the same experience?

Toni wrote

Approx. 30% of the WUs are outside the disk limits as it was noted. I will raise the limit for the next batch. For the time being, just let them fail.

Toni wrote

Approx. 30% of the WUs are outside the disk limits as it was noted. I will raise the limit for the next batch. For the time being, just let them fail.

1 of 4 tasks which I downloaded and crunched this afternoon showed the
"196 (0xc4) EXIT_DISK_LIMIT_EXCEEDED" error
again. How come?

http://gpugrid.net/result.php?resultid=20388212

Toni wrote

Approx. 30% of the WUs are outside the disk limits as it was noted. I will raise the limit for the next batch. For the time being, just let them fail.

1 of 4 tasks which I downloaded and crunched this afternoon showed the
"196 (0xc4) EXIT_DISK_LIMIT_EXCEEDED" error
again. How come?

http://gpugrid.net/result.php?resultid=20388212

That's because the WUs are from the current batch, not the next batch. You lucked out and got 3 good ones.

Toni wrote

Approx. 30% of the WUs are outside the disk limits as it was noted. I will raise the limit for the next batch. For the time being, just let them fail.

1 of 4 tasks which I downloaded and crunched this afternoon showed the
"196 (0xc4) EXIT_DISK_LIMIT_EXCEEDED" error
again. How come?

http://gpugrid.net/result.php?resultid=20388212

That's because the WUs are from the current batch, not the next batch. You lucked out and got 3 good ones.

yesterday, another 3 tasks were downloaded, all of them failed. Total wasted CPU time about 20.000 secs.
Obviously, all these tasks are still from the current faulty batch.

Question for Toni: Why are these tasks not withdrawn? And when will the next batch with the increased disk limit be sent out?
I hate to waste that much of my CPU work.

okay, I understand.

So, the question (for me) now is: how many more WUs are in the current batch?
Will they be sent out during one more week, one more month ... ?

And when will the new batch be launched?

As said before: I do understand that this is beta, and WUs can fail.
But knowing that due to the too law disk limit setting in the current batch most of the WUs are bound to fail, it doesn't make a whole lot of sense to me to download them and waste quite some CPU time.

Although no "unsent" tasks were shown in the Project Status Page this morning, my PC downloaded 3 WUs.
Again, all 3 failed from the disk limit error, after a total runtime of about 19.000 secs.
Failure rate: 100% :-(((

Hence, it would really be great to know for how much longer these faulty WUs from the "old" batch will be sent out, and when the new batch will be published.

a few minutes ago, the fourth WU within short time failed.
I am giving up now and switched them off in the settings. It's a waste of CPU time.

Perhaps Toni or someone else from the GPUGRID team could post a short message when the new batch which is supposed not to contain this disk limit error will be available.

Perhaps Toni or someone else from the GPUGRID team could post a short message when the new batch which is supposed not to contain this disk limit error will be available.

Perhaps Toni or someone else from the GPUGRID team could post a short message when the new batch which is supposed not to contain this disk limit error will be available.

It seems there is no hurry...

... Raimondas and Toni might finish up the app. According to their latest news the app works mostly fine now except lacking some optimization so it should not be too far off.
Patience please, we will make it.

My guess is that many crunchers will use it.

I think the Quantum Chem app should be a subproject and then there should be badges. If you offer badges, they will come.

thanks for the thorough information, this helps us to understand what's going on.

TN-Grid, for example, had a lot of help from two volunteers that helped the team to introduce the sse/avx extensions

If you ever visit Rosetta, you will quickly find out that rjs5 knows a whole lot about extensions and parallelism;

If you put out a call, they will come.

I just got another test WU - and again, it failed with the
"196 (0xc4) EXIT_DISK_LIMIT_EXCEEDED" error:
http://gpugrid.net/result.php?resultid=20424952

So, obviously the disk limit was not raised enough :-(

Actually the current beta QC version should not be too different from the final one. One common reason for failure is that the computations need huge temporary files. We can't make them need less space, but the disk limit will be raised (this occurred for the linux app too).

I installed 4TB HDD into 3 machines in anticipation of these tasks. Would like to see how they run.

How much disk space do the QC tasks actually need, on the average?

I've got 2 linux machines running QC tasks.

My 40 thread machine is running 10 tasks at once with 4 threads per task. According to boinc, GPUGrid is using 18.8 GB of disk space.

My other machine is running 4 tasks simultaneously at 4 threads per task and that one is taking 9.19 GB of disk space.

The linux app (3.31) is running very smoothly these days. Kudos to Toni et al for fixing the bugs of the earlier versions.

@[VENETO] boboviz: The code is open source https://github.com/psi4/psi4/

About a month ago, Stefan wrote:

We have stated this in the past but I'll explain it here again. GPUGRID is related to the lab. In the lab we don't have a C++ developer, they are all chemists, biologists and statisticians. Best bet would be me as a computer scientist but I've been busy with many other things recently. So the development of the Windows app has fallen on Raimondas who is very nice to do this as he works in a company (Acellera) which obviously has clients to take care of and does not have an immediate gain from making this work on GPUGRID and Toni who tests it out of personal interest. The current paper we are working on takes priority so once it is finished which is supposed to be this week Raimondas and Toni might finish up the app. According to their latest news the app works mostly fine now except lacking some optimization so it should not be too far off.
Patience please, we will make it.

I highly doubt anything has changed as they still don't have a C++ developer. I would just lay low and let it happen when it happens.

I highly doubt anything has changed as they still don't have a C++ developer.

Got this error message on the current batch of QC test for Windows:

3/28/2019 7:47:29 AM | GPUGRID | Aborting task 1445r30-TONI_TST11-0-1-RND5643_0: exceeded disk limit: 44800.51MB > 28610.23MB

And another:

3/28/2019 9:07:38 AM | GPUGRID | Aborting task 1445r7-TONI_TST11-0-1-RND1466_0: exceeded disk limit: 44800.51MB > 28610.23MB

Did a project reset and started three more tasks. Two task finished successfully and the other aborted with:

3/28/2019 10:37:48 AM | GPUGRID | Aborting task 170r5-TONI_TST11-0-1-RND6393_2: exceeded disk limit: 34955.30MB > 28610.23MB

<core_client_version>7.14.2</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code 195 (0xc3)</message>
<stderr_txt>
10:28:17 (15156): wrapper (7.9.26016): starting
10:28:17 (15156): wrapper: running install_miniconda.bat ()
10:28:18 (15156): install_miniconda.bat exited; CPU time 0.000000
10:28:18 (15156): app exit status: 0x1
10:28:18 (15156): called boinc_finish(195)

Can you reset the project?

17:51:36 (6372): install_miniconda.bat exited; CPU time 40.357459
17:51:36 (6372): wrapper: running .\qmml3\python.exe (run.py)
Traceback (most recent call last):
File "C:\ProgramData\BOINC\slots\0\qmml3\lib\site-packages\psi4\__init__.py", line 55, in <module>
from . import core
ImportError: DLL load failed with error code -1073741795

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "run.py", line 14, in <module>
import psi4
File "C:\ProgramData\BOINC\slots\0\qmml3\lib\site-packages\psi4\__init__.py", line 60, in <module>
raise ImportError("{0}".format(err))
ImportError: DLL load failed with error code -1073741795
17:51:37 (6372): .\qmml3\python.exe exited; CPU time 0.000000
17:51:37 (6372): app exit status: 0x1
17:51:37 (6372): called boinc_finish(195)

I try on a new machine (Intel Xeon and Windows 7)
All errors:

On my windows 7 computer, I am getting "195 (0xc3) EXIT_CHILD_FAILED" on all the beta WUs.

http://www.gpugrid.net/result.php?resultid=20741510


On my windows 10 computer,I some WUs succeed, while other have the "196 (0xc4) EXIT_DISK_LIMIT_EXCEEDED" error.

http://www.gpugrid.net/results.php?hostid=263612&offset=0&show_names=0&state=0&appid=35

Another common error is disk space. It's because some molecules we process are indeed too large, and we don't have prior indication on when this occurs.