Issues and Resolutions
log in

Advanced search

Message boards : News : Issues and Resolutions

1 · 2 · 3 · Next
Author Message
Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2528
Credit: 740,580,099
RAC: 0
Message 18361 - Posted: 6 Feb 2014, 15:35:29 UTC

One third of the lost credits have been restored. The remaining credits will be restored of the new several days. If your account does not exist because it was lost between the last full backup and now, please re-create your account using the same login name and email address and send me a PM and I will manually restore your stats.

A Windows ATI (CAL/Brook+) app was successfully compiled under VS2012 and is currently being tested so AMD HD 2400 thru HD 47xx GPUs will be able to participate.

The OpenCL CPU apps are not suspending properly. The issue has been replicated and is under investigation.

A number of volunteers are returning results with the error: "couldn't start app: CreateProcess() failed - (unknown error)". This is due to a missing Microsoft C++ Runtime library required by the nVidia OpenCL DLLs. The MSVC Runtimes are available from Microsoft at
http://www.microsoft.com/en-us/download/details.aspx?id=30679 or from here for 32-bit or 64-bit hosts.

Dirk Broer
Send message
Joined: 20 Aug 10
Posts: 33
Credit: 206,442,830
RAC: 1,010,458
Message 18368 - Posted: 6 Feb 2014, 19:11:31 UTC - in response to Message 18361.
Last modified: 6 Feb 2014, 19:24:21 UTC

Using my HD 4770 right at this very moment -see hostid=36135-, just as able as the HD 48xx series, so OpenCL 1.0 capable (only far cheaper to run 24/7/365!)

AMD ATI Radeon HD 4700/4800 (RV740/RV770) (512MB) driver: 1.4.1734 OpenCL: 1.00
____________

MiKe
Send message
Joined: 3 Jan 14
Posts: 2
Credit: 275,427
RAC: 0
Message 18369 - Posted: 6 Feb 2014, 19:39:42 UTC

Hey there,

found a strange behavior of the CAL App on my machine.
All WUs of the first bunch I was calculating on finished after
5 secs / WU with zero credit granted.

Take a look at:
http://boinc.thesonntags.com/collatz/results.php?hostid=138689

After a quick search on the statistics page I was not able to find other computers using the CAL App at the moment, so I cannot tell you,
if it is a general problem or not.
Nethertheless, I thought I'd better inform you.

Cheers,

Mike

Kombizahl
Send message
Joined: 29 Sep 09
Posts: 12
Credit: 158,620,174
RAC: 5
Message 18371 - Posted: 6 Feb 2014, 20:03:42 UTC

I've today some compute errors with my ATI Cards.yesterday it works fine.
Greetings
____________
Greetings

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2528
Credit: 740,580,099
RAC: 0
Message 18372 - Posted: 6 Feb 2014, 20:33:39 UTC - in response to Message 18371.

I've today some compute errors with my ATI Cards.yesterday it works fine.
Greetings


Reboot. My guess is that the driver crashed and the data being returned now is garbage. But, it proves the additional validation in the v6 apps is working properly for amd_gpu apps.

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2528
Credit: 740,580,099
RAC: 0
Message 18374 - Posted: 6 Feb 2014, 20:42:38 UTC - in response to Message 18369.

Hey there,

found a strange behavior of the CAL App on my machine.
All WUs of the first bunch I was calculating on finished after
5 secs / WU with zero credit granted.

Take a look at:
http://boinc.thesonntags.com/collatz/results.php?hostid=138689

After a quick search on the statistics page I was not able to find other computers using the CAL App at the moment, so I cannot tell you,
if it is a general problem or not.
Nethertheless, I thought I'd better inform you.

Cheers,

Mike



I removed the ATI app after only a couple minutes because the internal validation isn't working. If you look at the runtime and the output, it is pretty obvious that they didn't succeed. The question now is why did they return as successful when the app verifies the output and also why does it work on several of my ATI GPUs but they are running the 13.12 driver. It looks like 12.6 reports errors. Were you able to run the Collatz 2.09 ATI apps OK with that driver?

MiKe
Send message
Joined: 3 Jan 14
Posts: 2
Credit: 275,427
RAC: 0
Message 18375 - Posted: 6 Feb 2014, 21:12:55 UTC - in response to Message 18374.

Hey there,

found a strange behavior of the CAL App on my machine.
All WUs of the first bunch I was calculating on finished after
5 secs / WU with zero credit granted.

Take a look at:
http://boinc.thesonntags.com/collatz/results.php?hostid=138689

After a quick search on the statistics page I was not able to find other computers using the CAL App at the moment, so I cannot tell you,
if it is a general problem or not.
Nethertheless, I thought I'd better inform you.

Cheers,

Mike



I removed the ATI app after only a couple minutes because the internal validation isn't working. If you look at the runtime and the output, it is pretty obvious that they didn't succeed. The question now is why did they return as successful when the app verifies the output and also why does it work on several of my ATI GPUs but they are running the 13.12 driver. It looks like 12.6 reports errors. Were you able to run the Collatz 2.09 ATI apps OK with that driver?


Before the big crash of the server I crunched the mini_collatz WUs with this driver (Catalyst 13.9, CAL Vers. 1.4.1734) quite successfully.
The solo_collatz WUs only supported OpenCL, which is not supported by my card.
However, I never had problems with this driver up to now.
Has something changed while switching the CAL apps from mini_collatz to the solo_collatz calculations, which could cause this weird behavior?
I could try an older or newer driver (since you are using 13.12), but up to now I have not thought about that, since everything worked fine.

Cheers,
Mike

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2528
Credit: 740,580,099
RAC: 0
Message 18382 - Posted: 6 Feb 2014, 23:29:59 UTC - in response to Message 18375.

Has something changed while switching the CAL apps from mini_collatz to the solo_collatz calculations, which could cause this weird behavior?


You mean other than compiling an app with the Brook+ library which is known to be buggy and 10 versions behind the current AMD APP SDK version and hasn't been supported for almost 4 years? :-) On top of that, the library Gipsel used to create them back in 2009 had additional changes he made that are not available in the Brook 1.4.1 beta version on Sourceforge. Yes, I would say the ATI apps will be a challenge to get working properly as a solo app version.

Joe
Send message
Joined: 11 Jan 10
Posts: 26
Credit: 2,299,168,333
RAC: 1,970,505
Message 18383 - Posted: 7 Feb 2014, 2:42:50 UTC - in response to Message 18361.

I'm not a tech guy I just crunch various projects for the points.

I use nvida cards and unfortunately since the crash they don't work.

I understand you would rather folks like me don't put an undue burden on your server but I'm curious if what worked before will ever work again....

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2528
Credit: 740,580,099
RAC: 0
Message 18386 - Posted: 7 Feb 2014, 4:03:13 UTC - in response to Message 18383.

I'm not a tech guy I just crunch various projects for the points.

I use nvida cards and unfortunately since the crash they don't work.

I understand you would rather folks like me don't put an undue burden on your server but I'm curious if what worked before will ever work again....


It will take me a minimum of two days to repartition a box, install the OS and all the drivers, install all the devleopment tools, compile an app and then copy it to another box I also have to set up which doesn't have all the developer tools installed to see if it works. Then, if it does, I pray it works with all 8 bazillion flavors of drivers, service packs, etc. that everyone has installed.

The app was tested. It ran fine on my boxes. It doesn't run on yours. I don't know why. But, it will take time to figure out. On the bright side, when the solo apps error out, at least no one else is waiting for a wingman to complete it in order to get credit. So, I appreciate that you continue to try and re-try my attempts to get it working properly. Eventually, something will work. That, or I'll create a v6 CUDA app.

mathew
Send message
Joined: 23 Oct 12
Posts: 3
Credit: 50,064,546
RAC: 0
Message 18389 - Posted: 7 Feb 2014, 10:47:23 UTC

All WUs are crashing with "computation error" after collatz is back online.
Nothing changed to my configuration. It's a Dual Xeon MP running SuSE Linux 12.2 x86 and 3 additional GT 430 with OpenCL support.
Before the crash happend everything worked fine and I did not change anything to my system. stderr does not tell anything too.

So any ideas folks?

Profile [AF>Le_Pommier>MacBidouille.com]Prof
Send message
Joined: 27 Dec 11
Posts: 22
Credit: 1,453,940,875
RAC: 235,118
Message 18391 - Posted: 7 Feb 2014, 14:15:12 UTC

Hi Slicker!!

I've the same problem as Zombie MM with my Mac under Mavericks and an AMD 7950!! (Another Mac under 10.8.5 and an ATI 5770 works fine!!)

The stderr message is as usual:

Stderr output

<core_client_version>7.2.33</core_client_version>
<![CDATA[
<message>
process exited with code 245 (0xf5, -11)
</message>
<stderr_txt>
Collatz Conjecture v6.04 x86_64 for OpenCL
Based on the AMD Brook+ kernels by Gipsel
Platform: ADVANCED MICRO DEVICES
Device: 16915456
OpenCL context created
OpenCL command queue created
OpenCL program created
BUILD LOG

clBuildProgram() failed (-11) Program build failure
Error -11. Processing Aborted.
14:35:47 (1248): called boinc_finish

</stderr_txt>
]]>


But I've found this in my "console log". Perhaps something could help you solve that issue!!??

07/02/2014 14:36:54,338 solo_collatz_6.04_x86_64-apple-darwin__opencl_amd_gpu[1439]: solo_collatz_6.04_x86_64-apple-darwin__opencl_amd_gpu(1439,0x7fff77256310) malloc: *** error for object 0x107000018: incorrect checksum for freed object - object was probably modified after being freed.
*** set a breakpoint in malloc_error_break to debug


I cross my fingers.... ;-DD

Majestix
Send message
Joined: 7 Jun 10
Posts: 7
Credit: 17,923,550
RAC: 0
Message 18397 - Posted: 8 Feb 2014, 18:52:09 UTC

Hello

same problem here

,,,
08.02.2014 19:46:31 | Collatz Conjecture | task solo_collatz_2369823732050544011631_824633720832_0 resumed by user
08.02.2014 19:46:32 | Collatz Conjecture | [error] Process creation failed: (unknown error)
08.02.2014 19:46:32 | Collatz Conjecture | [error] Process creation failed: (unknown error)
08.02.2014 19:46:32 | Collatz Conjecture | [error] Process creation failed: (unknown error)
08.02.2014 19:46:32 | Collatz Conjecture | [error] Process creation failed: (unknown error)
08.02.2014 19:46:32 | Collatz Conjecture | [error] Process creation failed: (unknown error)
08.02.2014 19:46:35 | Collatz Conjecture | Computation for task solo_collatz_2369823732050544011631_824633720832_0 finished
08.02.2014 19:46:35 | Collatz Conjecture | Output file solo_collatz_2369823732050544011631_824633720832_0_0 for task solo_collatz_2369823732050544011631_824633720832_0 absent
...

OS: XP 32bit, fully patched
GPU NVIDIA GeForce 550GTX, Driver 331.21,
M$ runtimes reinstalled
project reset

any clues

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2528
Credit: 740,580,099
RAC: 0
Message 18398 - Posted: 8 Feb 2014, 21:07:47 UTC - in response to Message 18389.

All WUs are crashing with "computation error" after collatz is back online.
Nothing changed to my configuration. It's a Dual Xeon MP running SuSE Linux 12.2 x86 and 3 additional GT 430 with OpenCL support.
Before the crash happend everything worked fine and I did not change anything to my system. stderr does not tell anything too.

So any ideas folks?


Version 6.06 has just been released for Linux i686 which I hope will fix the error. I was able to reproduce the error on an Ubuntu 12.04 LTS 32-bit machine with an nVidia 9800 GTX+ GPU. Version 6.06 works OK on that machine so I'm hoping that it isn't a fluke.

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2528
Credit: 740,580,099
RAC: 0
Message 18399 - Posted: 9 Feb 2014, 2:03:42 UTC - in response to Message 18397.

Hello

same problem here

,,,
08.02.2014 19:46:31 | Collatz Conjecture | task solo_collatz_2369823732050544011631_824633720832_0 resumed by user
08.02.2014 19:46:32 | Collatz Conjecture | [error] Process creation failed: (unknown error)
08.02.2014 19:46:32 | Collatz Conjecture | [error] Process creation failed: (unknown error)
08.02.2014 19:46:32 | Collatz Conjecture | [error] Process creation failed: (unknown error)
08.02.2014 19:46:32 | Collatz Conjecture | [error] Process creation failed: (unknown error)
08.02.2014 19:46:32 | Collatz Conjecture | [error] Process creation failed: (unknown error)
08.02.2014 19:46:35 | Collatz Conjecture | Computation for task solo_collatz_2369823732050544011631_824633720832_0 finished
08.02.2014 19:46:35 | Collatz Conjecture | Output file solo_collatz_2369823732050544011631_824633720832_0_0 for task solo_collatz_2369823732050544011631_824633720832_0 absent
...

OS: XP 32bit, fully patched
GPU NVIDIA GeForce 550GTX, Driver 331.21,
M$ runtimes reinstalled
project reset

any clues



I cannot duplicate the error, at least not on a box running Win XP SP2. Chances are that there is a missing DLL and/or a DLL that is the wrong version. Running depends.exe from http://www.dependencywalker.com/ will list all the DLLs that the Collatz application depends upon and which, if any, are missing. Any chance you could run that and then post which DLL is reported as missing? (any DLL shown in yellow or red generally means it is missing or the wrong version)

Profile Zydor
Avatar
Send message
Joined: 19 Aug 09
Posts: 364
Credit: 840,811,292
RAC: 0
Message 18400 - Posted: 9 Feb 2014, 4:08:10 UTC
Last modified: 9 Feb 2014, 4:36:17 UTC

On mine it claims following are missing:

API-MS-WIN-APPMODEL-RUNTIME-L1-1-0.DLL
API-MS-WIN-CORE-WINRT-ERROR-L1-1-0.DLL
API-MS-WIN-CORE-WINRT-L1-1-0.DLL
API-MS-WIN-CORE-WINRT-ROBUFFER-L1-1-0.DLL
API-MS-WIN-CORE-WINRT-STRING-L1-1-0.DLL
API-MS-WIN-SHCORE-SCALING-L1-1-1.DLL
DCOMP.DLL
IESHIMS.DLL

Error message states (All error messages for each entry coloured in red):

Error opening file. The system cannot find the file specified (2).

Two warning messages shown, coloured in red:

Warning: At least one delay-load dependency module was not found.
Warning: At least one module has an unresolved import due to a missing export function in a delay-load dependent module.


Tree Root in the application was:
SOLO_COLLATZ_6.03_WINDOWS_X86_64__OPENCL_AMD_GPU.EXE

Majestix
Send message
Joined: 7 Jun 10
Posts: 7
Credit: 17,923,550
RAC: 0
Message 18401 - Posted: 9 Feb 2014, 8:04:49 UTC - in response to Message 18399.
Last modified: 9 Feb 2014, 8:05:19 UTC

Hello,


? IESHIMS.DLL Error .....
? WER.DLL Error .....

MPR.DLL

Warning: At least one delay-load dependency module was not found.
Warning: At least one module has an unresolved import due to a missing export function in a delay-load dependent module.

Profile Slicker
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 11 Jun 09
Posts: 2528
Credit: 740,580,099
RAC: 0
Message 18406 - Posted: 9 Feb 2014, 14:48:46 UTC - in response to Message 18401.

Hello,


? IESHIMS.DLL Error .....
? WER.DLL Error .....

MPR.DLL

Warning: At least one delay-load dependency module was not found.
Warning: At least one module has an unresolved import due to a missing export function in a delay-load dependent module.


MPR.DLL (multiple provider router) is for Windows Networking but is also used for functions such as WNetGetUser to get the logged on user's name. There is nothing in the Collatz app that calls any of those functions.

From StackOverflow:
ieshims.dll is an artifact of Vista/7 where a shim DLL is used to proxy certain calls (such as CreateProcess) to handle protected mode IE, which doesn't exist on XP, so it is unnecessary. wer.dll is related to Windows Error Reporting and again is probably unused on Windows XP which has a slightly different error reporting system than Vista and above.

I would say you shouldn't need either of them to be present on XP and would normally be delay loaded anyway.


In other words, those missing DLLs shouldn't matter. Obviously, something isn't working though. I guess I'll have to create an XP development machine and compile the app on it. When your operating system is older than the bourbon or scotch you drink, it is time to consider an upgrade. ;-)

pioneer1
Send message
Joined: 24 Oct 09
Posts: 27
Credit: 60,304,644
RAC: 1
Message 18407 - Posted: 9 Feb 2014, 14:55:27 UTC

IESHIMS.DLL Error opening the file. The system cannot find the file specified(2)
WER.DLL Error opening the file. The system cannot find the file specified(2)

Warning: At least one delay-load dependency module was not found.

On Win2K3

Majestix
Send message
Joined: 7 Jun 10
Posts: 7
Credit: 17,923,550
RAC: 0
Message 18408 - Posted: 9 Feb 2014, 15:00:28 UTC - in response to Message 18406.

Hello again,

MPR.DLL should not be the problem (was a copy n paste error)

regard

1 · 2 · 3 · Next
Post to thread

Message boards : News : Issues and Resolutions


Main page · Your account · Message boards


Copyright © 2018 Jon Sonntag; All rights reserved.