Windows 7 64-bit, Nvidia Drv v190, BOINC single user install.
- Rom
-Original Message-
From: Raistmer [mailto:raist...@mail.ru]
Sent: Saturday, October 17, 2009 2:33 AM
To: Rom Walton; David Anderson (BOINC)
Cc: boinc_dev@ssl.berkeley.edu
Subject: Re: [boinc_dev] BOINC 6.10.14 re mis
What OS & BOINC setup did you use?
Remote desktop not always lead to GPU hiding.
For example, no problem with RDP if OS is Windows 2003 server/XP and BOINC
installed as service.
While standalone run of CUDA app under user account will fail in such case,
app running under BOINC (running as service
I just tried to simulate this by using Remote Desktop, and BOINC was
still able to correctly detect the memory information for the Nvidia
GPU.
- Rom
-Original Message-
From: boinc_dev-boun...@ssl.berkeley.edu
[mailto:boinc_dev-boun...@ssl.berkeley.edu] On Behalf Of Raistmer
Sent: Frid
>Okay, instead of calling any of the CRT functions you can do this:
>#ifdef _WIN32
>TerminateProcess(GetCurrentProcess(), 0);
>#endif
>That'll prevent any form or error processing from the client after that
>piece of code is hit.
>- Rom
This solved my problem completely, thanks!
Now no
Don't know if it will work always but in case of "disappeared" GPU
cuMemGetInfo(&free,&total);
returns zero in both fields. Zero in total memory could be sign of error...
- Original Message -
From: "David Anderson"
To: "Raistmer"
Cc:
Sent: Saturday, October 17, 2009 1:37 AM
Subject:
That's what we're doing.
We have this working for ATI: the CAL init function fails
if the GPU is not usable.
We haven't figured how to tell if an NVIDIA GPU is still working;
none of the obvious things works.
We're waiting to hear from NVIDIA on this.
-- David
Raistmer wrote:
> If you really wan
"
- client: on startup, if a coproc needed by a job is missing, set a
"coproc_missing" flag rather than aborting the job. If use removes a GPU
board while there's a large queue of GPU jobs, they'll stay queued (until
their deadline passes).
Note: this doesn't fix the situation where user connec
Changing the app type from windows to console will cause a console
window to be displayed on Win9x which cannot be hidden by BOINC (which
probably isn't relevant anymore).
Okay, instead of calling any of the CRT functions you can do this:
#ifdef _WIN32
TerminateProcess(GetCurrentProcess(), 0)
1) _exit(0) still summons BOINC's debugger + Vista's dialog box.
2) no user intervention required in case of driver restart. Actually, I test
exception handling by running 3D-intensive game - it survives 3 or 4 driver
restarts w/o any problems already :)
But all CAL applications running at time o
Does the user have to restart the machine to recover from the
video-driver restart issues? Or will just re-launching the application
right after the crash help?
- Rom
-Original Message-
From: Raistmer [mailto:raist...@mail.ru]
Sent: Friday, October 16, 2009 4:43 PM
To: Rom Walton; bo
Agreed if "run into the same problem" is the case.
Problem of Vista's video-driver restart has not so permanent and
deterministic character.
It maybe days w/o such event so saving many-hours task from such incident
may be not so bad idea.
- Original Message -
From: "Rom Walton"
To: "Ra
El Vie 16 Oct 2009 05:39:08 Francisco Sanz García escribió:
> Hi Bernd.
>
> I modified your first shell script:
>
> #!/bin/sh
> LD_LIBRARY_PATH=../../projects/test.test/
> echo $LD_LIBRARY_PATH > /tmp/ldlibr.log
> echo $PWD >> /tmp/ldlibr.log
> ../../projects/test.test/hello_cuda2 "$@"
Try using _exit(0);
Although, to be perfectly honest you are going to be beating yourself up
later over this. For one reason or another app is causing a hardware
based exception to be thrown, setting things up so that BOINC will
re-launch the application later just so it can run into the same pro
Yes, it helped indeed, thanks!
Now I see next situation:
driver crash -> access to zero address exception -> app handler called ->
exit(0) called inside app exception handler and then exception re-catched by
BOINC again -> dump message in stderr and then windows (Vista) dilog box
appears with p
It'll convert SEH exceptions into C++ exceptions. You'll be able to use
the normal try/catch blocks to catch them.
- Rom
-Original Message-
From: boinc_dev-boun...@ssl.berkeley.edu
[mailto:boinc_dev-boun...@ssl.berkeley.edu] On Behalf Of Rom Walton
Sent: Friday, October 16, 2009 1:53
Go to Project Properties/C/C++/Code Generation section and change the
Enable C++ Exceptions option to /EHa.
That'll treat SEH exceptions like C++ exceptions.
- Rom
-Original Message-
From: Raistmer [mailto:raist...@mail.ru]
Sent: Friday, October 16, 2009 1:49 PM
To: Rom Walton; boin
yes.
- Original Message -
From: "Rom Walton"
To: "Raistmer" ;
Sent: Friday, October 16, 2009 9:34 PM
Subject: RE: [boinc_dev] Disabling BOINC's debugger
Are you building your app with Visual Studio?
- Rom
-Original Message-
From: Raistmer [mailto:raist...@mail.ru]
Sent:
Are you building your app with Visual Studio?
- Rom
-Original Message-
From: Raistmer [mailto:raist...@mail.ru]
Sent: Friday, October 16, 2009 1:32 PM
To: Rom Walton; boinc_dev@ssl.berkeley.edu
Subject: Re: [boinc_dev] Disabling BOINC's debugger
I use
try{
function_call();
}cat
I use
try{
function_call();
}catch(...)
{
handler();
}
but still got BOINC's debugger called...
- Original Message -
From: "Rom Walton"
To: "Raistmer" ;
Sent: Friday, October 16, 2009 8:59 PM
Subject: RE: [boinc_dev] Disabling BOINC's debugger
The exception handling code on
The exception handling code on Windows only kicks in if no other
exception handling mechanism is used.
Just define an exception handler of some sort and you should be good to
go.
Just remember, on Windows there are two different exceptions that can be
thrown. C++ exceptions and SEH exceptions.
Hello
Is it possible to give control for exception handling back to science app?
What BOINC API call I should comment out to do this?
Need exception handling inside appto catch exception instead of writing
to stderr and trashing task...
___
boinc_de
It sound like you need to add the to the file_info struct for
libcudart.so.
See:
http://boinc.berkeley.edu/trac/wiki/AppLibraries
That should clear up this issue.
- Rom
-Original Message-
From: boinc_dev-boun...@ssl.berkeley.edu
[mailto:boinc_dev-boun...@ssl.berkeley.edu] On Beha
This change has had a couple of weeks to shake down, and in general seems to
be working well. However, one unanticipated problem surfaced at AQUA last
night, and could possibly use a refinement.
AQUA had a CUDA app installed, but dormant. They applied the server update
for this issue, and I was
Hi Bernd.
I modified your first shell script:
#!/bin/sh
LD_LIBRARY_PATH=../../projects/test.test/
echo $LD_LIBRARY_PATH > /tmp/ldlibr.log
echo $PWD >> /tmp/ldlibr.log
../../projects/test.test/hello_cuda2 "$@"
then: before call boinc:
LD_LIBRARY_PATH=../../projects/test.test/
PWD=~/
24 matches
Mail list logo