Hello,
However, I am now migrating to Vista on a laptop with a integrated Quadro FX 3700M and seeing
I’ve also tried the following drivers:
devdriver_3.0_winvista-win7_32_197.16_notebook.exe
Both fail.
Any help would be appreciated!
This appears to be occuring when a worker thread that was created and shared w/my main
OpenGL context tries to create memory. I’ve looked at the release notes and it says:
[i] Attempting to map or unmap while a different context is bound than was
current during the buffer register operation will generally result in a
program error and should thus be avoided.[/i]
Does this mean that even though the documentation says you can access mapped memory from
different contexts as long as they are in the same OpenGL share group this should not work? It works for me on XP.
Jay
Hello,
We currently have a computer vision application that uses CUDA in a multi-threaded enviroment.
We’re using shared OpenGL buffers and cuda toolkit 3.0 to map memory to CUDA.
Everything works fine in Windows XP on my development machine with a descrete Quadro FX 580.
However, I am now migrating to Vista on a laptop with a integrated Quadro FX 3700M and seeing
painful issues. Every time a I attempt to create plain old CUDA device memory CUDA returns
“Out of Memory”. I can still populate my OpenGL buffer and register it with CUDA
fine but new any memory will not be created.
I’ve also tried the following drivers:
devdriver_3.0_winvista-win7_32_197.16_notebook.exe
257.21_notebook_winvista_win7_32bit_international_whql.exe
Both fail.
Any help would be appreciated!
Thanks,
Jay
Windows 7 also has the same issue on the XP machine that worked.
Is sharing OpenGL memory in multiple threads with CUDA NOT supposed to
be supported? If not then working in XP is just a fluke?
TMURRAY please help!
This appears to be occuring when a worker thread that was created and shared w/my main
OpenGL context tries to create memory. I’ve looked at the release notes and it says:
[i] Attempting to map or unmap while a different context is bound than was
current during the buffer register operation will generally result in a
program error and should thus be avoided.[/i]
Does this mean that even though the documentation says you can access mapped memory from
different contexts as long as they are in the same OpenGL share group this should not work? It works for me on XP.
Jay
Windows 7 also has the same issue on the XP machine that worked.
Is sharing OpenGL memory in multiple threads with CUDA NOT supposed to
be supported? If not then working in XP is just a fluke?
TMURRAY please help!
This appears to be occuring when a worker thread that was created and shared w/my main
OpenGL context tries to create memory. I’ve looked at the release notes and it says:
[i] Attempting to map or unmap while a different context is bound than was
current during the buffer register operation will generally result in a
program error and should thus be avoided.[/i]
Does this mean that even though the documentation says you can access mapped memory from
different contexts as long as they are in the same OpenGL share group this should not work? It works for me on XP.
Jay