Hello,
I am new with nvcc.
Is it possible to build application in windows, cross building with nvcc, and then run the binary in linux ?
We need to build in eclipse (windows) and run & debug a remote linux.
Thank you,
Ran
Hello,
I am new with nvcc.
Is it possible to build application in windows, cross building with nvcc, and then run the binary in linux ?
We need to build in eclipse (windows) and run & debug a remote linux.
Thank you,
Ran
It’s not a supported toolchain/process.
It should be possible to run a VM in windows that hosts linux, and build there.
Hi txbob,
Thank you.
If I understand correctly, nvcc for windows should be used (build+run binary) only in windows, and nvcc for linux only in linux(build+run binary) , Right ?
One more thing if I may.I see that it has dgb, gdbserver support.
So If I use the VM on windows, do you think it should it be possible to use eclipse in VM, and debug/deploy a remote traget ?
Thanks a lot!
Ran
If you use nsight eclipse edition, it should be possible to use it in the remote fashion that is supported:
Thank you,
I followed the documentation, and it is very interesting eclipse version.
Though, not sure what the additional value between this eclipse version, to just configuring a toolchain in any other eclipse version.
BR,
Ran
Is CUDA cross compilation Windows → Linux still impossible?
NVIDIA doesn’t provide any tools for this.
With the advent of WSL2, you can compile for linux using WSL2.
Yes I am evaluating this approach too. Unfortunately I couldn’t find an an easy way of integrating CUDA remote builds in the WSL environment for Visual Studio / MSBuild. Something like the CUDA target file that needs to be included in C++ projects to enable CUDA support…