In a previous blog, I shared news of work that we at Mentor Embedded had done with NVIDIA to help enable development of Android applications for NVIDIA Tegra devices, within the Microsoft Visual Studio environment. This technology is a key part of the NVIDIA NSight Tegra (https://developer.nvidia.com/nvidia-nsight-tegra) product.
Earlier in the month, we hosted a joint webinar with NVIDIA to talk about this offering. In case you missed the webinar, you can view the recording at http://www.mentor.com/embedded-software/multimedia/enabling-android-development-within-microsoft-visual-studio-for-nvidia-tegra-devices
Judging from the live attendance and archive views, this has been an extremely popular topic. There was a healthy QnA session towards the end of the webinar, which unfortunately is not captured in the recording. However, I have reproduced some of these here:
Q: Does Tegra4 have fpu? Also, can you use Cuda on the GPU? What is the compute compatibility? 2.0??
A: It does have an fpu, but tegra does not support cuda today
Q: Any specific visual studio version req’d?
A: VS2010 is required today, VS2012 support will be available soon. Sadly Visual Studio Express can’t be supported
Q: Could you talk a little bit about tools available for ODMs who build the full Android OS for their devices?
A: I’ll cover some of the available offerings at the end of the session. Also visit http://www.mentor.com/embedded-software/sourcery-tools-services/ for more information
Q: Is debugger breakpoint based or event based as well?
A: Both are supported – standard breakpoints, conditional breakpoints, etc
Q: Does the Nsight Tegra has any support for CUDA ?
A: Tegra processors don’t support CUDA today
Q: Is there any way to program tegra gpu for general purpose applications?
A: Yes, there are a number of gpgpu solutions available for Tegra. Anything based on OpenGL ES 2.0 should work
Q: Does the debugger support renderscript
A: I’m sorry, renderscript is not supported
Q: If I develop an app with this NVIDIA SDK, does it will only run on a device having NVIDIA platform ?
A: Nope – it will work on all platforms, but you can only debug on Tegra. For support for a non-NVIDIA platform, submit your request here.
We’d love to hear your comments and suggestions around this technology or any related topic. Post these on the Sourcery Communities at http://communities.mentor.com/mgcx/community/embedded_software/sourcery_codebench