I just update an old laptop with VS2013 and have a 9800gts, and it will be great to make some algorithms with it in home, but the GPU is no longer in Nsight Visual Studio Edition Supported GPUs (Full List). So there will never again compile any CUDA code with my old and reliable laptop?. Some workaround or further options will be great… Thanks and salve CUDA members!
I am sorry to say that we are not support 9800gts in the latest version. One option is to revert back an old version in order to use 9800.
Thanks for your fast response Victor,
So, that means install an older VS as 2010 and run CUDA/Nsight 5 package? Just wondering… is any way to run a old Nsight with VS2013? I tried but no avail… I know that a 9800gts is really old (but that thing is working like a clock!), but for testing some functions, algorithms, etc. it’s really great and save the file and open with the same VS makes the things easy (I have some issues with the files from VS2010 to VS2013 in C# and ASP.NET) Thanks and again Salve CUDA members!
the old NSight is not supported in 2013. you have to use VS2010.
Such a bummer…
Again thanks Victor, if you can answer other issues I’ll be gratefull:
- Why force the user/developer to buy a new machine for keep the pace of the software technology?
- VS2010 install 40^N c++ runtime .exe and VS2013 it’s the same, it´s overload a machine with troubles…
- I’m no expert in CUDA, I am in the capter 4 of GPU gems and Massively Parallel chapter 5, so why no make the things a little easy for students or learners (like me who learn in their free time) and improve some retrocompatibility? CUDA it’s not Haskell or something to install and run with or without a IDE… and learn all about it.
Thanks again for your time… Salve CUDA members!
Hi, User MaskedVll
- GTX9800 is a six years old product. You can see the new software technology is always driven by the hardware achievement. NVIDIA tried the best in the GPU market to support historic products.
- We run a lot test on machines with both VS2010 and VS2013 installed. They use disk space, but they work with each either. VS2013 does support set the compiler to VC2010 but manage the project in VS2013, that will make you happy to do “save and open in the same VS”.
Tesla architecture (GTX9800) was dropped from CUDA 6.5. That means you can use CUDA up to 6.0.
- As I said, NVIDIA tried the best to support old products and always trying to make things easier for developers.
If you have more questions about how to setting up your development with CUDA 6.0, VS2010/VS2013. Feel free to send emails to firstname.lastname@example.org. We will have people to support you in detail.
I didn’t know how old was mi GPU! my lap it`s only 4 years with me and I have 1 year old with CUDA, my bad… I check status and stadistics and other stuff, but never see when was released!.
I just want to make, compile, study basic algorithms and functions and trace what is happening with the tools that you offer in a great and free way (thanks to all in nVidia for this) not creating CG in realtime for that I have other machine, but not in home. I have readed about switching the compiler I’ll start testing. And I know that was a blind ambition to see the new software running in “my old and reliable lap” but… you can’t blame me for trying.
You are really a profesionals… salve CUDA members!