compute capability/CUDA Toolkit 3.1


I saw a new release of the CUDA Toolkit (3.1 Beta) which supports a lot of new features, but I am wondering if I am limited by my card’s compute capability. I read the following lines in the Programming guide:

“Devices with the same major revision number are of the same core architecture. The major revision number of devices based on the Fermi architecture is 2. Prior devices are all of compute capability 1.x (Their major revision number is 1).”

I am actually using a Tesla C1060 card, so its compute capability is 1.3.
Does it mean that I will not be able to use the new features available on devices of compute capability 2.0 ? Is it really an architecture limitation or can the card be used in a way to support them (for example use function pointers) ?

Thank you for your answer.


Compute x.y defines the hardware architecture. Compute 1.3 cards simply do not have the same features in silicon that compute 2.0 cards do. To utilize function pointers, recursion, etc… you must upgrade to a compute 2.0 card.

Thanks !