I was so excited about the news for Dynamic parallelism in CUDA 5.0. However, after I installed CUDA 5.0 I read the Dynamic Parallelism programming guide and it says that Dynamic parallelism is only supported on devices of compute capability 3.5 and higher. Taking a look at Nvidia’s products, the only 3.5 device I could find is Tesla K20, which is not available in any of the stores in my area. What’s the best and cheapest way to start developing with dynamic parallelism on my Toshiba Satellite notebook?
I have Windows 7 (64-bit)
NVIDIA GeForce GT 540M
Dynamic parallelism is available in any GPU that uses the GK110 chip. That is, GTX 780, GTX 780 Ti, GTX Titan, and the Tesla K20, K20X K40 cards.
There are GeForce devices with compute capability 3.5, but I don’t know if you can install any of them into your notebook system. (Assuming you even have a removable GPU in your notebook.)
If you have access to a desktop system, there are several options. On the expensive end of the scale, the GTX 780, 780 Ti and Titan are all compute capability 3.5 devices. On the very cheap end of the scale, there are some GK208-based cards described here:
Unfortunately, it can be a little tricky to figure out which desktop cards are actually using a GK208 chip due to confusing model numbering. The GK208 cards are also extremely slow, so if you can afford it, I would recommend at least the GTX 780.