Recently, Bill Dally (Bill Dally), CTO of NVIDIA, said that he was "aware of some projects, which aim to use CUDA on other platforms." Naturally, almost all, read the interview, decided that he was referring to graphics processors AMD. Nevertheless, according to representatives from NVIDIA, because there was something completely different.

Currently, there are several effective GPGPU common standards, including OpenCL and Microsoft DirectX. In addition to these standards, the chip for graphics production NVIDIA and ATI are actively used its own technology manufacturers, but NVIDIA CUDA is considered slightly better than the AMD Stream, and it is this consideration, together with a statement Dally led to the assumption that CUDA will be applied to graphics chips ATI. However, as indicated by representatives of NVIDIA, so that CUDA can be used in the GPU ATI, NVIDIA need to get access to ATI, which is certainly not going to happen. Thus, despite the expectations of users, NVIDIA will not adapt its technology to chip AMD, and the question of what was due to Dally, saying the country phrase, remains open.