Nvidia releases CUDA 2.2 with Windows 7 support

Santa Clara (CA) – Nvidia today released version 2.2 of its CUDA software development platform, which allows programmers to create multithreaded GPGPU applications that run on GeForce and Tesla cards.

The latest version of the SDK introduces support for Windows 7, which Nvidia describes as “the upcoming OS from Microsoft that embraces GPU computing.” We would not quite as far, since Vista has been supported in previous versions, and Windows 7 lacks direct GPU acceleration features.

The recently released OpenCL (pre-beta) driver (which we are looking forward to see the effects of this driver on Apple’s Snow Leopard OS, which will support OpenCL GPGPU acceleration has not made it into this version of CUDA, but Nvidia claims that 2.2 delivers improved performance for OpenGL applications running on Quadro GPUs when computing with CUDA and rendering OpenGL graphics functions are performed on different GPUs.

Other new features include a visual GPU profiler, a graphical tool that enables the profiling of C applications running on the GPU; enhanced bandwidth efficiency for video processing applications, zero-copy support, pinned shared sysmen, asynchronous memcopy for Vista, exclusive device mode and a hardware debugger for the GPU.