Nvidia releases Linux Cuda 2.2

NVIDIA HAS ANNOUNCED its latest version of Linux Cuda 2.2 Beta, a hardware level refresh the Green Goblin hopes will amaze and delight GDB users everywhere.

The software toolset, which exposes the GPU's processing abilities to x86 software, lets developers use a hardware-level debugger and profiler to help track down and kill problems in their code.

Nvidia reckons the new version will also allow developers - but only the ones with Nvidia graphics cards, of course - to isolate individual operations in real-time.

The Nvidia spinfo says the refresh means "Linux developers can now use a debugger on Cuda-enabled GPUs that offers both the familiar interface of the popular open-source GDB debugger and the ability to debug kernels as they execute on the GPU".

Purportedly, the GPU-side debugger also has all the features developers would expect from GDB in the first place, like the ability to have breakpoints, watch variables, inspect state and more. Oh, yeah, and it also packs "additional functions for Cuda-specific features", whatever they might be.

Nvidia is also touting the latest release of its Cuda Visual Profiler, another developer friendly tool allowing for full measurement of memory bandwidth within a kernel.

Lest we forget, Nvidia says Cuda 2.2 features better OpenGL interop performance, texture from pitch linear memory, Zero-copy support for direct access to system memory, pinned shared system memory and asynchronous memcopy operations in Vista.

The update also means Cuda 2.2 is now supported on Windows, Mac OS X, and all major Linux distributions. µ

L'INQ
Cuda wuda, shuda