NVIDIA has announced a new open beta for its Parallel Nsight plug-in, which enables Microsoft Visual Studio developers to create applications that optimize both graphical processing units (GPUs) and CPUs within a single development environment.
Announced in conjunction with the April 12 release of Microsoft Visual Studio 2010, this new beta adds features that are found in the latest GeForce and Tesla products based on the Fermi architecture, namely the Nsight Debugger, Nsight Analyzer, and Nsight Graphics Inspector, NVIDIA officials said.
The key new features for the open beta, which have been added underneath the three key technologies of Parallel Nsight, include Nsight Debugger, with NVIDIA Fermi architecture support for debugging Compute Unified Device Architecture (CUDA) C source code, and improved parallel debugging visualizations for DirectX 10 High Level Shading Language (HLSL) debugging.
The Nsight Analyzer brings OpenCL 1.0 call trace visualization correlated to CPU+GPU events on the timeline; timeline visualization of DirectX 11 API calls; and CUDA Profiler support for Fermi hardware counters. And the Nsight Graphics Inspector delivers DirectX11 frame debugging; DirectX11 frame profiling, with per-draw call bottleneck detection; and pixel history fully supported for Tesla (coming soon for Fermi).
With these additions, Parallel Nsight makes programming the GPU, for both computing and graphics applications, easy and accessible within the comfortable workflow of Microsoft Visual Studio, NVIDIA officials said.
Parallel Nsight supports Windows 7 and Windows Vista operating systems and full integration within Microsoft Visual Studio (2008 SP1 standard edition or later). The beta 2 version of Parallel Nsight supporting DirectX 11 is scheduled to be available in less than a month, with the first full production release scheduled for June. Both standard and professional versions of NVIDIA Parallel Nsight will be available upon final release.