NVIDIA Nsight is a powerful development toolkit that was especially designed to provide programmers with all the needed components for creating applications that are able to use the GPU and thus relieve some of the CPU stress.
The package includes debuggers for GPU computing and graphics, system and application tracers, as well as a graphics profiler.
Note: If CUDA Toolkit is needed, you can download it here.
Here are some key features of "NVIDIA Nsight":
· GPU CUDA debugging and memory checker
· Integrates with Visual Studio
· CUDA kernel debugger
· Direct3D shader and frame debugging
· Direct3D Pixel History and Dynamic Shader Editing
· System and applcaition trace
· Intel Pentium Dual-core CPU or equivalent 1.6 GHz minimum
· 2 GB minimum on host and target machines
· 32-bit machine: 240 MB for NVIDIA Nsight software
· 64-bit machine: 330 MB for NVIDIA Nsight software
· Monitor dedicated to target GPU
· One or more GPUs, must be a supported graphics card.
· CUDA Debugging - Minimum of 1 GPU
· Graphics Debugging - Minimum of 2 GPUs
· You must install an NVIDIA display driver that supports NVIDIA Nsight.
· NET Framework 3.5 with SP1
· Microsoft Visual Studio 2008 with SP1 Standard Edition or better
What's New in This Release: [ read full changelog ]
Graphics Debugging and Profiling:
· Support OpenGL 4.2 Head-Up-Display to overlay key performance statistics.
· Support OpenGL 4.2 for frame debugging, pixel history and frame profiling.
· Note: The OpenGL application must be 4.2 core compliant in order to use local shader debugging.
· Support for OpenGL GLSL GPU shader debugging.
· GLSL 3.3 and higher applications are supported on a remote debugging setup.
· GLSL 4.2 core applications are supported on a local debugging setup.
· Local, single GPU shader debugging and pixel history is now supported for HLSL and GLSL.
· DirectX frame capture generates Visual Studio project with source code.
· Improved frame debugger Visual Studio frame scrubbing performance.
· Support for the CUDA 5.0 Toolkit.
· Support for the Kepler GK110 architecture (for example, found in the Tesla K20).
· CUDA Dynamic Parallelism is now supported when building, debugging, and running analysis. For more information, see CUDA Dynamic Parallelism.
· The CUDA memory chec...