Parallel Nsight 1.51 - New Features
For a complete overview of all Nsight™ Visual Studio Edition features and access to resources, please visit the main Nsight™ Visual Studio Edition page.
Archived 05/09/2013: Nsight Visual Studio Edition 2.2 is no longer supported.
New NVIDIA Nsight™ Visual Studio Edition 2.2 Features:
Graphics Debugging and Profiling
- Support DirectX 9 for Frame Debugging, Frame Profiling and Analysis.
- Support for the new Kepler architecture (For example, GeForce GTX 680). Does not include support for GK110 based GPUs (For example, Tesla K20).
- Improved Frame Profiler with performance improvements and additional experiments.
- Improved Frame Debugger Visual Studio frame scrubbing performance.
Compute Debugging and Profiling
- Local CUDA debugging on a single GPU system.
- Support for the new Kepler architecture (For example, GeForce GTX 680). Does not include support for GK110 based GPUs (For example, Tesla K20).
- Support for CUDA Toolkit 4.2 (available at http://www.nvidia.com/getcuda).
- The CUDA information tool window now shows the number and size of memory allocations per context per device.
- Change focus to a specific thread within a warp in the lanes column of the CUDA information tool window.
- New freeze and thaw commands to control warp execution in the CUDA information tool window.
- Step through PTX or SASS disassembly if the CUDA application is built without debug information (no –G).
- Scroll through the disassembly in the entire address space.
- Ability to enter a function or symbol names in the disassembly toolbar.
- Support for creating a function breakpoint by using the stock Visual Studio breakpoint options.
- Ability to set a breakpoint from the Functions page in the CUDA information view.
- Drastic performance improvement when scrolling the warp page in the CUDA information tool window.
- Analysis summary pages have been redesigned to give a true high level picture of how the application utilized different areas.
Figure 1: Stepping through PTX or SASS disassembly if the CUDA application is built without debug information (no –G).
Figure 2: Redesigned Analysis Summary Pages
Application and System Trace
- Support for the new Kepler architecture (For example, GeForce GTX 680). Does not include support for GK110 based GPUs (For example, Tesla K20).
- Support for CUDA Toolkit 4.2 (available at http://www.nvidia.com/getcuda).
- Trace support for DirectX memory transfers.