Nsight Compute profiling CUDA 10.1 Task Graphs

OpenGL and Vulkan interoperability on the Graphics Debugger

Next-Gen CUDA Debugger with PTX+SASS source code correlation showing the state at a kernel breakpoint

NVIDIA® Nsight™ Visual Studio Edition 2019.3 is now available with these features and improvements:

Graphics Debugging

    API Enhancements
      • OpenGL 4.6 is now supported in Frame Debugging, C++ Capture, and Frame Profiling activities
      • Range Profiler now has OpenGL support has been added for Turing based GPU targets
      • New suport for GL_OVR_multiview and GL_OVR_multiview_multisampled_render_to_texture extensions
      • Basic support for OpenGL Immediate Mode has been added for the most common use cases
    Feature Enhancements
      HUD - Incompatible Capture UX
      • added the ability to accept a notification and proceed at the user’s own risk from the HUD when an incompatible capture is detected
      API Inspector
      • Meshlet pipeline stages have been added to the D3D12 API Inspector
      Frame Debugger
      • Improved capture and interception performance
    Other Improvements
      Windows 10 19H1
      • Support for applications using Variable Rate Shading APIs – ID3D12GraphicsCommandList5::RSSetShadingRate and ID3D12GraphicsCommandList5::RSSetShadingRateImage
      Multi-View Rendering support has been added

Compute Debugging and Analysis

    • Supports CUDA Toolkit 10.1 Update 2
    Next-Gen Debugger
    • Works with the latest Turing Super GPUs
    • OptiX applications can now be debugged (requires a 435 driver)
    • Core Dump Analysis added
    • Barrier stepping and end-of-kernel stepping control added
    Nsight Compute Profiler
      • Works with the latest Turing Super GPUs
      • OptiX applications can now be profiled (requires a 435 driver)
      • Reduced the profiling overhead, especially if no source metrics are collected
      • Reduced the overhead for non-profiled kernels
      • Trying to profile on an unsupported GPU now shows an "Unsupported GPU" error message
      • Added support for smsp__sass_* metrics on Volta and newer GPUs
      • The launch__occupancy_limit_shared_mem now reports the device block limit if no shared memory is used by the kernel
      NVIDIA Nsight Compute
      • The heatmap on the Source page now shows the represented metric in its tooltip
      • The Memory Workload Analysis Chart on the Details page now supports baselines
      • When applying rules, a message displaying the number of new rule results is shown in the status bar
      NVIDIA Nsight Compute Command Line Interface
      • Fixed an issue that reported the wrong executable name in the Session page when attaching
      • Fixed issues that chart labels were shown elided on the Details page
      • Fixed an issue that caused the cache hitrates to be shown incorrectly when baselines were added
      • Fixed an illegal memory access when collecting sass__*_histogram metrics for applications using PyTorch on Pascal GPUs
      • Fixed an issue when attempting to collect all smsp__* metrics on Volta and newer GPUs
      • Fixed an issue when profiling multi-context applications
      • Fixed that profiling start/stop settings from the connection dialog weren't properly passed to the interactive profile activity
      • Fixed that certain smsp__warp_cycles_per_issue_stall* metrics returned negative values on Pascal GPUs
      • Fixed that metric names were truncated in the --page details non-CSV command line output
      • Fixed that the target application could crash if a connection port was used by another application with higher privileges

For a complete overview of all Nsight™ Visual Studio Edition features and access to resources, please visit the main Nsight™ Visual Studio Edition page.

NVIDIA® Nsight™ Visual Studio Edition 2019.3 is available for download under the NVIDIA Registered Developer Program.

 Download   Documentation