We're pleased to announce that NVIDIA Parallel Nsight 1.0, our GPGPU and graphics development environment integrated into Visual Studio, has been released. Parallel Nsight bring GPGPU and graphics developers an unprecedented amount of power, and supports APIs like CUDA C/C++, OpenCL, DirectCompute, Direct3D, and OpenGL.
Getting the build:The build is now available from the
Parallel Nsight Developer Zone page.
Follow the instructions on the download page to download and then activate your build immediately.
Parallel Nsight Debugger for GPU Computing
- Debug your CUDA C/C++ and DirectCompute source code directly on the GPU hardware
- As the industry's only GPU hardware debugging solution, it
drastically increases debugging speed and accuracy
- Use the familiar Visual Studio Locals, Watches, Memory and
Breakpoints windows
|
Parallel Nsight Analyzer for GPU Computing
(Professional version only)
- Isolate performance bottlenecks by viewing system-wide CPU+GPU events
- Support for all major GPU Computing APIs, including CUDA C/C++, OpenCL, and Microsoft DirectCompute
|
Parallel Nsight Debugger for Graphics Development
- Debug HLSL shaders directly on the GPU hardware. Drastically increasing debugging speed and accuracy over emulated (SW) debugging
- Use the familiar Visual Studio Locals, Watches, Memory and Breakpoints windows with HLSL shaders, including DirectCompute code
- The Debugger supports all HLSL shader types: Vertex, Pixel, Geometry, and Tessellation
|
Parallel Nsight Graphics Inspector for Graphics Development
- Graphics Inspector captures Direct3D rendered frames for real-time examination
- The Frame Profiler automatically identifies bottlenecks and performance information on a per-draw call basis
- Pixel History shows you all operations that affected a given pixel
|
Once again, NVIDIA developers will be presenting at this year's GDC in San Francisco. This is your chance to see new material and success stories from developers that built games like Dark Void, EVE Online, and many others, as well as hands-on opportunities to try out our new tools, including NVIDIA Parallel Nsight (formerly codenamed "Nexus"). NVIDIA be located in the main expo area, booth 2008
To stay up-to-date on NVIDIA's presentations and demonstrations, please visit our special GDC page at http://developer.nvidia.com/object/gdc-2010.html
NVIDIA Fellow David Kirk and UIUC Professor Wen-mei Hwu have published their landmark textbook, Programming Massively Parallel Processors: A Hands-on Approach. This book will enable generations of programmers to understand and exploit the massively parallel architecture of GPUs. Various techniques for constructing parallel programs are explored in detail. The book also uses case studies to demonstrate the development process, which begins with computational thinking and ends with effective and efficient parallel programs. If you are looking to acquire a solid foundation in massively parallel development, this is an excellent place to start.
The book can be found at Amazon (http://is.gd/7bNYP) as well as many other bookstores.
We're proud to announce the trailer video for NEXUS, NVIDIA's upcoming Visual Studio-integrated toolset for graphics and GPU Computing. NEXUS includes powerful debugging and platform-wide performance tools.
More information about NEXUS will be available at the GPU Technology Conference, Sept 30th - Oct 2nd, 2009.
Beginning this week we'll be releasing the full text of GPU Gems 3 on the NVIDIA website -- Seven chapters to start, and then three chapters per week between now and GDC 2009.
Click here to begin reading the first section of GPU Gems 3,
Geometry.
Chapters this week:
Dan Amerson of Emergent Game Technologies presented "Stream Processing with CUDA: A Case Study Using Gamebryo's Floodgate Technology." Dan's presentation can now be found at http://developer.nvidia.com/object/nvision08-gamebryo.html
The latest version of NVIDIA CUDA now supports Apple Macintosh OS X 10.5 ("Leopard"). Have a an NVIDIA-powered recent Mac? See "High Peformance Computing for the Rest of Us" today! Details on the CUDA 1.1 release can be found here at the CUDA Zone.
The latest versions of our CUDA toolkit and SDK provide everything you need to harness the GPU's power for any kind of computing. This latest release of the CUDA SDK features several new samples, such as:
- Binomial Option Pricing
- Black-Scholes Option Pricing
- Monte-Carlo Option Pricing
- Parallel Mersenne Twister (random number generation)
- Parallel Histogram
- Image Denoising
- Sobel Edge Detection Filter
Altogether you'll find more than 40 code samples that you can learn from and extend for your own projects.
If you would like to work more closely with NVIDIA on a High-Performance Computing (HPC) project, please sign up to become a Tesla Registered Developer.
For the latest news about Tesla, please sign up for our newsletter.
CUDA Home page
We are pleased to unveil the cover of the upcoming “GPU Gems 3” book. This image has been rendered in NVIDIA’s “Human Head” demo, which features Doug Jones, who recently starred as the Silver Surfer in Fantastic Four: Rise of the Silver Surfer.
The book will be in full color, hard cover, and will contain 41 chapters, including “Advanced Skin Rendering”, which explains how to render ultra-realistic characters. GPU Gems 3 will be released at SIGGRAPH 2007 but can already be pre-ordered.
We will release more information about this project as we get closer to SIGGRAPH.
John Nickolls, Director of Architecture, NVIDIA will present “NVIDIA CUDA Software and GPU Parallel Computing Architecture” this presentation will be a first detailed disclosure of NVIDIA’s CUDA parallel software and NVIDIA's Massively Multi-Threaded Architecture.
Max Baron, In-Stat principal analyst and the content editor for this year’s Fall Microprocessor Forum, said, “We are pleased to have NVIDIA presenting one this year’s highly anticipated papers. Microprocessor Forum gathers together a sophisticated engineering audience from around the globe and we feel they won’t be disappointed this year when we are working with such high caliber companies such as NVIDIA.
Microprocessor Forum will take place May 22 – 23, 2007 in San Jose, California at the Doubletree Hotel.
In-Stat is proud to present Microprocessor Forum 2007 – check it out online at:
http://www.instat.com/mpf/07/index.htm
