NVIDIA Eclipse-Based IDE for GPU Computing on Linux and Mac OS Announced
On the first day of NVIDIA's GPU Technical Conference 2012, the company launches Nsight Eclipse Edition for both Linux and Mac operating systems. Not forgetting developers on Windows platform, the Nsight tools that plug into Microsoft Visual Studio development suite have also been updated.
With the release of this new edition of Nsight, NVIDIA has allowed CUDA programmers not only the option to work on either Linux- or Mac-based platforms, the company has also opened the avenue for these developers to compile their CUDA programs from within the open-source Eclipse Integrated Development Environment (IDE). Nsight leverages on the rich development tools of the Eclipse IDE to allow CUDA programmers to develop, debug as well as optimize their GPU-accelerated applications.
An important new feature of Nsight is its ability to perform code refactoring, allowing it to convert sequential CPU loops into parallel GPU kernels. This effectively speeds up the application by offloading such parallel tasks to the CUDA cores of the GPU. For a comprehensive list for GPUs supporting CUDA, please visit this official link of NVIDIA.
Nsight has the ability to syntax highlighting and auto-completion of reserved words and other items based on the program naming convention as it aims to increase the productivity of the programmers. Coupled with this new edition is a set of expert analysis tools, which provides automated performance analysis and step-by-step guidance to address application performance bottlenecks.
Not forgetting Windows developers, NVIDIA also announced the updated NVIDIA Nsight Visual Studio Edition. Formerly known as NVIDIA Parallel Nsight, this newer version adds a number of enhancements and updated features designed to make parallel programming on GPU-based Windows systems faster and easier. One of the most important improvement over the older Parallel Nsight is the newer Nsight is able to support local single GPU debugging. In the older version, at least two GPUs were needed during debugging as one GPU was required to execute the code and the other, the debugger. Debugging in NVIDIA Nsight Visual Studio Edition requires only a single CUDA-enabled GPU with CUDA 1.1 or higher version, saving money as developers wouldn't have to invest in another GPU just for debugging purposes.
Both versions of NVIDIA Nsight are now available for download without cost for NVIDIA registered developers. For more details pertaining to these software development suites, please visit the following links:- NVIDIA Nsight, Eclipse Edition and NVIDIA Nsight, Visual Studio Edition.
Read on for the full press release.
SINGAPORE — NVIDIA today introduced NVIDIA Nsight, Eclipse Edition, the world’s first integrated development environment (IDE) for developing GPU accelerated applications on Linux and Mac OS-based systems.
NVIDIA Nsight provides powerful debugging and profiling tools that enable high performance computing (HPC) and graphics developers to fully optimise the performance of CPUs and GPUs.
The new Nsight, Eclipse Edition, enables CUDA programmers to easily develop, debug and optimise the performance of GPU-accelerated applications within a familiar, highly productive IDE based on the open source Eclipse framework (www.eclipse.org).
Key features include:
- Automatic code refactoring – Helps convert slow sequential CPU loops into parallel GPU kernels
- Integrated expert analysis system – Provides automated performance analysis and step-by-step guidance to address application performance bottlenecks
- High-productivity development environment – Syntax highlighting and auto-completion for both CPU and GPU code helps developers program more efficiently
- Integrated code samples, online documentation – Makes it easy for developers to quickly get started
NVIDIA Nsight Visual Studio Edition for Windows
NVIDIA also announced an updated version of NVIDIA Nsight, Visual Studio Edition for Microsoft Windows developers. Nsight, Visual Studio Edition (formerly known as NVIDIA Parallel Nsight) adds a number of new enhancements and updated features designed to make parallel programming on GPUbased Windows systems faster and easier than ever.
Key among these features is local single GPU debugging, which enables CUDA developers to debug their CUDA C/C++ code natively on the hardware with any system equipped with any CUDA 1.1 or higher capable GPU. Other features include performance improvements to the frame profiler and debugger, and support for DirectX 9 frame debugging, frame profiling and analysis.