The latest update the NVIDIA HPC SDK is now available for download.
Overview
The NVIDIA HPC SDK 20.11 update introduces new features and enhancements for high-performance computing developers, including support for automatic GPU acceleration and new libraries. This comprehensive suite allows developers to program across the entire HPC platform, from GPUs to CPUs.
What You'll Learn
1
How to utilize the new nvlamath Fortran library for GPU acceleration
2
Why C++17 parallel algorithms are significant for HPC development
3
When to implement OpenMP target offload on Volta or newer GPUs
Prerequisites & Requirements
- Understanding of high-performance computing concepts
- Familiarity with CUDA Toolkit(optional)
Key Questions Answered
What new features are included in NVIDIA HPC SDK 20.11?
NVIDIA HPC SDK 20.11 includes beta support for OpenMP target offload on Volta or newer GPUs, GPU acceleration for Fortran DO CONCURRENT constructs, and the new nvlamath Fortran library for standard BLAS routines. Additionally, it supports CUDA Toolkit 11.1 and introduces new command-line options for better GCC compatibility.
How does the NVIDIA HPC SDK support automatic GPU acceleration?
The NVIDIA HPC SDK provides the first C++ and Fortran compilers that automatically accelerate standard language constructs, including C++17 parallel algorithms and Fortran intrinsics. This feature simplifies the process of leveraging GPU capabilities in HPC applications.
Technologies & Tools
Software Suite
Nvidia Hpc SDK
A comprehensive suite of compilers, libraries, and tools for HPC development.
Software Toolkit
Cuda Toolkit
Provides support for GPU programming and is integrated into the HPC SDK.
Key Actionable Insights
1Leverage the new nvlamath Fortran library to enhance the performance of your applications by accelerating standard BLAS routines.This is particularly useful for developers working on applications that require linear algebra computations, as it can significantly reduce execution time.
2Adopt C++17 parallel algorithms in your HPC projects to take advantage of automatic GPU acceleration.This allows for more efficient code that can run on both CPUs and GPUs, improving overall performance and resource utilization.
3Consider using the new command-line options for better compatibility with GCC when compiling your applications.This can help avoid potential issues during the build process, especially if your development environment relies on GCC.
Common Pitfalls
1
Failing to utilize the automatic GPU acceleration features can lead to suboptimal performance in HPC applications.
Developers may overlook these features, resulting in missed opportunities for performance gains, especially in compute-intensive tasks.