A Streamlined Route to HPC in the Cloud
Use of Cloud for High-Performance Computing (HPC) Applications Continues to Rise
HPC in the cloud market is expected to grow to USD $9 billion by 20241
HPC simulation is driving innovation in financial services, healthcare and transportation
Intel and AWS have collaborated for over 10 years to drive productive cloud environments that empower application developers and their users to meet HPC and artificial intelligence (AI) needs.
Intel and AWS’ standards-based approach provides a feature-rich cloud environment that can be tailored to meet the demands of a wide range of HPC and AI workloads.
Key Features of the Collaboration for AWS Developers
1. Innovative Intel® Hardware
Built-in Features Boost Performance for AI and HPC Workloads
Intel® Advanced Vector Extensions 512 (Intel® AVX-512) can benefit a wide range of workloads, such as scientific simulations, financial analytics, AI/deep learning, 3D modeling and analysis, image and audio/video processing, cryptography and data compression.
Speed Up Deep Learning Inference
Intel® Deep Learning Boost (Intel® DL Boost) enhances Intel AVX-512 with Vector Neural Network instructions (VNNI). Workloads like image classification, speech recognition, language translation or object detection can significantly benefit from this technology.
Get Even Higher Performance When You Need It Most2
Intel® Turbo Boost Technology automatically allows processor cores to run faster than the rated operating frequency if they’re operating below power, current, and temperature specification limits. The newly launched Amazon EC2 M5zn instances offer the fastest Intel® Xeon® Scalable processors in the cloud, with an all-core turbo frequency up to 4.5 GHz.
2. Explore Intel’s World-Class Developer Tools
Developers can employ Intel® architecture features to optimize scale-up and scale-out application performance. Intel’s unified, standards-based programming model, Intel® oneAPI (beta), offers:
- Analysis and debug tools to help find code optimizations and continually improve codes generation over generation.
- Optimized AI and machine-learning frameworks like TensorFlow, PyTorch and others take advantage of Intel architecture features such as Intel AVX 512 and Intel DL Boost.3
- Cross-architecture deployment using the Intel® Distribution of OpenVINO™ Toolkit.
- Increase performance of data science and AI frameworks using the Intel® Distribution for Python. Discover how the Intel Distribution for Python can speed up experimentation with predictable behavior.
- Intel® DevCloud lets developers try lots of technologies in a free development sandbox.
Read Industry Reviews of the Intel OneAPI Industry Initiative
3. Deploy Your Solution Easily
Hardware and software packaged into a standard cluster architecture.
Accelerate and simplify how users acquire and run HPC solutions.
Use a common interface to support the broad application needs of the target platform with the Intel® HPC Platform Specification.
Intel and AWS have collaborated to launch an Intel® Select Solution for Simulation and Modeling using AWS ParallelCluster to create an optimized solution to execute HPC workloads.
4. Additional Features Accelerate Deployment and Increase Return on Investment
- Developers and users have limited time to figure out optimizations and solution integration. They need an easier way.
- Intel and AWS have eliminated this barrier to cloud-based HPC by fine-tuning optimizations to increase performance using AWS Elastic Fabric Adapter and the Intel® MPI Library.
- Building on the Open Fabrics interface, this combination provides 100 GbE fabric performance with the low latency required to scale out application workloads across many instances.
- The solution is packaged using AWS ParallelCluster.