Exploring 6 Key Arm Innovations for August 2024

The Arm Editorial Team is always on the lookout for the latest most exciting and cutting-edge technology trends. In this edition of “Beyond the Arm Newsroom”, we highlight 6 groundbreaking innovations at Arm from August 2024. These include advancements in scaling IoT edge deployment, optimizing Spark SQL for power and efficiency, and boosting performance in media processing, HPC, and machine learning.

Revolutionizing Edge Computing with Scalable and Efficient Hybrid Processors

Arm advances edge computing with hybrid systems, combining Cortex-A and Cortex-M/R processors to make edge deployment more streamlined and scalable. Security Research Engineer Basma El Gaabouri details how applications can be deployed on Cortex-M from Cortex-A using cloud-native tools to simplify firmware updates and use resources more efficiently. This method ensures secure, scalable firmware updates, allowing applications to run on different cores based on their needs, improving resource use, and energy efficiency – all while simplifying management across varied hardware systems.

Supercharging Spark SQL on Arm with Meta’s Velox Engine

Apache Spark is designed to make big data processing quicker and easier. By leveraging the techniques implemented by the Meta-led Velox vectorized execution engine, users can achieve an effective solution for the Spark SQL accelerator on the Arm platform. Staff Software Engineer Yuqi Gu highlights why this advancement makes Spark SQL more efficient and powerful, improving performance and enabling the use of hardware accelerators like GPUs and FPGAs.

Matrix-Matrix Multiplication Boosts Media and ML Performance

Matrix multiplication refers to the operation of multiplying two matrices together to obtain a third matrix. To boost performance in media processing, HPC, and machine learning applications, matrix-matrix multiplication uses three powerful Arm technologies: Neon, SVE, and SME. Information Developer Khalid Saadi uses a side-by-side comparison to highlight key differences and offer insights to help developers transition from Neon or SVE/SVE2 to SME/SME2.

Leveraging Kubernetes for High-Performance IoT Edge Deployments

Kubernetes is the leading open-source container orchestration software built to automate software deployment, scaling, and management at the cluster level. Principal Research Software Architect Alexandre Peixoto Ferreira describes a compute capacity model tailored for Kubernetes that ensures efficient resource allocation in heterogeneous environments. Arm-based edge solutions enhance Quality of Service (QoS) management, crucial for developers working with multi-application systems in edge and cloud computing.

Harnessing Heterogeneity to Enhance Design for Edge Computing

A major challenge in advancing the Intelligent Edge is managing hardware heterogeneity. Principal Research Software Architect Alexandre Peixoto Ferreira details how hardware heterogeneity expands the design possibilities for edge computing. Solutions like reference cores and hybrid runtime deployments are proposed to mitigate complexity. These approaches aim to ensure portability, ease of programming, and manageable deployment, ultimately improving the efficiency and functionality of edge computing systems.

Demonstrating Hybrid Container Runtime for IoT Edge Applications

Implementing a container runtime for cloud-native deployment on high-performance IoT edge devices allows applications to be deployed on additional processors within a system. In edge environments, cloud-native technology helps tackle the challenges of developing, debugging, deploying, and securing software for these devices. Principal Engineer Chris Adeniyi-Jones demonstrates how deploying a new application component on the Arm Cortex-M4 processor within an SoC can lead to more energy-efficient computing.

Exploring Vulkan SDK, LLM Inference, and Advancing Mobile Graphics

In the latest episode of Innovation Coffee, where we go deep into the Arm developer ecosystem, Karen Ghavam and Richard Wright Jr. from LunarG discuss the Vulkan SDK, Jakub Sujak demonstrates the Learning Path for LLM inference on Android with KleidiAI, MediaPipe, and XNNPACK, Sam Martin covers advancements in mobile graphics with Arm, and Pareena Verma introduces the latest Learning Paths.

Building Computer Vision Applications with OpenCV on Android Devices

Arm Developer Evangelist Gabriel Peterson takes you through the fundamentals of building a computer vision application on Android using OpenCV. Based on the Learning Path by Dawid Borycki, this tutorial serves as an excellent introduction to OpenCV, with advice for adding support to a project, and processing images.