Blogs

The CPU Utilization metric can be misleading when comparing Arm Neoverse processors with SMT-enabled architectures. This blog post explains why and how to address it.

Nairobi's hardware development scene is discussed by Arm Ambassador, Kevin Odero.

Read about the Arm Developer Community showcase event from Arm Ambassador, Fidel Makatia.

In this blog post, we show the steps for running a Llama 3.1 demo on AWS Graviton4, and how Arm Kleidi Technology can be used to improve the performance of LLMs running on PyTorch.

In this blog post, we review Kleidi Technology contributions Arm has made to the PyTorch framework and how these greatly improve PyTorch inference performance.

In this blog post, we show how Kleidi Technology improves the performance of Automatic Speech Recognition on Neoverse N2-based Alibaba Yitian 710 CPUs, outperforming x86-based options.

In this blog post, we explore a neural network based solution for improving low light video and optimizing for mobile.

In this blog post, we explore how the Arm KSA Framework can help optimize learning and we describe some of its possible applications.

This blog post highlights that G8Y outperforms G7, showing up to a 36% improvement in Elasticsearch performance on Alibaba Cloud Yitian 710 instances.

OpenRNG the new open-source RNG optimised for Arm; easily port apps with a drop-in replacement for VSL RNG calls. Get the best performance on Arm.

This blog post presents the KSA framework. It aligns with renewed government, industry, and educational institutions’ interest in workforce development in the semiconductor sector.

In this blog post, we show how a hybrid runtime and k3s can be used to deploy an application onto an edge platform that includes an embedded processor.

This blog post identifies heterogeneity as an opportunity to create better edge computing systems.

In this blog post, we address heterogeneity in IoT edge deployments using Kubernetes.

The hybrid runtime allows the deployment of software onto embedded processors within a single SoC using cloud-native orchestration tools, such as k3s.

This blog post introduces an accelerator for Spark SQL. Experimental results highlight the potential of this approach for Spark SQL on Neoverse N2.

This blog post describes how to implement the same matrix-matrix multiplication algorithm using three different Arm technologies: Neon, SVE, and SME.

In this blog post, we show how AWS Graviton4 delivers leadership HPC performance across a range of HPC workloads compared to AWS Graviton3E and 4th Gen AMD EPYC.

Apache Flink is a framework for stateful computations over data streams, performance benchmark shows Yitian 710 outperforms Ice Lake significantly.

Demonstrating the performance benefits of the NPU by running two TinyML applications on a modern microcontroller that is equipped with an Arm Ethos-U55 NPU.

Developing AI applications for camera trap data to non-invasively monitor bears and other wildlife.