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.

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

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

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

The Streamline CLI Tools are free cloud-native performance analysis tools for software running on Arm Neoverse  CPUs. Read this blog to learn more.

In this blog, we show how Alibaba Cloud customers can reduce LLM Inferencing cost per token by using Yitian 710-based instances.

Samsung and Arm have announced a strategic partnership to develop next-generation communication technologies. Read more here.

In this blog post, we show how sentiment analysis can be added to existing applications using Hugging Face and PyTorch models running on AWS Graviton3.

BOLT is a post-link optimization technology enabling performance improvement for various workloads. Read more in this post.

In this blog post, we demonstrate the performance and cost effectiveness of using Arm Neoverse-based CPUs for smaller large language models (LLMs) like Llama3.

In this blog, we explore the latest release of Red Hat Enterprise Linux (RHEL) with improved support for Arm architecture.

Dataplane stack is an open source reference solution for building networking application and user cases with optimal configuration on Arm Neoverse platform.

This blog illustrates how to enable BOLT on arm platform and the performance uplift after enabling it.

Firewalls need to be scalable, efficient, and cost-effective. The whitepaper shows the performance optimizations to make it effective.

In this blog, Capgemini shares the cost savings and carbon reduction benefits possible through their AWS Graviton workload migration assessment.

Read about how Arm builds with Docker Build Cloud.

In this blog, we highlight the benefits of Neoverse S3 System IP as a foundation for building custom silicon.

In this blog, we highlight the benefits of Neoverse CSS N3 for developing performance-per-watt optimized custom silicon for cloud-to-edge markets.

In this blog, we highlight the benefits of Neoverse CSS V3 for developing TCO-optimized, Confidential Compute-enabled custom silicon for cloud, HPC, and AI/ML use cases.