Arm Blog

See how KleidiCV 26.03 improves OpenCV acceleration on Arm CPUs with new optical-flow support, platform updates, and up to 14x gains.

SGLang, the popular high-throughput LLM serving engine, now runs on Arm Neoverse CPUs with W8A8 quantization added for dense and MoE models, and a dedicated Arm64 CI pipeline established for production…

Explore how Arm and Google optimize on-device AI audio workloads to improve performance, efficiency, and user experiences across edge devices

How the Arm supports early developers, inspires learners, and shapes responsible global citizens

Struggling with messy data? Discover how digital cleanup helps both people and AI find the right answers.

Why the rise of enterprise AI agents demands a new class of infrastructure and why the Arm-native stack, anchored by the Arm AGI CPU, is the architecture built for this moment

Optimize exponential functions on Arm with FEXPA, SVE, and polynomial approximation to improve throughput in AI and HPC applications.

ATfL replaces ACfL with an LLVM-based toolchain, delivering optimized performance for AI, HPC, and cloud workloads on Arm platforms

In this blog, we announce perf-libs-sparse, a new open-source project that contains the sparse linear algebra source code included in Arm Performance Libraries.

Selecting kernel build options and maintaining them with a custom Kconfig.

Explore how Arm and CDT Cloud advance embedded debugging in VS Code using open ecosystems, shared tools, and community-driven innovation

Optimize RAG workloads on Arm with Performix, using profiling insights to tune BLAS performance and improve efficiency on Neoverse platforms.

Optimize TensorFlow Lite on AWS Graviton with Arm Performix. Identify bottlenecks, boost throughput by 40%, and reduce ML inference costs.

Optimize Arm workloads with Performix. Detect bottlenecks, apply SIMD, and validate improvements through structured and repeatable analysis workflows.

Find out about Arm contribution to the latest LLVM 22 release: Latest architecture and CPU support, code generation improvements.

Moving Mobile Graphics is a half-day course providing a technical introduction to the state of the art in mobile graphics from practitioners at the forefront

The Arm Neoverse CMN Telemetry Solution is here, explore the whitepaper to learn how to apply the topdown methodology and use the tools for system level performance analysis

Learn how AMBA AXI works in real SoCs. Understand data flow, debug issues, and improve performance and correctness in system design.

Arm Performance Libraries (ArmPL) 26.01 is now available for Linux, macOS and Windows. It can be used to accelerate a wide variety of use cases, including engineering, scientific and machine learning cases

Arm CPU 和开源技术如何借助 faster-whisper 与 vLLM 实现实时、隐私保护的语音对话

Discover The Architecture Speaks, a new AI tool that guides you through the Arm ARM and helps answer complex technical questions.

Explore how NauWork uses AI agents, Snowflake, and Arm-based compute to deliver scalable, cost-efficient talent matching for semiconductor teams

UT Austin have developed ultra-thin e-tattoos using Arm-enabled weightless neural networks for low-power, on-device analysis of vital signs in medical-grade wearables

Cloud‑ready embedded development that is standardized, portable, and built for teams

Arm enables robotics training at scale with Isaac Sim, supporting full AI workflows from build to deployment on a single DGX Spark system

How Arm is unleashing rack-scale performance and efficiency for the agentic AI data center

This release delivers a major upgrade to Vulkan capture, introducing a new capture pipeline built on GFXReconstruct