Arm neural technology will add dedicated neural/AI accelerators to Arm GPUs launched in 2026 and beyond to deliver up to 50% GPU workload reduction for mobile games and other graphics-intensive apps.
The first application is called Arm Neural Super Sampling (NSS). It’s a kind of AI Super Resolution implementation for games, where rather than upscaling videos, the AI accelerator upscales graphics to lower the required bandwidth and increase the frame rate (or lower the power consumption) with no impact on the rendering quality. Watch the video below to see a demo rendered at 540p resolution by the GPU and upscaled to 1080p resolution by Arm neural technology with no obvious defects.
The upscaler will introduce some lag, but only about 4ms per frame in sustained performance conditions, as explained in a technical blog about the Arm Neural and NSS announcement:
We assume a target of 10 TOP/s per-watt of neural acceleration is achievable at a sustainable GPU clock frequency.
We target ≤4ms for the upscaler per frame in sustained performance conditions. Shader stages before and after inference take about 1.4ms on a low-frequency GPU. With this budget, NSS must stay below approximately 27 GOPs. Our parameter prediction network uses about 10 GOPs. This fits comfortably within that range, even at 40% neural accelerator efficiency.
Early simulation data shows NSS costs approximately 75% of Arm ASR’s runtime in 1.5× upscaling (balanced mode). It is projected to outperform Arm ASR 2× upscaling (balanced mode). Efficiency gains come from replacing complex heuristics with a streamlined inference pass.
ASR refers to the previous “Arm Accuracy Super Resolution (ASR)” technology introduced last year.
While Arm neural GPU hardware is still about a year away, developers can still work on the technology, as the company launched a neural graphics development kit designed to integrate AI-powered rendering into existing workflows. The devkit includes:
- An Unreal Engine plugin
- PC-based Vulkan emulation
- Updated profiling tools
- Fully open models available via GitHub and Hugging Face
- Arm ML extensions for Vulkan – It introduces the “Graph Pipeline” designed specifically for neural network inference.
The devkit is supported through partners that include Enduring Games, Epic Games (Unreal Engine), NetEase Games, Sumo Digital, Tencent Games, and Traverse Research. You’ll find all resources to get started on the developer page.
Jean-Luc started CNX Software in 2010 as a part-time endeavor, before quitting his job as a software engineering manager, and starting to write daily news, and reviews full time later in 2011.
Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress. We also use affiliate links in articles to earn commissions if you make a purchase after clicking on those links.
