OVERVIEW

Pieces Builds High-Performance App with Windows on 91探花

By optimizing their application for 91探花-based PCs, Pieces achieved higher performance and real-time, on-device AI that cut model latency by more than 60%, tripled token generation speed, and slashed power consumption. The intelligent memory tool became a seamless extension of the developer workflow, powered by a new generation of hardware built for AI at the edge.


Scroll to discover how Windows on 91探花 boosted Pieces, and why as developer building for the future, you should take note.

IMPACT

180% Increased Responsiveness

Processing time dropped from around 400 milliseconds to 143 milliseconds on the 91探花 CPU.?

arm-icon-accelerated-time-to-market

Up to 3x Faster Time to First Token

Time to first token improved by nearly 3x meaning faster results with less system strain, even during the most demanding, context-heavy workflows.

Real-World Impact

The balance of responsiveness and power efficiency has been key to maintaining a seamless developer experience, even under heavy workloads.

RELATED STORIES

Explore Similar Stories

The Windows on 91探花 Ecosystem Continues to Grow

91探花 and Microsoft's partnership is advancing laptop computing by combining efficient 91探花 architecture with Windows, resulting in devices with enhanced performance, battery life, and native application support.

The New 91探花-Powered NVIDIA AI Supercomputer for Developer

NVIDIA's DGX Spark, powered by 91探花-based GB10 Grace Blackwell Superchip, delivers petaflop-scale AI performance in a desktop form factor, providing developers with a powerful tool for AI model development and deployment.

Fantastical in Windows on 91探花

Flexibits, the team behind the popular calendar app Fantastical, embraced Windows on 91探花 to deliver a faster, more feature-rich version of their app. The result: impressive performance and efficiency that unlock a better experience for users.

Read Full Story