TechWatch

First look at the Arm Machine Learning Processor

Arm chose this year’s IEEE Hot Chips Symposium to show off their first-generation Machine Learning Processor claiming a design that delivers ‘massive’ efficiency improvements over GPU, CPU, or DSP implementations as well as the ability to scale from IoT to server applications. Being an English design, it probably serves tea in a porcelain cup, as well. Leaving aside the issue ...

Peter McGuinness

Arm chose this year’s IEEE Hot Chips Symposium to show off their first-generation Machine Learning Processor claiming a design that delivers ‘massive’ efficiency improvements over GPU, CPU, or DSP implementations as well as the ability to scale from IoT to server applications. Being an English design, it probably serves tea in a porcelain cup, as well. Leaving aside the issue of whether any architecture can deliver all of that, the implementation described is aimed explicitly at mobile applications and delivers a theoretical MAC throughput of 4096 OPS per clock for a 4 TOPS machine if running at 1 GHz, which
...

Enjoy full access with a TechWatch subscription!

TechWatch is the front line of JPR information gathering service, comprising current stories of interest to the graphics industry spanning the core areas of graphics hardware and software, workstations, gaming, and design.

A subscription to TechWatch includes 4 hours of consulting time to be used over the course of the subscription.

Already a subscriber? Login below

This content is restricted

Subscribe to TechWatch