These chips want to revolutionize and democratizing visual computing

Visual Computing

I may not sound much the name ofMovidius, but this is the companyresponsiblefor the production of chipsMyriadwho have been instrumental in the famousTango Project.We have had the opportunity to talk to those responsible at the Mobile World Congress and we have to understand where technology is aimed.

Jack Dashwood and Gary Brown, responsible for the marketing department, explained to us the origins of a company that started directing his steps toward the high-performance computing and has taken a radical turn to do one thing, but doing it very well.The idea:to democratize visual computing, and all sorts of devices take a performance hitherto inaccessible to the market.

VPUs not GPUs

Project Tango was really just a demonstration of where you could go in the footsteps of Myriad chips, which they define asVisual Processing Units and contrasting two types of more generic processors, CPUs and GPUs that we use in day by day both in our desktops and laptops and our mobile devices.

In Movidius they have advanced significantly since that first edition of these chips:the new Myriad 2and have 12 processing cores intended specifically related to visual computing tasks: 3D Sensing, recognition depth (similar to what we saw with Kinect), gesture recognition or recognition and detection of objects between them.

All these tasks have the peculiarity of being able torun in parallelon these chips that have benefited from a reduction in the manufacturing process and the micro architecture developed by the company itself was born in Ireland and end shortly after moving its headquarters to Silicon Valley.

Visual Computing
Image Source: Google Image

Visual computing with energy-efficient

As we explained those responsible, it is true that a GPU has certain similarities with the concept that their VPUs are based, but the clear difference is the energy efficiency: the Myriad have anenergy roof just 1.2Wmaking that can run in parallel to both the CPU and the GPU. Relieving them of the work for which these general-purpose processors (or at least not so specific) are prepared.

You may also like to read another article on DuckysDesktop: The seven best projects with a Raspberri Pi to build your own gaming console

The fields of application of these chips are very numerous, but for example away in the beginning of scenarios such as the autonomous cars: neither the size nor the energy consumption requirements are particularly limited, and Movidiusdirect their steps towards other fields.

Thevirtual realityis one of those fields, and these developments Movidius avoid e.g. having to rely on external cameras have to be calibrated to enjoy an adequate virtual environment.Something similar occurs with increased and certainly the realitydrones: careful because these chips provide the ability todetect objects and avoid collisions, which is crucial for the future of some unmanned vehicles that could theoretically take to the skies in the not too distant future.

The company could not give details on specific products that were working with several partners, but could confirm thattwo of them will appear before the end of the first quarter of the year.We look forward, it is evident that Movidius and chips Myriad could have a major impact in areas that generate so many expectations and where visual computing and combining them with areas such as neural networks and Deep Learning promise many (and good) surprises.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.