Monday, January 2nd 2017

The Laceli AI Compute Stick is Here to Compete Against Intel's Movidius

Gyrfalcon Technology Inc, an emerging AI chip maker in Silicon Valley, CA, launches its Laceli AI Compute Stick after Intel Movidius announced its deep learning Neural Compute Stick in July of last year. With the company's first ultra-low power, high performance AI processor Lightspeeur 2801S, the Laceli AI Compute Stick runs a 2.8 TOPS performance within 0.3 Watt of power, which is 90 times more efficient than the Movidius USB Stick (0.1 TOPS within 1 Watt of power.)

Lightspeeur is based on Gyrfalcon Technology Inc's APiM architecture, which uses memory as the AI processing unit. This eliminates the huge data movement that results in high power consumption. The architecture features true, on-chip parallelism, in situ computing, and eliminates memory bottlenecks. It has roughly 28,000 parallel computing cores and does not require external memory for AI inference.
The standard USB 3.0 Stick can be connected to a range of devices such as PCs, laptops and mobile phones, and can enhance their image-based deep learning capabilities to unprecedented levels without them having to tap into a cloud based system. With Caffe and TensorFlow support, the Stick enables various application development backgrounds.

Powered by a specialized AI Processor, Laceli can be used in many smart applications such as image and video recognition, natural speech understanding, natural language processing, and day-to-day AI applications.

The Laceli AI Compute Stick will be shown in public for the first time in January 2018 at CES 2018 in Las Vegas, Nevada. Lightspeeur 2810S is in production at the moment and is available today to qualified customers.
Add your own comment

2 Comments on The Laceli AI Compute Stick is Here to Compete Against Intel's Movidius

#1
theoneandonlymrk
I am looking forward to a time where buying something like this makes cortana shine at last and my game badies brighter is there a general windows api yet?
Posted on Reply
#2
lexluthermiester
The mining community is going to love this! 20 of these together is equal to a good GPU at a fraction of the power usage(if I'm doing my math right).
Posted on Reply