Kalray Announces the Release of an Efficient Manycore Processing Solution Dedicated to Deep Learning | cross pond high tech | Scoop.it

Roundingout its embedded technology offer, Kalray has expanded its processingcapabilities to the world of artificial intelligence. The company is introducinga highly optimized deep learning solution targeting embedded applications likeautonomous cars, avionics, drones, robotics and more. The solution is capableof supporting all of the most commonly-used deep learning neural networks andframework, such as GoogleNet, Squeezenet, CAFFE and more.

Kalray’s deep learning solution includes:

  • MPPA®2-256 Bostan manycore processor:industry-recognized 288-core processor
  • Kalray Neural Network (KaNN): deep learningsoftware tool used in the development and evaluation of neural networks on MPPA®.KaNN is compatible with all commonly-used deep learning networks.

In terms of pure performance, Kalray is able to leverage the288 cores of its MPPA®2-256 Bostan processor in order to efficiently processnotoriously compute-heavy deep learning algorithms. To do this, the solutionuses the extensive on-chip memory of the Bostan processor and it spreads thecompute-heavy aspects of deep learning – data dependent layers and weightparameters – across the MPPA®’s numerous cores. The result is particularlyefficient processing – up to 60 frames per second while running “GoogleNet”– outperformingthe most efficient GPUs addressing today’s embedded market.