Nvidia has been leading the new neural network charge for the last few years, even releasing dedicated chipsets for autonomous cars, but Intel has decided not to let the GPU maker run away alone with the whole market.
At the Wall Street Journal’s D.Live event, Intel CEO Brian Krzanich announced today that Intel will soon be shipping the world’s first family of processors designed from the ground up for artificial intelligence (AI): the Intel® Nervana™ Neural Network Processor family.
Intel says the Nervana NNP is a purpose-built architecture for deep learning, designed to free users from the limitations imposed by existing hardware, which wasn’t explicitly designed for AI.
The chips feature:
- New memory architecture designed for maximizing utilization of silicon computation
The Intel Nervana NNP does not have a standard cache hierarchy and on-chip memory is managed by software directly, allowing better memory management which enables the chip to achieve high levels of utilization of the massive amount of compute on each die. This translates to achieving faster training time for Deep Learning models.
- New level of scalability for AI models
Designed with high speed on- and off-chip interconnects, the Intel Nervana NNP enables massive bi-directional data transfer. A stated design goal was to achieve true model parallelism where neural network parameters are distributed across multiple chips. This makes multiple chips act as one large virtual chip that can accommodate larger models, allowing customers to capture more insight from their data.
- High degree of numerical parallelism: Flexpoint
Neural network computations on a single chip are largely constrained by power and memory bandwidth. To achieve higher degrees of throughput for neural network workloads, in addition to the above memory innovations, Intel invented a new numeric format called Flexpoint. Flexpoint allows scalar computations to be implemented as fixed-point multiplications and additions while allowing for large dynamic range using a shared exponent. Since each circuit is smaller, this results in a vast increase in parallelism on a die while simultaneously decreasing power per computation.
- Meaningful performance
Intel says the chip is the first of a series of silicon aimed at achieving a 100x increase in deep learning training performance by 2020.
In designing the chip, Intel worked with Facebook to gain their technical insights with the aim of nurturing a new class of AI applications that are limited only by the imagination of developers.
Intel says cognitive and artificial intelligence (AI) technology are on a trajectory to reach $46 billion in industry revenue by 20201, with Intel’s first-generation Intel® Nervana™ Neural Network Processor to join the party by the end of this year.
Source: Intel Newsroom