When the AI boom came a-knocking, Intel wasn't around to answer the call. Now, the company is attempting to reassert its authority in the silicon business by unveiling a new family of chips designed especially for artificial intelligence: the Intel Nervana Neural Network Processor family, or NNP for short. The NNP family is meant as a response to the needs of machine learning, and is destined for the data center, not your PC. Intel's CPUs may still be a stalwart of server stacks (by some estimates, it has a 96 percent market share in data centers), but the workloads of contemporary AI are much better served by the graphical processors or GPUs coming from firms like Nvidia and ARM. Consequently, demand for these companies' chips has skyrocketed.
It's been nearly a year since I published my first Special Report on artificial intelligence and urged readers to buy the processor maker NVIDIA (NVDA) at $68.80. With US annual auto production at 17 million, and global car and commercial vehicle production at a record 94.64 million, that is a lot of processors. All new AI startups comprise small teams of experts from private labs and universities financed by big venture capital firms like Sequoia Capital, Kleiner Perkins, and Andreeson Horowitz. The global artificial intelligence market is expected to grow at an annual rate of 44.3% a year to $23.5 billion by 2025.
Neural networks are "slow" for many reasons, including load/store latency, shuffling data in and out of the GPU pipeline, the limited width of the pipeline in the GPU (as mapped by the compiler), the unnecessary extra precision in most neural network calculations (lots of tiny numbers that make no difference to the outcome), the sparsity of input data (lots of 0s), and many many other factors. How can we make deep neural network training, testing, and predictions faster? One way is to write faster algorithms, like the relu activation function, which is much faster than tanh and sigmoid, and another is to write better compilers to map the neural network into the hardware. Way back in 2011, my collaborators and I built custom processors on FPGAs to speed up neural network computations.
Recently, analyst Trip Chowdhry of Global Equities Research wrote in an investor note that Wal-Mart Stores (NYSE:WMT) will ramp up its focus on deep neural networks for its OneOps cloud business and that the retailer will tap NVIDIA's (NASDAQ:NVDA) graphics processing units (GPUs) to make this happen. Deep neural networks, and the broader deep learning segment, are part of a growing artificial intelligence market. Additionally, NVIDIA said in its second-quarter fiscal 2018 report that it forged new partnerships with Microsoft, Google, Tencent, IBM, Baidu, and Facebook to help them bring new deep learning and artificial intelligence services online. Aside from NVIDIA's deep-learning total addressable market, adding more of these customers is important, because the company's data center revenue segment (which includes GPU sales for deep-learning technologies) is becoming a larger part of the business.
In 2015, the company revealed the Nvidia Drive PX -- the world's first in-car super computer. "That has now snowballed into them being basically the leader for autonomous driving hardware for the auto industry," said Egil Juliussen, director of research and principal analyst for automotive technology at IHS Markit. Intel, so far Nvidia's biggest competitor in self-driving cars, plans to have autonomous BMWs powered by Intel processors in production by 2021. Nvidia already has more than 80 automakers, software firms, transportation network providers and other companies using its chips to develop self-driving car technology, and last month Nvidia added Volvo as one of its newest partners.
Chipmakers such as Nvidia Corporation (NVDA), Intel Corporation (INTC), Advanced Micro Devices, Inc. (AMD) and ARM Holdings could face significant headwinds as their general-purpose processors lose market share to hardware designed for specific tasks, according to Barron's. Thus far, these chips have outperformed similar ones produced by Nvidia and Intel when used for specific tasks such as machine learning. The TPU is more specialized than general-purpose chips, making it better-designed to complete specific tasks, according to Barron's. Not everyone is convinced of Patterson's point of view, as some think that designing processors for specific tasks will confine them to niche functions, according to Barron's.
Nvidia has released a new version of TensorRT, a runtime system for serving inferences using deep learning models through Nvidia's own GPUs. Serving inferences from GPUs is part of Nvidia's strategy to get greater adoption of its processors, countering what AMD is doing to break Nvidia's stranglehold on the machine learning GPU market. Nvidia claims the GPU-based TensorRT is better across the board for inferencing than CPU-only approaches. One of Nvidia's proffered benchmarks, the AlexNet image classification test under the Caffe framework, claims TensorRT to be 42 times faster than a CPU-only version of the same test -- 16,041 images per second vs. 374--when run on Nvidia's Tesla P40 processor.
TAIPEI: Chipmakers switched focus at Taiwan s top tech fair this week with bets on new areas such as driverless cars, virtual reality and artificial intelligence, shifting away from smartphones where intense competition has pushed down components prices. A push to boost artificial intelligence processors, a key technology behind driverless cars, has been a prominent part of the fair. Nvidia Corp, a visual computing company, focused on its Volta Graphics Processing Unit, the product of $3 billion investment in research and development. Nvidia is part-owned by Japan s SoftBank Group Corp, which last month announced its Vision Fund - investing in technologies such as artificial intelligence and robotics - had raised $93 billion.
It's a big accomplishment for Alphabet (NASDAQ:GOOGL) (NASDAQ:GOOG). To match the intuitive skills of human players, programmers taught AlphaGo pattern recognition. The latest version of AlphaGo that beat number-one-ranked Ke Jie is even more impressive than the one that defeated legendary player Lee Sedol last year. The Motley Fool owns shares of and recommends Alphabet (A shares), Alphabet (C shares), and Nvidia.
This chip may operate along the same lines as Google's (NASDAQ:GOOG) Tensor Processing Unit (TPU) that accelerates certain specialized mathematical operations useful in AI. While he avoided revealing any specific Apple initiatives, his comments reveal his concerns that AI software requires enormous computing power. Gurman's article incorrectly states that Qualcomm (NASDAQ:QCOM) already has hardware AI acceleration in its Snapdragon 835. Qualcomm's Snapdragon Neural Processing Engine that Gurman refers to is in fact a set of software APIs.