News

NovuMind Releases Details of Breakthrough AI Chip

Benchmark Tests Show Dramatic Advantages in Latency and Power-Efficiency Achieved by Innovative Design Approach  Santa Clara, CA, USA, October 25, 2018 – NovuMind Inc., a leading innovator in full-stack Artificial Intelligence technologies, today released performance details of its NovuTensor AI chip, with class-leading power efficiency.  NovuTensor, purpose-built for deep neural network inference computations, will expand the frontiers of AI by enabling mass deployment of AI inference models at the edge of the network. NovuMind has produced its first silicon chips and is releasing benchmark test results showing performance superior to the most advanced GPUs currently available for AI inference applications. The benchmark tests were run with a variety of networks including ResNet18, ResNet34, ResNet50, ResNet70, VGG16, and Yolo2. According to Dr. Ren Wu, founder and CEO of NovuMind, “Until now, GPUs have powered the advances in AI, particularly around the training of deep neural network models from large sets of data. Once models are trained, however, the challenge is to deploy them at scale. GPUs and other processors are expensive and consume large amounts of power. Their architectures are optimized for 2-dimensional matrix computation. While they perform well when processing large batches of data, these chips are not suited for real-time…More DetailsNovuMind Releases Details of Breakthrough AI Chip

Ren Wu’s blog

Disrupting the Giants

As we get close to launching our first chip, one of the most frequent questions I get asked is, “How will you guys compete against other AI chips, including those from giant companies like Intel and Nvidia?”  It’s a question I love, because it leads to some great discussions around computing architectures, chip design, and how AI will be deployed to improve peoples’ lives. These can all be lengthy discussions, but for this post let me just focus on some of the design considerations that have led NovuMind to a unique, disruptive approach. We are just at the beginning of an AI revolution, brought about by Deep Neural Networks. As this revolution unfolds, it is not surprising that there are many design approaches, building on existing technologies. The processing units that power most of today’s DNN solutions are CPUs, GPUs, and DSPs.  These are all derivatives of the von Neumann model, based on the concept of executing a computer instruction set, able to execute loops, subroutines, arithmetic functions, and so on. But DNNs are forcing us to think about processor architectures in new ways. DNNs are not organized around the traditional instruction set model. They are more about interconnected layers…More DetailsDisrupting the Giants

News

NovuMind Welcomes John Wei as VP of Design Technologies

Semiconductor Industry Veteran to Lead Initiatives Taking NovuTensor™️ AI Chip to Broad Range of Applications Santa Clara, CA, USA, October 8, 2018 – NovuMind Inc., a leading innovator in full-stack Artificial Intelligence technologies, announced today that Dr. John Wei has joined the company in the role of Vice President of Design Technologies. John will lead full-customization initiatives to ensure that NovuMind can deliver highly customized solutions based on its NovuTensor™️chip. NovuTensoris purpose-built for AI inferencing, with breakthrough performance-power ratios that make it ideal for deployment in embedded applications and edge servers. According to Dr. Ren Wu, CEO and founder of NovuMind, “We are very pleased to welcome John to our team. We already have the most advanced AI inference chip, with a patented design that brings unsurpassed performance-to-power ratios. With John’s new initiatives, we plan to migrate this design to smaller-geometry semiconductor processes that will drive even more performance gains.” “We are at the dawn of an era of massive AI deployment,” continued Dr. Wu. “Advances in deep learning and the semiconductor technology to deploy it will mean that we can put the powers of perception and thinking in all kinds of network-edge applications. We want to provide our NovuTensor…More DetailsNovuMind Welcomes John Wei as VP of Design Technologies

News

NovuMind Awarded Patent for Disruptive AI Chip

Native Tensor Processor is an Industry First, Providing Dramatic Performance Advantages for Deep Neural Networks Santa Clara, CA, USA, October 1, 2018 – NovuMind Inc., a leading innovator in full-stack Artificial Intelligence technologies, announced today that it has been awarded a US Patent for the novel design of its NovuTensor AI chip. NovuTensor is purpose-built for AI inferencing, with breakthrough performance-power ratios that make it ideal for deployment in embedded applications and edge servers. “This patent validates the unique, domain-specific architecture of our NovuTensor chip,” said NovuMind’s founder, Dr. Ren Wu. “We have designed NovuTensor to run convolutional neural network AI models with unsurpassed speed and efficiency. We have taken a disruptive approach, with an architecture that performs 3D tensor computations natively. This gives NovuTensor a huge performance advantage over conventional architectures, which are designed around 2D matrix operations and incur extra overhead when they have to unfold tensors into a series of matrices and perform a series of matrix multiplications.” US Patent No. 10,073,816, “Native Tensor Processor and Partitioning of Tensor Contractions,” was issued on Sep. 11, 2018.  NovuMind has numerous other patent applications pending. Dr. Wu further commented on the state of the AI industry. “With advances in…More DetailsNovuMind Awarded Patent for Disruptive AI Chip