NEUCHIPS' Purpose-Built Accelerator Designed to Be Industry's Most Efficient Recommendation Inference Engine
Team comes out of Stealth to Introduce Innovative AI Inference Platform
LOS ALTOS, Calif., June 2, 2022 - NEUCHIPS is excited to announce its first ASIC, RecAccelTM N3000 using TSMC 7nm process, and specifically designed for accelerating deep learning recommendation models (DLRM). NEUCHIPS has partnered with industry leaders in Taiwan's semiconductor and cloud server ecosystem and plans to deliver its RecAccel™ N3000 AI inference platform on Dual M.2 modules for Open Compute Platform compliant servers as well as PCIe Gen 5 cards for standard data center servers during the 2H'2022.
"In 2019, when Facebook open sourced their Deep Learning Recommendation Model and challenged the industry to deliver a balanced AI inference chip platform, we decided to pursue the challenge," said Dr. Lin, NEUCHIPS CEO, Co-Founder of Global Unichip Corp, subsidiary of TSMC and Professor at National Tsing Hua University, Taiwan. "Our continued improvements in MLPerf DLRM benchmarking and whole-chip emulation give us confidence that our RecAccel™ AI hardware architecture co-designed with our software will scale to deliver industry leadership and exceed our target of 20M inferences per second at 20 Watts."
NEUCHIPS RecAccel™ N3000 Inference platform includes sophisticated hardwired accelerators, patented query scheduling and a comprehensive software stack optimized to provide high accuracy and hardware utilization while maintaining energy efficiency required in data centers. Other key features include the following:
- Proprietary 8-bit coefficient quantization, calibration and hardware support that deliver 99.95% of FP32 accuracy.
- Patented embedding engine with novel cache design and DRAM traffic optimization that reduces LPDDR5 access by 50% and increases bandwidth utilization by 30%.
- Dedicated MLP compute engines that deliver state-of-the-art energy efficiency at engine level, and 1 microjoule per inference at SOC level.
- Proven software stack that delivers very high scalability across multiple cards.
- Support for leading recommender AI models including DLRM, WND, DCN, and NCF.
- Robust security based on hardware root of trust.
About Neuchips:
NEUCHIPS develops purpose-built AI inference chip platforms from the ground up, by co-developing hardware and software to meet our customer's requirements for performance, accuracy, power, and cost-efficiency. NEUCHIPS is a founding member of MLCommons™. For more information, please visit https://www.neuchips.ai or contact contact@neuchips.ai
|
Neuchips Hot IP
Related News
- NEUCHIPS Secures $20 Million in Series B2 Funding to Deliver AI Inference Platform for Deep Learning Recommendation
- Xilinx Launches Alveo U55C, Its Most Powerful Accelerator Card Ever, Purpose-Built for HPC and Big Data Workloads
- Flex Logix Announces Working Silicon Of Fastest And Most Efficient AI Edge Inference Chip
- NEUCHIPS Announces World's First Deep Learning Recommendation Model (DLRM) Accelerator: RecAccel
- P.A. Semi Successfully Develops the Most Power-Efficient High-Performance Processor Ever Designed
Breaking News
- VeriSilicon introduces AcuityPercept: an AI-powered automatic ISP tuning system
- Avant Technology Partners with COSEDA Technologies to Enhance System-Level Software Solutions
- intoPIX Powers Ikegami's New IPX-100 with JPEG XS for Seamless & Low-Latency IP Production
- Tower Semiconductor and Alcyon Photonics Announce Collaboration to Accelerate Integrated Photonics Innovation
- Qualcomm initiates global anti-trust complaint about Arm
Most Popular
- Qualcomm initiates global anti-trust complaint about Arm
- Sarcina Technology launches AI platform to enable cost-effective customizable AI packaging solutions
- EnSilica Agrees $18m 7 Year Design and Supply ASIC Contract
- Siemens to accelerate customer time to market with advanced silicon IP through new Alphawave Semi partnership
- Tower Semiconductor and Alcyon Photonics Announce Collaboration to Accelerate Integrated Photonics Innovation
![]() |
E-mail This Article | ![]() |
![]() |
Printer-Friendly Page |