1-112Gbps Medium Reach (MR) and Very Short Reach (VSR) SerDes
ARM, IBM team on low power analog AI chip
By Nick Flaherty, eeNews Europe (September 9, 2022)
Researchers at ARM and IBM have developed a 14nm analog compute in memory chip for low power always on machine learning.
These always-on perception tasks in IoT applications, dubbed TinyML, require very high energy efficiency. Analog compute-in-memory (CiM) using non-volatile memory (NVM) promises high energy efficiency and self-contained on-chip model storage.
However, analog CiM introduces new practical challenges, including conductance drift, read/write noise, fixed analog-to-digital (ADC) converter gain, etc. These must be addressed to achieve models that can be deployed on analog CiM with acceptable accuracy loss.
Researchers from ARM and IBM Research Zurich looked at the TinyML models for the popular always-on tasks of keyword spotting (KWS) and visual wake words (VWW). The model architectures are specifically designed for analog CiM, and detail a comprehensive training methodology, to retain accuracy in the face of analog issues and low-precision data converters at inference time.
E-mail This Article | Printer-Friendly Page |
Related News
- Blumind Harnesses Analog for Ultra Low Power Intelligence
- TI delivers leading analog integration, best-in-class low power and floating-point performance in new Stellaris ARM Cortex-M4F microcontrollers
- ARM Artisan Low Power IP Offered By IBM, Chartered To Support 65-Nanometer Common Platform
- Nordic Semiconductor and Arm reaffirm partnership with licensing agreement for latest low power processor designs, software platforms, and security IP
- NVIDIA Collaborates With SoftBank Corp. to Power SoftBank's Next-Gen Data Centers Using Grace Hopper Superchip for Generative AI and 5G/6G
Breaking News
- Ceva multi-protocol wireless IP could simplify IoT MCU and SoC development
- Controversial former Arm China CEO founds RISC-V chip startup
- Fundamental Inventions Enable the Best PPA and Most Portable eFPGA/DSP/SDR/AI IP for Adaptable SoCs
- Cadence and TSMC Collaborate on Wide-Ranging Innovations to Transform System and Semiconductor Design
- Numem at the Design & Reuse IP SoC Silicon Valley 2024
Most Popular
- GUC provides 3DIC ASIC total service package to AI/HPC/Networking customers
- Qualitas Semiconductor Appoints HSRP as its Distributor for the China Markets
- Siemens collaborates with TSMC on design tool certifications for the foundry's newest processes and other enablement milestones
- Credo at TSMC 2024 North America Technology Symposium
- Huawei Mate 60 Pro processor made on SMIC 7nm N+2 process