Week In Review: Design, Low Power

IP for TSMC 5nm, 22nm; MIPI RFFE v3.0 IP; Nvidia’s new GPU.


Tools & IP
Synopsys released a range of IP for TSMC’s 5nm process technology. It includes interface PHY IP such as 112G/56G Ethernet, Die-to-Die, PCIe 5.0, CXL, and CCIX; memory interface IP for DDR5, LPDDR5, and HBM2/2E; die-to-die PHYs for 112G USR/XSR connectivity and High-Bandwidth Interconnect; and foundation IP including logic libraries, multi-port memory compilers, and TCAMs.

SmartDV launched both design and verification IP for the MIPI RF Front End Control Interface (MIPI RFFE) v3.0 specification. The portfolio includes simulation IP, assertion IP, post-silicon validation IP and SystemC models, along with RFFE master and slave Design IP. Synthesizable Transactors for accelerating SoC testing on hardware emulators or FPGA prototyping platforms is also part of the portfolio.

CAST released a 100Gbps UDP/IP core. It includes a complete UDP/IP hardware stack, with support for related networking standards and functions (such as ARP with cache, IGMP, and VLAN). It operates in stand-alone mode, offloading the demanding task of UDP/IP encapsulation/decapsulation from the system processor while transmitting and receiving data at speeds up to 100Gbps.

Arasan Chip Systems uncorked MIPI D-PHY IP supporting speeds of up to 2.5 Gbps on TSMC 22nm ultra-low power (22ULP) and 22nm ultra-low leakage (22ULL) processes. The MIPI D-PHY IP is integrated with Arasan’s CSI Tx, CSI Rx, DSI Tx and DSI Rx IP.

Dolphin Design announced a portfolio of IP for TSMC 22nm ultra-low leakage (22ULL) process targeted at battery-operated IoT devices. It includes power management IP, including low-leakage and quiescent LDOs and DC/DC converter, as well as audio IP including ADC/DAC/codecs and voice activity detection solution.

Nvidia made a slew of announcements along with the company’s virtual GTC keynote. Primary among them was its new GPU architecture, Ampere. The new architecture aims to unify AI training and inference and boost performance by up to 20x over its predecessors. It adds automatic mixed precision and support for both Tensor Float (TF32) and Floating Point 64 (FP64). The first Ampere GPU is the A100, a universal workload accelerator built for AI, data analytics, scientific computing and cloud graphics. It can be partitioned into as many as seven independent instances for inferencing tasks or combine with other A100s as a single GPU.

Nvidia’s EGX Edge AI platform was expanded with the addition of the EGX A100 for larger commercial off-the-shelf servers (which incorporates newly-acquired Mellanox technology and is and based on the Ampere architecture) and the credit card-sized EGX Jetson Xavier NX for micro servers and edge AI. And with health on the forefront of many minds, the company’s Clara healthcare platform was updated for faster genomic sequencing, new AI models, and integration of sensors for smart hospitals.

Picocom licensed Arteris IP’s FlexNoC Interconnect IP for use in its upcoming 5G New Radio (5G NR) small cell baseband SoC. Picocom, which specializes in 5G wireless baseband technologies for open small cell radio access networks, cited the ability to manage on-chip bandwidth and complexity and retain design flexibility.

GigaDevice inked a patent licensing agreement with Rambus giving GigaDevice, a non-volatile memory and 32-bit MCU maker, access to more than 180 resistive RAM (RRAM) technology-related patents and applications from Rambus and Reliance Memory (a joint venture between Rambus, GigaDevice and several strategic investment partners).

Arm TechCon has been renamed Arm DevSummit and will be held virtually October 5-9, 2020. The call for papers and presentations is now open through June 9.

DAC will be a virtual event this year due to the COVID-19 pandemic. It will still take place July 19 – 23, 2020. More details on the new virtual format will be available at a later date.

Many conferences have now been cancelled, postposed, or moved online. Find out what’s happening with each at our events page. How about checking out a webinar instead?

If you have an event planned but not sure how to start moving it online, ACM published a report detailing some best practices for virtual conferences. It discusses the tasks required of organizers, platforms, and financial considerations, alongside examples of conferences that have gone virtual and lessons learned from their experiences.

Leave a Reply

(Note: This name will be displayed publicly)