Mentor’s nmLVS; Siemens buys Avatar; Analog Devices buys Maxim; Arm rumors

Acquisitions – either officially announced or just rumored about – make up most of our news summary this week. We will then move to some AI chip updates; but first, let’s take a look at one of the EDA announcements that are going to be in the spotlight at this year’s Virtual DAC, running from July 20 to 24.

Early short circuits fixing with Mentor’s Calibre nmLVS-Recon

Mentor has announced the Calibre nmLVS-Recon technology, aimed at speeding overall circuit verification turnaround time by helping designers identify and resolve selected systemic errors early in the development phase. As explained in the announcement’s press release, early design versions typically contain many gross systemic violations. For example, a “shorted nets” class of violation generates millions of errors and is very compute intensive. Circuit verification engineers can use the Calibre nmLVS-Recon short isolation configuration to find and fix these types of violations quickly and efficiently.

Some interesting insights on this new Mentor technology are offered by one of the latest posts from John Cooley, who describes Calibre nmLVS-Recon as “a linter for circuit verification engineers”. Quoting an anonymous source, Cooley adds that “later revs of Recon will supposedly lint for other types of violations, but right now it’s only about finding and fixing shorts.” The reason for focusing on shorts is that – according to a Mentor slide published by Cooley – “as transistor counts move from 30 billion at 7 nanometer to ~70 billion at 5 nanometer, the number of short circuits are exploding.” (…) “Fixing shorts may not sound like a big deal overall, but for LVS debug, 80% of the time is spent chasing and fixing shorts.  Any way to cut that cycle time is a win for your team’s overall schedule.”, Cooley concludes.

Siemens acquires Avatar

Siemens has signed an agreement to acquire Avatar Integrated Systems (Santa Clara, CA), a developer of place and route software for IC design. Siemens plans to integrate Avatar technology with several Mentor products, including the Calibre platform, Tessent software, and Catapult HLS software. According to Siemens, Avatar pioneered a detailed-route-centric architecture that has been built bottom-up on a unified in-memory data model, designed to enable all engines to access full design data and attributes at any time. This empowers each engine (placement, routing, timing, optimization, clock tree synthesis, etc.) to dynamically invoke other engines incrementally. Avatar’s products are built on technologies acquired from ATopTech in 2017. The product line includes Aprisa, a netlist-to-GDS block-level physical implementation tool, and Apogee, a top-level prototyping, floorplanning and chip assembly tool.

Analog Devices acquires Maxim

As jointly announced by the two companies, Analog Devices will acquire Maxim in an all-stock transaction that values the combined enterprise at over $68 billion. The transaction is expected to close in the summer of 2021; upon closing, two Maxim directors will join ADI’s Board of Directors, including Maxim President and CEO, Tunç Doluca. Maxim contribution is expected to strengthen ADI’s analog semiconductor leadership position especially in the automotive, data center and power management markets. The combined enterprise is forecasted to have a revenue of $8.2 billion, more than 50,000 products, more than 125,000 customers, more than 10,000 engineers and approximately $1.5 billion in annual research and development investment. Cost savings will include lower operating expenses and cost of goods sold, by the end of year two subsequent to closing; and manufacturing optimization, by the end of year three subsequent to closing.

SoftBank reportedly exploring Arm sale or IPO

According to The Wall Street Journal, SoftBank “is exploring alternatives including a full or partial sale or public offering” of Arm. For its part, Arm has recently announced proposed organizational changes to strengthen growth and profitability of its core semiconductor IP business. The changes would consist in transferring two IoT Services Group businesses from Arm to SoftBank. As observed by EETimes, “if Arm is heading for either an IPO or being acquired, transferring an operation that is a drag on profitability is a sensible preparatory step.”

Synaptics-Broadcom wireless IoT deal

Synaptics will acquire “certain rights” to Broadcom’s existing Wi-Fi, Bluetooth and GPS/GNSS products and business in the IoT market – as well as future roadmap devices designed in advanced process nodes – for approximately $250 million in an all-cash transaction. Synaptics expects the acquisition to add approximately $65 million in current annualized sales.

STMicroelectronics acquires BeSpoon and Riot Micro

STMicroelectronics will acquire the entire share capital of Ultra Wide Band specialist BeSpoon (Le Bourget du Lac, France) and the cellular IoT connectivity assets of Riot Micro (Vancouver, Canada) with the goal of strengthening the roadmap for its STM32 microcontrollers and secure MCUs. By integrating these technologies, STM32 MCUs will gain accurate indoor positioning and cellular communication capabilities.

AI chip updates

Belgian research institute Imec and Globalfoundries have built a new AI chip based on imec’s Analog in Memory Computing (AiMC) architecture utilizing GF’s 22FDX process, achieving an energy efficiency up to 2,900 TOPS/W. The two companies describe this achievement as a “breakthrough” in AI chips, enabling inference-on-the-edge for low-power devices.

Graphcore has recently unveiled its second-generation Colossus MK2 IPU (Intelligence Processing Unit) targeted at AI workloads in datacenters, the GC200. The company describes its IPU as “a completely new kind of massively parallel processor, co-designed from the ground up with the Poplar SDK, to accelerate machine intelligence.” According to Graphcore, the GC200 is the world’s most complex processor and achieves an 8x step up in real-world performance compared to first generation Colossus MK1 IPU. Built using a TSMC 7nm process technology, the new GC200 has 59.4 billion transistors and 1472 cores, running nearly 9,000 independent parallel program threads, and is equipped with 900MB of In-Processor-Memory. Performance claimed by Graphcore is 250 TeraFlops at FP16.16 and FP16.SR (stochastic rounding).

And Tenstorrent has achieved first-pass silicon success for its Grayskull AI processor SoC using Synopsys’ DesignWare IP portfolio.

Imec’s analog AI chip mounted on the PCB used for measurement and characterization. Image credit: Imec

This entry was posted on Monday, July 20th, 2020 at 6:41 am. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.