By Duncan Bremner, CTO SureCore Limited
Editor’s note: sureCore just announced availability of its 28nm FD-SOI memory compiler (press release here), which supports the company’s low-power, Single and Dual Port SRAM IP. Here, the company’s CTO explains why this IP is getting such impressive results.
~ ~ ~
Recently, sureCore announced results from a 28nm FD-SOI test chip that showed dynamic power savings exceeding 75% and static power cuts up to 35% (when compared against a number of current commercial offerings), while only incurring a 5-10% area penalty for its ultra-low power SRAM IP.
And while this data is easily substantiated as shown in Figure 1, the sceptical industry pundits have raised questions that fall into two camps: (a) That can’t be done; or (b) How did they manage that? In answer to both of these questions, here’s a quick look at the history and engineering strategy that we adopted to deliver these results.
Looking back to the early days of sureCore, SRAM fascinated us because despite many process iterations, the SRAM in use today bears a striking resemblance to the SRAM architectures that existed in the ’70s and ’80s. We concluded that no one had really taken a “blank-sheet-of-paper” look at the architecture for over 40 years. Recognising the growing importance of power efficiency for SoCs targeting forward-looking applications such as wearables, IoT, and other mobile devices, we examined power consumption in detail, and began by investigating how we could reduce SRAM power to a level attractive to the next generation of power critical, SoC designers.
Our starting point differed significantly from the traditional approach to SRAM R&D that typically starts at the bit cell. We recognised that the basic bit cell is fixed by the foundry; it’s a piece of electronics that is carefully optimised for fabrication. Modern bit cells are designed by the foundries who tend to put an emphasis on the broadest possible manufacturability drivers; yield and faster-time-to-volume as opposed to more performance-centric metrics. Their focus is on the front-end process optimisation, area and yield.
The basic rule of R&D fabless foundry engagement has been, “use the storage array – you won’t get a better packing density.” Consequently, the application use model had become separated from the technology — ‘faster or cheaper’ became the industry’s mantra instead of ‘faster and better’. This resulted in SRAM design teams focusing on how to build more sensitive read amplifiers to detect the signals, and better write amplifiers to drive the signal on to the bit cell. Not much time was spent looking at the fundamental architecture and asking: “Is this the best way?”
sureCore decided to take a more holistic view and stood back from the whole problem. We started with a clean sheet of paper and asked, “Where does the power go when you start storing data on SRAM?”
We discovered that a lot of the power is consumed hauling parasitic capacitance around. Our design strategy was therefore very simple; we developed a system architecture to optimize power while still retaining the area advantages of the standard foundry bit cell.
Simply stated, we architected the internal block architecture of SRAM by splitting the read amplifier function into a local and global read amplifier, thus dividing the capacitive load from the word-line, only driving the areas being addressed and not the whole array. This resulted in significant dynamic power savings during the read cycle. In a similar fashion, we reduced the write cycle power by a similar amount. Whilst hierarchical solutions are not new, the sureCore “secret sauce” is at circuit level developed by our engineering teams leading to not only significant power savings, but also comparable performance levels.
Our “blank sheet” approach delved deep; right down to the fundamental device physics level. Our strategic partners, Gold Standard Simulations — recognised world leaders in modelling devices at the atomic level and experts in nano-scale process nodes, helped us to understand the behaviour and limitations of processes at nodes below 28nm at a device level and bit cell level. Combining this fundamental device understanding with excellent circuit design and system analysis skills, we’ve identified where existing SRAM solutions waste power, and architected our solution to avoid this; we deliver power savings without the added complexity of write and read-assist.
At the outset, we determined it was important that our IP be process-independent. sureCore IP is based on architecture and circuit techniques rather than a reliance on process features. The result of this is technology that can reduce power in standard bulk CMOS, but is equally applicable to newer FinFET or FD-SOI processes and across all geometries, even down to 16nm and below. We believe our approach is paying off and, because we insisted in retaining the foundry optimised bit cell, sureCore’s technology can be retrofitted into existing designs enabling extended product life cycles.
This is our basic technology story… a start-up deciding to take a fresh look at an old technology and dramatically improving power performance over 75% compared with existing solutions. This is a new approach to SRAM power consumption for power sensitive applications and it delivers tangible battery life benefits to both the end user and the FD-SOI designers. Today’s FD-SOI technology is optimised for low power applications, bringing extended battery life to the nascent markets of wearables and IoT.
Gold Standard Simulations Ltd. (GSS) announced a multimillion dollar contract to license its complete TCAD/EDA tool suite to GlobalFoundries (see press release here). The fully integrated and automated tool chain includes GARAND, the GSS ‘atomistic’ TCAD simulator; Mystic, the GSS statistical compact model extractor; and RandomSpice, the GSS statistical circuit simulator. The GSS tool suite is the world’s only fully integrated tool chain that performs simulation-based Design/Technology Co-Optimisation (DTCO) in advanced bulk, FD-SOI and FinFET technologies, including statistical variability and reliability.
“At 14nm FD-SOI is much cheaper, 30-40% cheaper, than Intel’s technology,” Asen Asenov told David Manners in a recent Electronics Weekly post (see full post here). Asenov is CEO and Founder of Gold Standard Simulations (GSS). The subject of the post was how TSMC has turned to GSS for statistical analysis tools. Professor Asenov is a fan of ST’s FD-SOI, noted Manners. The main challenge is building the ecosystem, he concluded.
Targeting low-power SRAM for FD-SOI and FinFETs, UK physical IP start-up sureCore has received a £250K grant (about 292K Euros or $380.5K) from the Technology Strategy Board SMART. Working with the major foundries developing FD-SOI and FinFET technologies, the grant will be used in the development of a demonstrator chip to showcase sureCore’s patented array control and sensing scheme, which significantly lowers active power consumption. Through a combination of detailed analysis and using advanced statistical models, sureCore has designed an SRAM memory consuming less than half the power of existing solutions. SureCore is working closely with Gold Standard Simulations (GSS) Ltd. (GSS Founder/CEO Asen Asenov is a sureCore director).
Following investigations and simulations, GSS has declared, “Metal-gate-first FD-SOI will be very good but metal-gate-last could be spectacular.” “…the technologist who that could develop and deliver metal-gate-last FD-SOI at 28nm will be able to offer you supply voltage below 0.5V,” they explained. They also noted, “The statistical variability introduced by the random discrete dopants in the FD-SOI MOSFETs is significantly lower compared to bulk MOSFETs with equivalent dimensions.”
The 38th annual SOI Conference is coming up in just a few weeks. Sponsored by IEEE Electron Devices Society, this is the only dedicated SOI conference covering the full technology chain from materials to devices, circuits and system applications.
Chaired this year by Gosia Jurczak (manager of the Memories Program at imec), this excellent conference is well worth attending. It’s where the giants of the SOI-related research community meet the leading edge of industry. But there are also excellent courses for those new to the technology. And it’s all in an atmosphere that’s at once high-powered yet intimate and collegial, out of the media spotlight.
This year it will be held 1-4 October at the Meritage Resort and Spa, a Napa Valley luxury hotel and resort, set against rolling hills with its own private vineyards. Finding the right spot for this conference is key. One of the things that people really like about it is that in addition to the excellent speakers and presentations, the locations are conducive to informal discussions and networking across multiple fields. This year’s spot looks like the perfect setting, with easy access to Silicon Valley.
The Conference includes a three-day Technical Program, a Short Course, a Fundamentals Class, and an evening Panel Discussion. Here’s a look at what’s on tap for this year.
(To register at the discounted rate, be sure to send in your registration by September 17th. You can get the pdf of the full program & registration information from the website.)
ARM’s SOI guru Jean-Luc Pelloie chaired this year’s Technical Program committee, which selected 33 papers for the technical sessions. There will also be 18 invited talks given by world renowned experts in process, SOI device and circuits design and architectures and SOI-specific applications like MEMS, high temperature and rad-hard.
Here’s a rundown of the sessions:
Short course: Design Enablement for Planar FD & FinFET/Multi-gates (chaired by UCL & Leti) The conference kicks off on Monday with six sessions by experts in technological trends, the physics of fully depleted devices, technology design kits as well as digital, analog and RF designs specific for FD-SOI.
The fundamentals course: FinFET physics (chaired by Intel): on Wednesday afternoon, three hour-long sessions will give comprehensive insights into the physics and processes related to multi-gate FETs.
Panel: Is FinFET the only option at 14nm? (chaired by Soitec) Following the always-popular Wednesday evening cookout, the panel discussion is a lively favorite event. This year’s invited distinguished experts will share their views on the industry’s FinFET roadmap.
All in all, it’s a great event. If you go, why not share your impressions on Twitter with #SOIconf12, @followASN and @IEEEorg? And of course ASN will follow-up with summaries of the top papers in our PaperLinks section. See you there?
Are FinFETs better on SOI? In a series of papers, high-profile blogs and subsequent media coverage, Gold Standard Simulations (aka GSS) has indicated that, yes, FinFETs should indeed be better on SOI.
To those of us not deeply involved in the research world, much of this may seem to come out of nowhere. But there’s a lot of history here, and in this blog we’ll take a look at what it’s all about, and connect a few dots.
GSS is a recent spin-off of Scotland’s University of Glasgow – but there’s nothing new to the research community about these folks. The core GSS-U.Glasgow team has been presenting important papers on device modeling at IEDM (which is one of the most prestigious of our industry’s conferences) and elsewhere for many years.
At the risk of stating the obvious, accurate simulations are incredibly important. Technologists need to be able to predict what results they can expect from different possible transistor design options before selecting the most promising ones. Then they also need to provide reliable models to designers who will use them before committing chips to silicon. One of the biggest challenges is predicting variability, which as we all know is getting worse as transistors scale to ever-smaller dimensions.
At IEDM ’11 last December, GSS-U.Glasgow presented Statistical variability and reliability in nanoscale FinFETs. This covered “A comprehensive full-scale 3D simulation study of statistical variability and reliability in emerging, scaled FinFETs on SOI substrate with gate-lengths of 20nm, 14nm and 10nm and low channel doping…”. Essentially they concluded that scaling FinFETs on SOI should be no problem – and in fact the statistical variability of a 10nm FinFET on SOI would be about the same as the industry’s currently seeing in 45nm bulk CMOS.
That paper was based on work that the GSS-U.Glasgow team had done on two major European projects: the EU ENIAC MODERN project, and the EU FP7 TRAMS project. It’s perhaps worth looking a little more closely at what those projects are about – and who’s involved:
A few months later, when Chipworks published pictures of the (bulk silicon) Intel 22nm FinFETs, the folks at GSS started a series of blogs that caught the attention of major tech pubs such as EE Times, Electronics Weekly and EDN. For reference, here are the blogs and basically what they concluded:
Specifically, the July 27th blog indicated that if FinFETs are rectangular in shape, drive current would be 12-15% better. Would that be easier to do on an SOI wafer? Soitec has argued that their “fin-first” SOI-based approach to FinFET manufacturing will save both time & money while getting better results (see Soitec’s Wafer Roadmap for Fully Depleted Planar and 3D/FinFET in Semiconductor Manufacturing & Design).
The GSS blog also reminded readers that the company’s CEO and founder, Asen Asenov (an extremely heavy hitter who’s published over 550 papers), has hinted that “…SOI FinFETs with an almost ideal rectangular shape may be a better solution for future FinFET scaling”. GSS has noted previously that “FinFETs built on an SOI substrate could have significant advantages terms of simpler processing, better process control and reduced statistical variability”.
Fin shape aside, GSS said that by virtue of the layer of insulation, SOI would give another 5% boost to FinFET drive current. But perhaps more importantly, that layer of insulation in SOI-based FinFETs would deliver on average 2.5 times less leakage – which would translate into a doubling of battery-life for your cell phone.
IBM has now entered into an agreement with GSS et al on a project called StatDES, for Statistical Design and Verification of Analogue Systems – see last month’s IBM blog by IBM Research Scientist Dr. Sani Nassif, entitled “Fins on transistors change processor power and performance”.
Dr. Nassif writes, “IBM, University of Glasgow and the Scottish Funding Council are collaborating on a project to simulate 3D microprocessor transistors at a mere 14 nanometer scale (the virus that causes the common cold is more than twice as large at 32 nanometers). Using a silicon-on-insulator (SOI) substrate, the FinFET (fin field-effect transistor) project, called StatDES, promises to keep improving microprocessor performance and energy conservation.”
The steering group also includes folks from ST, Freescale, Wolfson and Cadence, so one would guess we’ll be hearing more from this project – and others like it, to be sure – in the future, wouldn’t you think?