High-tech

Due Diligence: Pulsic

By The Chilli analysts

Introduction

Pulsic is a post Chilli R3 EDA (electronic design automation) software vendor, founded late 1999. The company currently has fully functional products shipping to customers, and is generating measurable revenue from its customer base.

Pulsic's heritage actually dates back before 1999. During the 1980s, Racal-Redac was involved in the development of software technology for the design of PCBs (printed circuit boards). The company's researchers noticed that PCBs were getting simpler, as the complexity was moving onto the ICs (integrated circuits) as Moore's law was allowing more transistors, and hence functionality, to be integrated. The company initiated a research and development programme to look into techniques for place and route (the placement and interconnection of blocks of logic).

Racal-Redac's was acquired by Zuken of Japan in 1994 and Zuken continued the research, using the technology for its products targeted at PCB development. Zuken developed the technology for IC development, creating the Rexsir product for the Japanese market (packaging it with analogue IC design tools from Seiko Instruments). At this point, the development team wanted to aggressively apply the technology to IC design, whereas Zuken was focused on PCB development. Zuken licensed the technology to EDA vendor Chip & Chip, which then changed focus, allowing the Pulsic team to license the IC technology.

Vital statistics

Value proposition

The Chilli perspective

Vital statistics

The company was founded by Mark Williams (COO), Jeremy Birch (CTO), Fumiaki Sato (VP of business development) and Mark Waller (VP of R & D), all formerly from Zuken, apart from Birch, who joined from ST Microelectronics. Ken Roberts (former VP and general manager for Magma Europe) was appointed CEO in 2003.

Seed funding of $300k was provided by William (Wai-Yan) Ho, a founder of several EDA startups including Chip & Chip (acquired by Aspec Technology) and Archer Systems (acquired by Epic and ultimately Synopsys). The founders bootstrapped the company by investing sweat equity and relying on the cheapest form of funding, namely customer revenues from an OEM relationship with Seiko for the first three years.

Pulsic received €5.3m of R3 funding from Prime Technology Ventures and Index Ventures to expand and strengthen sales and marketing channels. Roberts expects Pulsic to breakeven during Q2 2004.

Headcount is currently 15, with 12 staff involved in development and customer support. Pulsic has two offices, in Bristol and Tokyo. The company is in the process of appointing reps in Korea, Israel and the USA. Roberts aims to have reps in key geographies, backed up by a Pulsic employee to ensure share of mind and demonstrate commitment.

Blue chip customers include Elpida (ex NEC and Hitachi commodity memory divisions), Fujitsu, Renesas (Mitsubishi and Hitachi semiconductor divisions) and Sony. At the time Pulsic was created out of Zuken, the company had 8-10 customers with no tapeouts at that point. By June 2002, the company estimated 100 tapeouts with 20 customers and 30 product licenses. The company currently has 28 customers, 51 licenses in production use and well over 110 tapeouts.

Value proposition

The design of complex ICs requires the use of EDA (electronic design automation) tools. Designs are done at a high-level of abstraction, to ease design productivity. The high-level design is eventually translated into a physical description of millions of transistors and their interconnection, which is then imprinted on masks, which are used for manufacturing semiconductors at silicon foundries.

The placement and interconnection (routing) of blocks of logic circuits affects their performance, in terms of speed, power, size (of overall IC) and susceptibility to interference, impacting cost and reliability. This is not as critical in digital IC design as it is in analogue and mixed-signal (analogue and digital circuitry on the same design) IC design, where the exact value of a signal is critical, in contrast to digital signals, which are either on or off.

There are different approaches to IC routing - grid-based, graph-based and shape-based. Their suitability is determined in part by the kind of circuitry (digital, analogue, mixed-signal) to be routed, as well as by the constraints, e.g. timing, added to the routing operation. Analogue and mixed-signal routing requires additional constraints in terms of signal integrity, crosstalk, etc.

Grid routers work by superimposing a mesh-like template over the area to be routed, where evenly spaced horizontal and vertical tracks form a grid. The major disadvantage of a grid-based router is that not everything can be resolved to a grid, and they are not good at handling signal-integrity issues. A shape based router works without a grid, and can therefore work with complex geometries and get a higher routing density. A shape-based router can handle complex requirements such as differential pair routing, shielding, bus interleaving, 45-degree routing, etc. This makes them ideal for analogue and mixed-signal design.

Shape-based routers have traditionally had problems handling large data sizes, making them suitable only for small designs. Pulsic has overcome this with an approach known as 'T-shape', which is a data representation for storing routing information. The company is in the process of filing patents for its technology.

The company has two products based on the same core technology. 'Lyric' was released in June 2002, although "early adopter" versions had been available since April 2001 through Seiko Instruments. In June 2003 Pulsic introduced 'Prelude', a new physical design suite including routing based on the same technology, using distributed computing. Lyric is positioned at those designs where blocks to be placed and routed are up to 100,000 gates, whereas Prelude partitions larger designs into blocks, which are then spawned off to Lyric running on a network of machines.

The performance of place and route tools is traditionally assessed in the size and speed of the resulting device. In the analogue and mixed-signal domain, design productivity is a key issue, as much of place and route is manual, and any automation that deals with signal integrity issues and angled wires can result in an order of magnitude improvement. According to Roberts, "Pulsic is targeting unmatched automation for place and route in the analogue space. For mixed-signal, we bring a unified interactive, semi-automatic and fully-automatic environment offering better usability than existing solutions." Pulsic's business model relies on one and three year time-based licenses as well as a perpetual license. Pricing for Lyric begins at $100k for a basic, single-node license.

The Chilli perspective

Technology shifts: as digital circuits shrink, more analogue effects come into play, where each wire becomes a potential internal antenna, and signal integrity becomes more of a challenge. With more high-frequency communication standards and designs emerging, including UWB (Uweeba), Wi-Max, etc, greater attention to mixed signal design integrity will be necessary.

COCS: the total cost of customers serviced and supported is an important metric for all EDA, IP and software vendors. Many such companies fall into a trap of supporting customers which are ultimately unprofitable, as the true cost of support exceeds the profit margin from the product deal. Pulsic is well positioned in this respect, both through it's history through Zuken and the OEM deal with Seiko, which gave it a number of steady blue chip customers, as well it's complete focus on a specialist niche. The company has developed a number of additional modules for specific customer applications, e.g. datapath, memory and flip-chip, that provide some vertical technologies alongside the broad analogue mixed-signal capability, and this will allow further penetration into it's blue chip customer base.

Partners: in EDA, partners and competitors are often the same companies. For an EDA product, it's vital to be able to integrate with the design flows in common use by customers. Pulsic's products interface to tools from Cadence, Synopsys and other vendors, and the company is a member of the Synopsys and Cadence partner programmes.

Competition: Pulsic has chosen to concentrate on the niche area of analogue and mixed-signal routing, which has a lower number of players than the digital realm. Cadence currently dominates this space with it's own shaped-based router, IC Craftsman (from it's acquisition of Cooper & Chyan). While Cadence is an 800-pound gorilla, it is better to have one large competitor than three, as is the case with other segments of the EDA industry.

Way forward & outlook: Pulsic has been something of a secret, so far. Being in stealth mode is great for gaining initial traction in a target niche, but the company must now raise its profile and scream and shout about its customers, revenues, and prepare for an exit. The company needs to demonstrate customers it can use as references - it's often tricky to convince conservative Japanese customers to provide testimonials, but this comes with time as the relationship develops deeper.

The company is a good example of the adage that it is wise to bury the NIH (not invented here) syndrome and get to market quickly with a unique proposition, whatever its provenance. This is a common story in North America, but in Europe, especially the UK, there has been a tendency to be preoccupied with original technology, rather than solving customer issues and commercial aspects.

Pulsic is well positioned with existing blue chip customers, to prepare for an exit within the next 12-18 months. This is most likely through acquisition by a larger EDA company looking to enter or strengthen their share of the analogue and mixed-signal place and route market. On current form, it could be a stretch for Pulsic to expand its headcount, sales and service centres from its existing base.

Comments on this story? Send an e-mail to editor@thechilli.com

© Chilli Publishing Ltd 2004

11FEB2004