ChipCrunch.com

Some things just don’t go away easily, and in some rare cases this might actually be a good thing. We wrote about Ambric and their promising Am2045B processor and development environment back at the end of 2007. Subsequently, at the end of 2008 after failing to raise an additional $15 million in funding, Ambric folded. Now it seems that there will be an additional chapter in Ambric’s history for the company’s assets were recently acquired by Nethra, as I was recently informed by Nethra’s Senior Director of Business Development, Manu Pallai, via a note titled Ambric Continued. According to a press release on the company’s website, Nethra intends to complete the integration of the Ambric team by mid-April and continue working with current Ambric customers to ensure a seamless transition and continued support.

So who is Nethra and what use could they have for the Ambric team and assets? Let’s start with the basics: Nethra is a privately held fabless semiconductor company based out of Santa Clara, CA. The company specializes in imaging and video solutions for the digital consumer market, as well as surveillance and industrial applications. It is backed by Artiman Ventures, GunnAllen Financial, iLabs Capital, and Quicksilver Ventures. The amount of funding the company has received is presently not being published, but digging around the net it seems that the company received $14.2 million at the end of 2006. IC development costs are rather high and with this acquisition I would think that the company must have obtained some additional funding in the mean time. Nehtra’s current flagship processor is the Nethra Imaging (NI) 2065 which integrates a 32-bit RISC CPU core and supports CMOS sensors up to 3.2 megapixels. The key advantages the NI2065 offers are algorithms the company has been able to implement for adaptive lighting, low-lighting, and image stabilization to improve image quality. The NI2065 consumes a maximum of 0.2 Watts, measures 5 mm x 6 mm in a 68 pin package, and as such is targeted at mobile applications. Below is a chart summarizing the features for several of the NI 20XX processors.

Now consider the following: Nethra is heavily vested in improving the image quality captured by CMOS sensors, essentially striving to enable performance comparable to CCD sensors but at a much lower price point. Their current products max out at 3.2 megapixels and rely on algorithms to compensate for the image quality captured by the CMOS sensors. With higher image resolutions just around the corner it is likely that the 32-bit RISC core will become overwhelmed regardless of how optimized the code for the algorithms might be. The good news is that image processing is a highly parallel task where the same filter can be applied to different areas of the picture simultaneously. Now, if these algorithms can be combined with the massively parallel architecture that Ambric was pursuing, and the power can be kept reasonable, then Nethra might indeed be plotting a course that should enable it to scale well as the picture and video demands become more complex. It seems like a showdown between Nethra and Movidia might be looming in the near future.

Let’s face it, as much as we would like our chips to work at the desired frequency and voltage levels, when the in initial silicon comes back, we are happy if we can simply get past a basic boot. Things are not likely to improve with the ever smaller process sizes which are no friend of variation, nor are they going to get better with developers integrating an increasing number of cores and IP from many sources into a single System-on-Chip (SoC). With a few exceptions, most engineers dread spending weeks in the debug labs trying to figure out which paths are causing the issues. So what would make sense going forward? How about a vastly improved application-level debugging and profiling platform specifically targeted at SoC devices that is generic enough to work with different architectures while still enabling the capture of detailed system operation? Enter UltraSoC Technologies, a startup based out of Cambridge, UK that was spun out from the University of Kent in 2005.  The company is developing what it calls UltraDebug, a tool that is supposed to provide all the features we mentioned above and is specifically targeted at embedded systems and multi-processor SoCs. Given the current career opportunities, it seems that the product will be based on Eclipse C/C++ Development Tooling (CDT). Unfortunately, this is just about all UltraSoC is revealing for now. In addition to UltraDebug they are also working on a tool called UltraOptimize that is supposed to aid the engineer with profiling and performance measurements. UltraSoC recently raised about $600K in capital from the South East Seed Fund and Iceni Seedcorn Fund, which is to be utilized for the continued development and marketing of UltraDebug.

Last week NVIDIA launched the GPU Ventures Program, which strives to identify, support, and invest in early stage startups that are able to leverage NVIDIA's GPUs for parallel computing applications.  Clearly, the goal that NVIDIA is trying to accomplish here is to show that the processing power of GPUs can be utilized for many more tasks than just rendering polygons in video games. Finding some killer applications for all this parallel processing power would be a significant step in growing the market demand for GPUs. But finding killer applications is not exactly an easy task. Many technologies finally took off long after the companies that pioneered them disappeared. To avoid a similar fate, the clever guys at NVIDIA figured out that instead of imposing their idea of what a killer application ought to be, it might be smarter to stick to what they know how to do best, building massively parallel chips, and to leave the pursuit of finding killer applications for these chips to an ecosystems of startups centered around their technology.

Essentially, this venture program is more or less an extension to the Emerging Companies Summit - a private showcase of startup companies that NVIDIA held in 2008 for the first time. This summit is scheduled to take place again sometime this fall, although an exact date has not been set yet. NVIDIA plans to showcase the developments by these startups on their newly launched Venture Zone website. Currently, in addition to being able to browse the profiles of companies that have already become part of the ecosystem, the site also allows you to submit your business proposal on how you plan to build your company around NVIDIA's Compute Unified Device Architecture (CUDA) computing platform. Investments in early startups are expected to span from $500K to $5M, and NVIDIA also intends to help out with marketing, development, and even product distribution. So if you have a good idea that can take advantage of the CUDA platform it might be time to dust of that business plan textbook and get to work!

These days everything seems to be centered on portable and wireless technologies. There are a great number of startups specializing in RF technologies, digital amplifiers, and software defined radio. Even wireless power harnessing for low power applications has appeared on the horizon. With 3G, 4G, WiMAX, Ultra Wide Band, Wireless USB, and beyond it seems that the entire wire-line infrastructure that has been built over the last century at a cost of billions, if not trillions, of dollars, might become obsolete. Given the media’s infatuation these days with everything wireless, someone not well versed in the data communication business can easily be mislead to believe that wire-line obsoleteness is inevitable. Thus, when a startup appears and claims that it has developed technology that will revolutionize the speed and distances at which wire-line networks can operate, one quickly takes notice. Enter LightWaves, an equipment startup that originated out of Cedar Rapids, Iowa and has recently moved its headquarters to Austin, TX. The company has developed what it calls TimeFlux technology, which enables data transmission on any wire type at greater speed than any other currently existing technology. I recently had the pleasure of meeting with Drew Childress, LightWaves’ Vice President of Business Development and Marketing, to speak a little bit about LightWaves. During our meeting Drew was kind enough shed some light on the company’s TimeFlux technology and data communication businesses in general. What follows below is a summary of our conversation.

ChipCrunch:
"Cedar Rapids, Iowa is a rather unusual place for a startup to emerge from?"
Drew Childress:
"That’s true. But what you might not know is that Cedar Rapids actually has the highest number of engineers per capita of any major metropolitan area in the US with Rockwell Collins  being the major employer."

ChipCrunch:
"I didn’t know that. What prompted the company to move its headquarters to Austin, TX recently?"
Drew Childress:
"Although we have plenty of EEs in Cedar Rapids, what we lacked were people that really understood telecom, and people that wanted to be engaged in startups. Additionally, when talking to Venture Capitalists, they very much considered Cedar Rapids a fly over zone, meaning, they urged us to move to an area where both talent and VCs were more readily available."

ChipCrunch:
"Obviously you had other choices than Austin, such as Boston or San Jose. So why did you choose Austin?"
Drew Childress:
"Boston seems rather more focused on the medical startups. San Jose and Austin both feature a lot of electrical engineering talent, however being from the mid-west we felt a lot more at home in Austin. As a matter of fact, if you could pick up Austin and place it somewhere in the mid-west it would fit right in, it is very much a mid-west kind of town. Additionally, other factors such as the cost of living and receptiveness of VCs also influenced our decision. "

ChipCrunch:
"What are the plans for your site in Cedar Rapids?"
Drew Childress:
"We’re definitely keeping the Cedar Rapids site. We currently employ several outstanding engineers there who did not want to relocate down to Austin. Further, we lease a lot of land up there. Land we most likely could not afford down here. We have over five miles of buried wire up there and as such, unlike other companies which rely on simulations to test their theories, we are able to test everything on actual wires. "

ChipCrunch:
"Your company web-site is rather vague on how many employees you currently have, how much funding you have received and from whom, and so on. Can you elaborate?"
Drew Childress:
"Sure. We currently employ sixteen people almost all of whom are electrical engineers. We just recently hired three engineers in Austin, and plan to hire an additional eight or so throughout the year. We have had several rounds of funding in which we were able to raise a total of 16 million. We’ve been able to use that money to run the company for about five years, so we are very conservative cash wise. We are currently working on Round D for our funding."

ChipCrunch:
"With the economy being as it is, do you have any concerns about raising more money?"
Drew Childress:
"Not really. If you have a promising technology that can solve some real problems, there is still plenty of money out there. On a related note, an interesting thing about the economy is that the demand for home services such as internet and cable TV tends increase when times are bad. People spend more time at home and look for cheaper entertainment alternatives to going out."

ChipCrunch:
"Now to the real question. Everybody seems to be enamored with wireless these days. So why do you believe that wired is the way of the future?"
Drew Childress:
"If you understand data communications and the industry, you will quickly come to understand that in the industry nobody believes in wireless for large data transfers. In order to deliver IPTV or related video services you need high quality, low jitter and low latency. In other words, you need guaranteed quality of service. A customer might tolerate mediocre video quality on a small portable device screen, but when the same customer watches content on a big screen television the quality being delivered better be top notch or the customer will be very unhappy. You can look at air as one large coaxial cable where all the wireless devices share bandwidth, and while this will work, the throughput will suffer and so will the content quality that is being delivered."

ChipCrunch:
"Why not buffer the content then? For example, the same way that providers such as  Netflix do before streaming a movie?"
Drew Childress:
"Buffering is certainly a solution, if you only want to watch a specific stream. What happens when you want to browse channels? You can’t expect the consumer to wait several seconds while switching channels. Nor can you expect them to have enough storage space and bandwidth to buffer all of the channels simultaneously. And to complicate things even further, bandwidth requirements will only increase from here on. 3D television is coming down the pipes and that by itself will double the required data rates."

ChipCrunch:
"Ok. So now that we have an idea about the problems, what do you guys bring to the table to solve them?"
Drew Childress:
"We have developed a new wire-line technology for home networking that we call TimeFlux. TimeFlux works on any type of wire and at speeds significantly faster than anything that is currently available. The type of wiring and quality of wiring is unique in every house, and as such for us to be able to use any kind of wire and our ability to maintain high throughput even on lower quality wires is a huge advantage. As a matter of fact, just last month we successfully completed a mock home test with a potential customer, and we achieved speeds twice as fast as anything they have ever seen."

ChipCrunch:
"I understand you don’t want to reveal all the details about TimeFlux, but can you expand a little bit on how you are able to achieve these high speeds?"
Drew Childress:
"Sure. But let me first start out by taking a step back and explaining a couple more problems that telecom providers currently face. Many technologies, such as vdsl2 for example, promised 100 Mbps at a range of up to 5000 feet. As these technologies exited the labs, reality quickly kicked in and currently vdsl2 can at best do about 25 Mbps at up to 3000 feet. This shortcoming in performance stems from the fact that current technologies are hitting fundamental physical limits of carrier based modulation schemes, which cannot be overcome by increasing the signal power or by more sophisticated encoding techniques. In the case of vdsl2, as signal power is increased, waves begin jumping off the actual wire resulting in a lack of throughput. What current schemes essentially do is modulate the frequency and multiplex in time. With TimeFlux we do exactly the opposite: we modulate time and multiplex in frequency. Based on the math alone, we are able to achieve twice the throughput of competing technologies simply because we are twice as efficient with bandwidth."

ChipCrunch:
"I can imagine that that there are quite a few companies interested in your TimeFlux technology. How do you intend to bring it to market?"
Drew Childress:
"We are currently designing TimeFlux based products for the commercial market (hotels, hospitals and multi-tenant buildings). Our UltraJack family of wall-mounted network adapters can utilize the existing wiring in these buildings to deliver high quality connectivity throughout the building. Next up will be products for the residential market and we may also offer TimeFlux chipsets a little further down the road."

ChipCrunch:
"Thank you so much for your time and all the information that you shared. Best of luck with TimeFlux and please keep us abreast of any new developments at LightWaves."