A look back to the Cell processor and solid-state drives and forward to the driverless car.
As I do every January, I am looking back 5, 10, and 15 years to see what predictions did and did not turn out to be right, and how that relates to design technologies enabling those developments. Looking back five years reveals just how key system-development technologies were for what IEEE dubbed the “Top 11 Technologies of the Decade”. Looking back 10 years shows how they enabled communications and multicore processing. And it appears that the 15-year-old prediction of the car as a “Web browser with tires” has become reality, again enabled by system design and verification techniques.
The January 2011 issue of IEEE Spectrum ran a special on the “Top 11 Technologies of the Decade” from 2000 to 2010. Four of the top five—smartphones, social networking, voice over IP, and multicore CPUs—clearly were enabled by the advances in system design technologies during that timeframe. Without advanced verification, emulation, and virtual and FPGA-based prototyping, as well as advanced hardware/software enablement, development of these technologies would not have been possible. And the list is even longer when expanded to the top 10 and the “Next to the Best Technologies“.
The January 2006 issue of IEEE Spectrum talks about that year’s winners and losers. It was hit and miss. In “Multimedia Monster“, Samuel K. Moore described how the cell processor was 2006’s winner. The article is very insightful about the capabilities and shows how its compute power would be applicable to graphics—for example in high-end game consoles. According to “End of the Line for IBM’s Cell“, calling it quits in 2009 “confirms suspicions that Cell’s overall commercial success has been limited, and there will be a number of cheaper, higher performance, more widely supported alternatives to the processor starting in 2010.” However, I still love my Playstation 3, and from a design perspective it was certainly a trailblazer project. The design was fabricated using a 90nm SOI process. The 45nm Cell processor was introduced in concert with Sony’s PlayStation 3 Slim in August 2009. What is cool is that the design technologies used to develop the Cell processor mimic what has become mainstream today with the Cadence System Development Suite. It is probably safe to assume that the Cell processor was emulated in hardware, and as virtual platforms even several software emulators were available. In comparison, all of today’s major game consoles use Palladium emulation—part of the System Development Suite—for verification and early software bring-up.
And software development was key. And complicated. One of the predictions from that time frame definitely came true—programming complex multicore systems can be an inhibitor of their adoption. In “Holy Chip!”, Daniel Lyons accurately predicts:
If anything stops Cell’s commercial success, it is likely to be the chip’s very power. It is, to put it politely, a challenging platform for software creators. “The programming model is a nightmare,” says Marc Tremblay of Sun, chief architect of a rival chip called Niagara, which uses a more traditional approach. He argues Cell’s balky design will snag widespread adoption beyond gaming.
I couldn’t agree more (of course 20:20 hindsight always helps), and if you are interested in a review of the rise and fall of some multicore architectures, check out “Real World Multicore Embedded Systems” to which I contributed chapter 3 on that topic.
So was the Cell a technology winner? I would say so, especially with its trailblazer characteristics.
Another prediction from the 2006 issue was about solid-state discs (SSDs). In “Too Little, Too Soon“, the prediction states that “Samsung’s solid-state disks will be puny, pricey, and impractical,” marking them as a loser. Well, for 2006, perhaps. There is an interesting lesson about timing here. According to data from Statista, SSDs are expected to be 32% of the global shipments of HDDs and SSDs in PCs in 2016. It simply took time. And they used Palladium emulation, Cadence Accelerated VIP, and Cadence VIP as well, significantly increasing their productivity. Just another example of System Development Suite technologies enabling innovation.
The prediction from that time was correct in the short term… but sometimes technology is just early. Ten years on, I would not want to work without the SSDs in my devices, selfishly driven for me by performance and noise, actually.
How hard predictions are and how little “coolness” of technology has to do with commercial success becomes clear when reading “Black, White, and Readable“, in which NanoChromics displays were marked as a clear winner because the display “looks like ink on paper, and it has a high contrast ratio and a wide viewing angle. It’s visible everywhere, even in bright sunlight. It consumes little power and is cheap to manufacture”. Similar technology must have been used in Amazon’s Kindle and other readers. But the last reference I could find to the technology was in the 2010 article “Ntera Prints a Display on Almost Any Surface”. Oh well. It sounds like the basic technology at least was a winner.
In closing, when scanning the 15-year-old references in the “Always on” IEEE Spectrum from 2001, the article “The Car: Just a Web Browser with Tires” caught my attention. Sounds like that prediction has been implemented by now and the foresight from that time was pretty impressive—I am writing this as I have Waze already set to get me to my next destination and am told that I need to leave soon.
Of course, soon enough I will be writing blogs while my car is driving me. And what’s cool is that the system design and verification technologies that my team and I are involved in creating are key enablers!