|November 2028 is the 40th anniversary of ESD. Click here read other 2028 lookbacks.|
Remember those days when electronic engineers, called hardware designers, designed those chips called ASICs and ASSPs? They had to pay millions of dollars to buy EDA tools and then spend 18 months and even years designing those chips for a few or even just one product in an application space.
To make matters worse, if the specifications of the design changed or if a flaw was found late in the design cycle, the design team had spend millions of dollars to respin masks or even completely redesign the chip. That was hell…that is until they found out they could really replace the chip along with a few others on the printed-circuit board with a single FPGA.
Thank goodness FPGA technology, which was originally pioneered and today still overwhelmingly dominated by Xilinx, advanced at the rapid pace it did. Today, a consumer can program new functionality into their iBrain that years ago would have required the expertise of a software engineer or an embedded systems programmer.
And 25 years ago, back in 2003, it would also have taken several hardware engineers–fresh from realizing there were no longer any benefits (be it gate counts, performance, power, die size, or electromagnetic compatibility) to designing an ASIC or ASSP over an FPGA–to program an FPGA.
Today, if you want to upgrade your iBrain sensing system, your hydrogen-electric hybrid automobile, or your home's control system, you simply contact the vendor on your iBrain, Insto-pay them, and have them download the upgrade to your iBrain to make it more intuitive, or to your auto to make it more fuel efficient as it flows in autonomous unity and safety with the rest of traffic at 120 MPH, or to your home to ensure it waters the plants, turns on the heater, and prepares to play your favorite song as the garage door opens automatically, of course, as you pull into the driveway.
Indeed, today we take a lot for granted. FPGAs touch every electronic and electronic-bio system today. FPGA platforms allow inventors to quickly and easily implement their ideas in silicon or biosilicon by describing in English, not code (as in yesteryear), what they want their invention to do.
The Xilinx IDS 28.0 software simply interprets that description, and under the hood, assembles the building blocks–what used to be called IP–stitching the blocks together and automatically writing the firmware, software, and operating system to implement your invention. The IDS 28.0 software even determines on its own which functions to run in hardware and which ones to run in software.
Of course, you never see any of this automation taking place; you just know that when you want your iBrain to Insto-Translate the conversation you're having with your new friend in China, all you have to do is contact your iBrain carrier and say “add this feature to my iBrain.” And we take for granted that this allows vendors to offer you tons of new features and functionality to a huge range of products, enabling you to customize their products (and of course charge for those customizations) to your individual needs.
But, as we write for the anniversary issue of Embedded Sensing Design (formerly Embedded.com, formerly Embedded Systems Design , formerly Embedded Systems Programming ), we should also recognize that the path that led to this degree of automation wasn't an easy one.
Although Xilinx is one of the oldest semiconductor companies (45 years) and the most successful, it has had its share of bumps and bruises. Remember when we thought that embedded engineers would be willing to learn these things called hardware description languages? That didn't turn out to be the case. Remember when we only offered chips?
Thank goodness Xilinx started offering the tools and blocks, as well as the silicon in what we call FPGA platforms. Doing that cut down TTI (time to innovation) considerably–and spurred growth certainly much faster than the older model. Remember when we (Xilinx) bought our closest competitor because at one time, there was all this hype that their technology was as good as our own.
Turns out it wasn't even close. Oh well, why spend time looking back when you can look forward. And we here at Xilinx do look forward to the inventions you will continue to come up with using our FPGA platforms. Thank you, friends for sensing this old-timey article in 200 languages across the globe simultaneously. Peace and prosperity to you all.
Vin Ratford is vice president and general manager of the solutions development group (SDG) at Xilinx, with responsibility for design tools, methodologies, IP, and hardware platforms. Ratford joined Xilinx in 2006 when the company acquired Accelchip, where he served as president and chief executive officer.