Controlling the World
I sometimes call this invention a "control architecture." This is not to confuse it with the world of microcontrollers, although it will perform very well in such a capacity. However, the design's low-level orientation primarily aims it at control of high speed logic that is closely associated physically. I have compared it to ASIC technology, but I believe the architecture and its simulator provide advantages over such Application Specific Integrated Circuits.
As an architecture, its "microassembly" language has been designed from the ground up to provide that low level control. It is not just a third-party RISC or bit-slice core offered as a silicon library item. In addition, it provides a well thought-out structure, a paradigm if you will, to tie the circuits it controls into a larger parallel processing environment, if a given application requires that much computing power.
Much of the engineering effort to date has been applied towards the creation of a software simulation of the FSA. The strategic reason for investing this time is that no programmable chip gets sold without being accompanied by an easy way to develop the programs which are to run on it, and this simulator is the precursor of such a development system, one in which any custom logic will be able to be designed in the well known C language prior to its conversion into hardware.
Because the machine language is powerful yet simple to master, and since the logic specification language is the ubiquitous C, I believe this architecture can support very rapid product design cycles. In addition, the high degree of internal parallelism should allow the final product, especially if it is complex, to run extremely fast in its integrated circuit form.
Product development, ASIC or otherwise, can be aimed in two different directions. There has always been a good aftermarket in the computer industry; engineers and purchasing agents usually prefer second- or multiple- sourced parts. This architecture can certainly be put to work duplicating the operation of other devices. Indeed, it could tackle the most complex, since one of its strengths will be the emulation of other processors' instruction sets.
Higher profit, on the other hand, comes from creating new, successful, cutting-edge products. As far as microcontrollers go, if the Flexible System Architecture can command ultra fast, highly complex circuits, it can directly control, through its own instruction set and powerful ALU system, virtually anything else: an industrial process, an engine, an appliance, a heating system, a robot arm, etc. Other applications and products may require the greater horsepower of a parallel approach, or the inclusion of additional specialized circuitry.
To oversimplify, the basic FSA could be described as a "standard" section, as opposed to any circuits it controls, which exist in a "development," or "custom" section. These terms are somewhat misleading, as they imply specific physical locations for each. The standard section is actually a core which can be replicated and dropped in anywhere, hence the internal parallelism of the architecture. Nevertheless, for discussion purposes here, the term "section" will serve.
Returning to marketing, another entry for the FSA would be that of HC add-in boards and components (I use here the abbreviation HC, for "home computer," rather than PC, both to avoid brand confusion, and because "home" is a broader concept than "personal"). For a simpler application such as a disk drive controller, a device implementing only the standard section of the architecture should have ample power to handle the task through software alone.
A more complex product such as a sound card might need a chip with more computing power designed into the development section, but this would consist of standard DSP (Digital Signal Processing) circuitry, and once created, could be further adapted into other off-the-shelf versions of the same basic part.
These are the sorts of deployments which would be the logical first choices for a flexible architecture to target in the HC arena. No matter how dissimilar the jobs these various add-in cards accomplish, just a few versions of the FSA should be able to handle most of them. A small set of similar chips, selling at commodity IC prices, should find its way onto many different kinds of boards performing many different kinds of functions, because the software on each of them can be tailored to the specific task that card does.
While such software will not be cheap to develop, the current chips on the current boards are complex enough that they also require a programming effort, or its equivalent in hardware design, so costs run more or less the same. As time progresses, the new architecture will gradually augment its hardware price advantage in the software area as well, as more and more previously written code is reused in upgrades and new applications.
Once an engineering and manufacturing infrastructure is built through servicing these simpler board-level functions, add-ins with tougher computing requirements, such as exotic video or voice-recognition cards, can be developed on a case-by-case basis with specialized custom logic in the architecture's development section. Of course, even these custom circuits will still be driven by programs running out of the standard section and written in the standard FSA language, so the cost benefits of software reuse will continue to apply.
Thus, the Flexible System Architecture arguably will become one of the best tools for developing home computer System-On-Chip (SOC) logic in the future.