I read with great interest this forum’s debates when they turn to the science of PET/POET technology.If you read the output of very successful investors, then you quickly learn that part of their recommendations involve in depth understanding of the products of the company you are investing in.This is generally straight forward when you deal with single or even multiple products we all use daily, Coca Cola, motor cars, mining, etc.It is much more difficult when it comes to products which involve advanced technology.
Add to this the claims of a revolutionary product, in a historically highly successful field such as microprocessors, and it is not difficult to grasp that we are all reliant on what dedicated experts tell us.Even if you have an advanced scientific background, it can be extraordinarily difficult to understand where this technology stands in relationship to others.Hereby hangs the difficulty we all have.
If you hunt through the literature to compare Group 111-V elements used as semi-conductors compared with Silicon, you come against a wall of information which tells you all sorts of pros and cons which make comparison impossible in theoretical terms.
My take on all this centres around what we know from the output of the chief scientist of this company.If you can understand that using a POET/PET microprocessor alters the way and speed with which it handles classical information inside the processor.All microprocessors shunt around groups of digits (BYTES) from one place in a microprocessor to another for storage or arithmetic procedures and a programme tells the processor where to put all this information and how to use it.It does this one step at a time according to a clock which times each step.A Silicon transistor can do this 3 – 4 billion times a second – we have impressive results in any home computer to confirm this.
The problem that emerges is the increasing demand for faster computing, and everyone knows that mobile technology is a worldwide phenomenon.So, if you ask your microprocessor to do lots of things at once so, for example, if you want to process repeatedly a series of large frame images, attach sound at the right point, send this to a screen with the highest definition and the sound to amplifiers and speakers which miss nothing then you require a fast microprocessor and well designed programming.Here’s the trouble: Silicon has to address this issue in a number of ways.It puts as fast a clock speed as possible (too fast and the chip overheats and fries even if is cooled).It puts more and more transistors in the microprocessor by miniaturising circuits.This addressed some issues.Additionally you can add extra microprocessors so you hear terms such as dual or quad core computing.
Now, we hear of the problems from all this effort emerging.You cannot turn the clock speed up because this overheats the Silicon and so clock speeds have stagnated at less than 4 billion cycle per second.Miniaturising is reaching its limits because of what are called short channel effects – as you make the transistor control smaller then quantum effects on the electrons mess up the on/off mechanism of a transistor, causing signal error.Hugely expensive research programmes involving billions of dollars have been used to help with these effects which exponentially increase as you make transistors smaller.The effect has been that to cover the cost of this research now causes chip prices to rise and reduce ability to produce reliable microprocessors (prices used to fall at one time as processing improved).The effect has been colossally costly for Silicon to proceed and increasingly the “end of Moore’s Law” is heard.Even increasing the number of microprocessors (known as parallel processing) doesn’t solve the problem because the software programming involved in controlling all the processors cannot resolve issues where single bottlenecks occur and programming is hugely expensive too.
Then, uniquely as far as we are aware, a process (POET technology) has been developed on a different substrate (Gallium Arsenide) which allows much higher clock speeds – we are talking about a 20 fold increase in clock speed.The heat generated is much lower because of the substrate’s chemistry allows less resisted electron flow and so it is possible to miniaturise much more easily and very much lower heat loss demands (it saves on a battery by a factor of 5).Additionally, because you can create mini lasers and organise internal light information transfer, you can move BYTES at the speed of light inside and outside the microprocessor, so you can transfer at the highest possible speed in the physical world.
My last paragraph states what PET/POET is able to achieve now.It implies not just an increase by a microprocessor of computing power but a huge step upward.Intel have just spent billions producing its Broadwell chip, shortly to be available.This has a 25% increase in computing power and a 5% saving on power use.It had to redesign the entire chip architecture and design different transistors to do this.Reviews were notable for their underwhelming commentary in the face of this effort.We are told that 100 nM POET chip is as fast.When the POET chip is manufactured at similar transistor densities and clock speeds tweaked.The processing speed will be many times taster than this.That is achieved at much lower cost because fabrication costs are designed to be very substantially lower – the processes are subject to multiple patents and this is what we are told.
To summarise: Yes! This is revolutionary, so buy in; buy in now.Not many understand the technology but, when they do, it will certainly interest institutional investors.The tipping point is nigh, BUY!
David
