POET Technologies Inc.

For those who read this forum but who are not up on the technology.

It is now received wisdom that PTK will produce a 40nM microprocessor by early 2015, possibly sooner. This term 40NM is important because it refers to the length of the gate control in a transistor. All processors contain million or even billions of transistors etched onto a base substrate made up of layers, by a variety of manufacturing processes that are highly complex but not relevant to this discussion.

Imagine that a transistor is a tiny piece of material that can conduct electricity under certain circumstances. These circumstances depend on a second electric current that switches the transistor off or on (it’s why binary code can be represented by transistors). This area of a transistor, where the off/on effect occurs, is known as the gate. If you read the electronic literature, the methods of gate control have altered over the years to produce better control by altering the gate’s size and shape. Currently INTEL are making a transistor with 3 fins to improve control, hence a FINFET (the FET in this acronym is a Field Effect Transistor) and is present in their latest processors, it sticks up, out of the surface, and is, hence, 3 dimensional (3D) and not flat as in a POET chip (it is flat and this is known as Planar).

So, the length of this gate, determines the quoted length. POET is going to produce a 40nM gate which is longer than current gates in Silicon at 14nM, with efforts to go shorter. By the way, a nano-metre (nM) is 1 x 10-9 metre, none the wiser, then try .0000000001 metres. Thus POET’s gate length is 40 x 10-9 metres (or, before the mathematicians kick off, 4 x 10-8 metres).

Also, you may be delighted to know that, over the last 12 months or so, claims have been made about the speed of a POET chip, compared with the ubiquitous Silicon, is much improved at equivalent gate lengths. Concerns have been raised that a 40nM POET processor only improves performance by 3 or 4 nodal points. A node is the gate length and as they improve or shorten the gate length then this length decreases. It is done in a step by step process often at intervals of 18 to 24 months. So the size goes 40, 25, 15, 10 7 nM and the number of transistors increase as the gate length shortens – these lengths I have given are are the nodal points. So POET is claiming that a 40nM PET processor will behave as a 14nM Intel or other manufactures’ processors. Its performance is theoretically equivalent to the best processor on the market currently – this will only be confirmed when produced and tested and matched against other processors, towards the end of this year.

There are some qualifications however: the POET processor can work at much higher clock speeds than a Silicon one. Silicon, at best, works at 4 billion cycles per second (4 GHz). PET can work at 70 GHz (it is claimed) and this can be made even faster (I’ve seen figures for the future as high as 500 GHz). So the PET chip can be made to have more transistors, over time, and at much higher clock speeds, also over time. This will certainly make huge increases in processing data. This is why PET or POET can take over from Silicon and preserve Moore’s so called Law. Also, remember that POET has internal advantages such as light transfer of information and the ability to integrate other novel circuits within a single microprocessor that cannot be done in any other processor and it makes its internal speeds efficeint, cool and very, very fast.

What is proposed at 40NM is a serious and realistic attempt to produce a microprocessor at least as fast as the best Silicon processors on the market. POET is quite deliberately designed to be reproducible in existing foundries, with very modest alterations, and at low cost, whereas it has taken INTEL 10’s of billions of dollars to produce their latest processor. Also, POET can have its gate length reduced by known and standard techniques and can increase its clock speed. Silicon will not be able to do this, or only at extreme cost and very slowly. It’s why the potential is huge.

Once again, I apologise to the some who know all this, but there are many on this forum who find the messages difficult to follow – I hope it helps somewhat.

David