You Want Real Quantum Computing? Follow the Rainbow

Meet Roy G. Biv

Red, orange, yellow, green, blue, indigo and violet. The seven basic colors of the rainbow. Throw in black, grey and white. Now you have a Base 10 hardware architecture using optical transistors and interconnects.

A Space Odyssey

Binary / Octal / Hexadecimal are all fine and dandy. They exist because current hardware architectures are simple binary zeroes and ones.

Take a Byte

A binary byte is composed of eight bits. 2⁸ = 256
Let's say a decimal byte is also composed of eight bits. 10⁸ = 100,000,000

That is a 390,625% increase in the amount of information that can be stored within a single byte.

Theory of General Relativity

In relative terms, what kind of data savings can be achieved by using a Base 10 hardware architecture?

A widely used hardware data type is a 32-bit integer.

• 2³² = 4,294,967,296
• 10¹⁰ = 10,000,000,000

There are two data saving components here:

Exponent

1. Binary requires 32 bits to represent numbers between zero and 4,294,967,296
2. Decimal covers this numerical range at just 10 bits

Savings = 320%

Numerical Range

The decimal numerical range is larger than the binary numerical range

Savings = 233%

Total data space savings = 553%

Applicability

An optical hardware architecture can be applied anywhere electrons are currently used to represent data state including, but not limited to:

• CPU / GPU
• RAM
• Non-mechanical storage
• Data Bus (PCI-Express / RAM Bus / NVM Express / etc)
• Fiber Optical Networking

How Fast is Fast?

There is fast, then there is really fast.

The speed of light is 299,792,458 m/s.

Technically both electrons and photons travel at this speed. However, electrons traveling across a conductive surface generates heat. Which is the same principle as friction and slows an electron down to some percentage of its maximum attainable speed.

Electrons are also influenced by electromagnetic fields which can introduce errors when sending data from one place to another (e.g. CPU to RAM).

Photo(n) Selfie

I am sure there are challenges physics incurs on photons just as it does for electrons. However I am not specifically aware of what they are.

For the sake of this argument, I will say that using photons within the context of a traditional computer architecture (CPU / GPU / RAM / Storage / Networking / etc) does not incur:

1. A friction penalty: Pushing photons back and forth in the visible wavelength does not generate heat
2. A data corruption penalty: Photons are not influenced by electromagnetic forces like electrons

Beep, Beep!

Let's say we will use a 40nm optical architecture for representing transistors and interconnects within a CPU. The theoretical maximum frequency a CPU can operate with half-duplex interconnects would be:

• 299,792,458 * (1,000,000,000 / 40) = 7.49481145×10¹⁵ = 7.5 PHz

That would be 7.5 Petahertz. Or 1,742,976% faster than an Intel Core i7-7740X Kaby Lake-X Quad-Core 4.3 GHz processor.

Pipe Dream

Maybe so. But this idea for quantum computing is infinitely better than the current one being pushed where they want to use bosons within the nuclei of atoms to represent data state.