Shared tool developed for quantum and supercomputer systems

Posted by donutloop@reddit | programming | View on Reddit | 9 comments

[-]

Big_Combination9890@reddit

Quantum computers are a key emerging technology, particularly for specific kinds of problems that require enormous computing power.

Key for what? And what problems are we talking about specifically, where QCs have proven, in real world scenarios that they will make a contribution on par with the hype surrounding this "key emerging technology"?

https://en.wikipedia.org/wiki/Quantum_algorithm

Here is a list of known quantum algorithms. You might notice: It's not very long. So, how is this technology, which, after many decades of research is still not practically viable, and for which only a very limited number of algorithms are actually known, "key" for something?

Please elaborate.

https://www.scientificamerican.com/article/will-quantum-computing-ever-live-up-to-its-hype/

[-]

Determinant@reddit

The larger quasi-quantum computers from companies like D-Wave are great for optimization problems since they're good at simulated annealing but they're not truly fully quantum in the traditional sense so they don't bring the elusive exponential speedups that most quantum hype refers to.

The true quantum computers (which unfortunately operate at tiny scales) aren't useful for most applications today but are great at generating truly random noise. High-quality random number generators have many current benefits.

[-]

Big_Combination9890@reddit

but are great at generating truly random noise

So is a large shelf full of lava-lamps, which costs a tiny fraction of a QC setup, and actually works.

[-]

Determinant@reddit

CloudFlare uses lava lamps as a small part that feeds into and is mixed with their larger entropy pipeline.

Lava lamps are fundamentally deterministic from a physics perspective if you run it through a simulator that has the exact initial conditions. The problem is that we can't measure the conditions precisely enough so it seems high enough quality from a practical standpoint.

Quantum RNGs are fundamentally indeterministic due to the fundamental physics regardless of how precise you measure initial conditions.

[-]

Big_Combination9890@reddit

The problem is that we can't measure the conditions precisely enough

Wrong. We cannot measure them at the necessary precision at all.

Why? Because of quantum mechanics.

Ironic, isn't it? 😎

Quantum RNGs are fundamentally indeterministic

Yes, so is radioactive decay, thermic variations, certain electrical phenomena and a whole bunch of other things that doesn't require millions of dollars and a specialized lab to set up.

Sorry, but your argument simply doesn't work. Just because I can use an entire freight train to transport a single Banana, doesn't mean it makes sense to do so, or is practical in any given situation.

[-]

Ameisen@reddit

but are great at generating truly random noise

So is a large shelf full of lava-lamps.

So is a couple of thermistors... which can be part of the COU package. Random temperature fluctuations are a valid source of entropy.

[-]

Wonderful-Wind-5736@reddit

Even D-Waves claims are questionable. QUBOs are comparatively difficult to create compared SAT or constraint linear problems. In some problem classes like TSP, the encoding kills the theoretical quadratic speedup completely.

You need to remember compute is completely fungible. There's no material difference wether a problem was solved on a quantum computer or a classical one. Whatever is cheaper per unit problem solved wins. Development cost factor into that, too.

I also don't see how measuring some random quantum state is different from generating random numbers by hashing the image of a camera pointed at lava lamps or just measuring thermal noise.

[-]

Determinant@reddit

D-Wave has had over $7 million annual revenue for the last 3 consecutive years so they must be providing real value.

Regarding the quality of the randomness, Google and IBM seem to think it's much better than other sources.