Why Are We Here? : George Ellis

I seemed to remember that you set N different calculations going, and only one of those can return a result, and then you collapse the wavefunction and extract the answer - but I don't remember for sure. I never had access to a QC to motivate me to make the effort!!!

However, a quantum computer must do something in parallel otherwise what is the point of it - even theoretically?

A quantum computer can evalaute 2^n input combinations to unitary functions in parallel where n is the number of qubits. Already for n = 100 that's 10^30 function evaluations in parallel which is completely crazy compared to traditional computing. That's why there's such a hype around it. At the same time you shouldn't really compare it to how traditional computation works as the results can't directly be observed. Whether the calculations really happen or they only happen for the measurement you read out is a philosophical question.

A quantum computer can evalaute 2^n input combinations to unitary functions in parallel where n is the number of qubits. Already for n = 100 that's 10^30 function evaluations in parallel which is completely crazy compared to traditional computing. That's why there's such a hype around it. At the same time you shouldn't really compare it to how traditional computation works as the results can't directly be observed. Whether the calculations really happen or they only happen for the measurement you read out is a philosophical question.

Well my hazy recollection of Shor's algorithm, which factors a very large integer, is that you perform a sequence of steps (in parallel) but then you get some useful information out of the wavefunction by finally collapsing it. However the information (presumably one of the factors of the integer) only has a statistical chance of being correct, so you have to repeat the whole process a number of times. Also, if the coherence of the 100-qubit system only lasts for about 1 microsecond the question is how many steps can be crammed into that microsecond?

I guess another way to put it, is what if any interesting computations has it performed, and are the certain encryption schemes (such as the RSA algorithm) now vulnerable using this machine?

David

Well my hazy recollection of Shor's algorithm, which factors a very large integer, is that you perform a sequence of steps (in parallel) but then you get some useful information out of the wavefunction by finally collapsing it. However the information (presumably one of the factors of the integer) only has a statistical chance of being correct, so you have to repeat the whole process a number of times. Also, if the coherence of the 100-qubit system only lasts for about 1 microsecond the question is how many steps can be crammed into that microsecond?

I guess another way to put it, is what if any interesting computations has it performed, and are the certain encryption schemes (such as the RSA algorithm) now vulnerable using this machine?

David

The threat from Shor’s algorithm is not imminent. But presumably hostile foreign countries like China is pursuing a “store now, decrypt later” strategy even with today’s internet traffic: https://www.nextgov.com/emerging-tech/20...er/187020/

The threat from Shor’s algorithm is not imminent. But presumably hostile foreign countries like China is pursuing a “store now, decrypt later” strategy even with today’s internet traffic: https://www.nextgov.com/emerging-tech/20...er/187020/

Thanks, OK - my next question is whether there are many other useful algorithms that run on a QC, and whether there is any prospect of a compiler that will let any old program (possibly written in a new language) to run.

I did come across one comment that academics like working on QC ideas because the subject is well funded even though they see the prospect of a viable machine as unlikely.

Do you know how the IBM machine represents its qubits - electron spins, nuclear spins, polarised photons, etc? Is there any convergence in the field towards one specific type of representation?

Also, since your answers are a bit terse, I'm wondering if your job restricts what you can divulge here.

David