In Qiskit, where is the floating point operation performed?

Question

I am using Qiskit - the simple circuits I am writing can either run on the hardware backend (using least_busy) or the QASM simulator backend.

If I have statements like the following - (random_numbers were generated using QRNG in the circuits - not shown here)

b = .6
a = .3
flt = []

for rn_num in random_numbers:
     flt.append( b*(a-rn_num) + a * rn_num)

Where are these floating point operations performed? I am connected to IBM Quantum Experience - using Jupyter. Are these operations performed on a traditional machine or a quantum backend?

My understanding is unless it is related to a Quantum circuit - everything gets executed on a traditional machine (similar to print or sockets?)

Is there a document that explains this further? I read release notes - I found them a bit ambiguous. I found this comment in Aer 0.3 release note

Increased performance with simulations that require less floating point numerical precision.

Maybe this note is referring to statevectors in floating point ...??

Answer 1

The Qiskit backends (quantum devices or simulators) work only when you explicitly invoke them, usually with execute. The code in your snippet does not call qiskit, and runs on a traditional machine.