Hi, I am trying to construct specific unitary for initial state that fits my needs and goal is to simulate a parameterized circuit starting from that initial state.
And I’ve found 16 qubit unitary is numpy array with its size about 64 GB.
So my workstation has 256GB rams, considering simulating circuit is done by multiplying matrices,
- I wonder how that is feasible,
I’ve seen many others in thins forum who simulated 16+ qubit circuits with own laptops.
Hey @K_Hwang, welcome to the forum!
There’s a lot of tricks under the hood that let us do matrix-matrix multiplication with an amount of memory that seems impossible. But, these tricks are battle tested and quite old! Check this out: Basic Linear Algebra Subprograms - Wikipedia. BLAS, or some variation of it, is at the heart of any linear algebra operation that a computer performs. 
Let me know if that helps!
1 Like
Appreciate your help, your answer made me clear.
I have no backgrounds in CS, ingenious ideas always surprises me.