qEOM taking too long to run for BeH2 (STO-3G minimal basis). #1387
Description
Hi, I'm currently running a qEOM simulation for BeH2 in the STO-3G basis. I've been able to get the qubit count down to 8 by first freezing core electrons and thereafter employing the parity mapping scheme. I use VQE (with UCCSD) to compute the ground state, which is thereafter fed into qEOM to construct excited states. I've noticed that for BeH2 (and similar 8-qubit systems), qEOM takes forever to run. These simulations are being on AerSimulator as the backend with the MPS method, which in principle should be more efficient than the dense statevector simulator (the backend options are passed to the Estimator primitive). Is there a reason as to why qEOM is taking so long to run?