Estimating the Energy Crossover for Quantum Computing

A driving concern for beyond Moore’s Law computing is the need to reduce energy consumption across device types. In this contribution, we present a framework by which quantum computing performance is bench-marked by energy... [ view full abstract ]

A driving concern for beyond Moore’s Law computing is the need to reduce energy consumption across device types. In this contribution, we present a framework by which quantum computing performance is bench-marked by energy consumption. We present an analysis for Grover’s search algorithm for an unsorted database in comparison to a classical sequential search.
Our analysis considers how both approaches implement the query to the database. Because the cost of a query may effectively trump any speedup for the quantum algorithm, we define a break-even energy for the query cost as G_o=(2c/π)√N-G_q where c is the full life-cycle cost of a single examination of an element in the classical sequential search and G_q is the cost of the Grover iteration. We estimate this crossover by calculating expectations for c, G_q, and G_o culled from publicly available vendor data and theoretical proposals of quantum processing units. We find that as the size of the database increases, the crossover energy also increases and the breakeven effort point is found at a relatively small value of N. The resulting estimate for the energy provides an upper limit in terms of N and hardware constants that give system designers and engineers power budget targets.