Emerging post-CMOS switch options - a product reliability perspective

Summary form only given: Sooner or later, CMOS scaling will come to an end. What do we do next? A number of very different switches have been proposed as replacements, some of which in fact do not even use electron charge as the state variable. Instead, these switches pass tokens in the spin, excitonic, photonic, magnetic, qubit, or heat domains. The emergent physical behaviors and idiosyncrasies of these novel switches can compliment the execution of specific task algorithms or workloads, and improve overall thru-put in high performance computing. They also present new reliability challenges and perils, however. This talk will describe some of these potential CMOS replacements, how they may be used in design, and the concerns they raise for those who manage product reliability.