“…Here, we used the listed benchmarks to compare the model sizes of XBMs and STGs, as well as compare the literal counts between the XBM tools (MINIMALIST and 3D) and the STG tools (PETRIFY and MPSAT via WORKCRAFT) based on their complex-gate implementations. These benchmarks include published XBMs found in the MINIMALIST framework [6], [23] (concur-mixer, dme, dram-ctrl, hp-ir, token-distributor, pe-send-ifc, pscsi, scsi, tangram-mixer), in the 3D [5] tool (biu-dma2fifo, diffeq, fifocellctrl), and in various publications (ack-xbm-si [24], biu-fifo2dma [5], imec-alloc-outbound [25], nowick-basic [4], pooffice-sbuf-send-ctl [17]). For fairness, we have factorised the literal counts produced by MINIMALIST and 3D using LOGIC FRIDAY [26] (a graphical frontend to the ESPRESSO logic minimiser [27]) before counting them, as they were originally in the sum-of-products form.…”