At least right now, Apple only bins the M1 on graphics cores (7 vs. 8). My guess is that those have the most complex lithography, so it would make sense that th

At least right now, Apple only bins the M1 on graphics cores (7 vs. 8). My guess is that those have the most complex lithography, so it would make sense that those would be the biggest yield variable. My guess is that any multivariate binning will be lowest-common denominator, e.g. an 8-performance core / 12-GPU core part with a single, second-tier bin of 6 performance cores and 10 GPU cores.