I was thinking specifically of Apple's M1 and recent upgrades from AMD and Intel, which all "go wider" (more execution units) and it works well for them. In som

I was thinking specifically of Apple's M1 and recent upgrades from AMD and Intel, which all "go wider" (more execution units) and it works well for them. In some cases, these units can't even possibly be fed in steady state, only e.g. while catching up from some misprediction or cache miss.