So you just want VLIW architecture? It turns out it's better to have that abstract bottleneck ISA in practice, so CPU hardware can iterate microarchitectures wi

So you just want VLIW architecture? It turns out it's better to have that abstract bottleneck ISA in practice, so CPU hardware can iterate microarchitectures without breaking software compatibility. Intel learned about this hard way with Itanium.