SIMD-0401: Stake program Pinocchio migration (p-stake)

[grod220]

Proposal to replace the current version of the Stake Program (Stake11111111111111111111111111111111111111) with a CU-optimized Pinocchio version: p-stake.

[jstarry]

I don't feel convinced by the motivation section. Are stake instructions really using up so much block space that we need to optimize them? This feels like a much smaller win compared to optimizing the token program.

[grod220]

This feels like a much smaller win compared to optimizing the token program

True, but perhaps the SIMD should have put the second point in the motivation section first. There is a no-std wave coming where there will be increasing requirements to adopt it else be marooned w/ old tooling. Alignment with the upstream eBPF toolchain is a primary goal for folks working on the compiler/VM. Aligning with this earlier will allow us to not build new features that eventually have to be re-written in no-std rust. A quite nice side effect of this work is the CU savings.

[jstarry]

Ok, I'm definitely out of the loop on no-std plans but that feels like something we can do later in favor of doing other things that are more impactful to developers now.

[deanmlittle]

lgtm