> ... So your tight loops all look like "increment I; compare I to 0; compare I to X; do real work" ...
Compilers are generally smart enough to hoist these comparisons out of the loop, at least in a static, AOT-compiled language like Rust.
You're right though that in-order chips are a lot more power-efficient, and the in-order approach is the one that's taken in most RISC-V implementations (which seems to be a highly comparable architecture to the Alpha you mention).
Compilers are generally smart enough to hoist these comparisons out of the loop, at least in a static, AOT-compiled language like Rust.
You're right though that in-order chips are a lot more power-efficient, and the in-order approach is the one that's taken in most RISC-V implementations (which seems to be a highly comparable architecture to the Alpha you mention).