[fulafel]

Speculative execution enables continuing along the predicted branch without stopping. You do need to have the ~2 instructions to get the length test input on hand but that usually can be eaten by insn level parallelism without hurting the latency of the array operation.

[errantmind]

Do you have any evidence of this claim? Perhaps a benchmark?

This doesn't align with any of my performance optimization experience.

[fulafel]

I went looking, and seems I have to walk my claim back somewhat. Wide issue OoO processors hide a lot of the overhead but not all of it.