4 ms·
Historically the delta between 16 and 32 GPRs has not made a huge difference to most algorithms (where the move from 8 to 16 was worth maybe 20% on typical code
by ajross 13y ago
Historically the delta between 16 and 32 GPRs has not made a huge difference to most algorithms (where the move from 8 to 16 was worth maybe 20% on typical code). Obviously there are outliers, but no: this was all hashed out in the early days of x86_64, when the early Athlons were turning in per-transistor-cycle performance numbers at or above what was seen in the RISC world.
Basically: more registers means less spilling and filling to memory, but more work in context switches and for non-leaf function calls to save the extra data. It's all about tradeoffs. Note that no one ever went to a 64 GPR ISA (Itanium and Sparc have more than 32, but they do it with a stack engine that takes the work off the user code for the most part).
ARMv8 is clean and nice, and 64 bit computation is clearly the future just as it is in the desktop world. But it's not a revolution (again, just as it wasn't in the desktop world) and we shouldn't try to make it one.