1 branch per cycle
no delay slots
static branch prediction bit (e.g., bne+ bne-)
3 cycle mispredict penalty
The only non-pipelined
instructions in the ISA.
Uses FD and HI/LO registers
with full/empty bits.
Proc, IntDiv, FPDiv State
Machines run independently
Processor will only stall if
you read MFFD, MFLO, MFHI
before dividers have finished.