To provide a seed value for a Newton-Raphson routine to quickly calculate a sufficiently-accurate value. Per notes in the RVV spec:
The 7 bit accuracy was chosen as it requires 0,1,2,3 Newton-Raphson iterations to converge to close to bfloat16, FP16, FP32, FP64 accuracy respectively.
To provide a seed value for a Newton-Raphson routine to quickly calculate a sufficiently-accurate value. Per notes in the RVV spec: