The proof for putnum was inductive, with the base case being putnum(value,base) working correctly when 0 <= value < base, and the inductive step assuming that the function output the correct digit string for values of value where 0 <= value < basek, and showing that the output is still correct for values of value where basek <= value < basek+1.
To see the inductive step, consider the number
The proof for getnum was based on the observation that
New material covered in class: pipelining. See Chapter 2 of Kane & Heinrich to see the R2000/3000 pipeline.
The participatory demonstration that I used in class was the computation of
Latency and throughput were defined in class: latency is the time until the output is done, and throughput is the aggregate rate at which output appears (this is sometimes an asymptotic value -- supercomputer manufacturers talk about ``Peak GFLOPS'' -- and more properly it is measured based on ``average-case input''.)
I started to talk about data and control hazards, and will go over this again next time.
bsy@cse.ucsd.edu, last updated