[General boards] [Fall 2018 courses] [Summer 2018 courses] [Winter 2018 courses] [Older or newer terms]

Question About Assignment 1, Question 3


#1

I am working on computing the scaled efficiency in question (3.b), and I am a bit confused about the data available in the table.

Namely, I am using the process described in the lecture notes (pages 1-69–1-72) to compute $n_4, n_8$, and $n_{16}$, which are required in the computation of scaled efficiency.

However, given that matrix-matrix multiplication has cubic complexity, I end up having to take the cubic root of $p$, which means I end up with values for $n_4, n_8$, etc. which are not in the table.

Should I simply approximate to the closest values in the table, or is there something else that I may be missing?

Alternatively, if we were to follow the example from pages 1-78 and 1-79, is it sufficient to only provide an analytical formula for the resulting efficiency?

Thank you,
Andrei


#2

We talked about this in class. See the different formula
of page 71 versus page 69.
Essentially the formula of pg 71 is a generalization of the one in pg 69. It is fine to approximate, as long as you correct your approximation. You need to give a number eventually.


#3

Thank you for the clarifications!

I had misunderstood the question, and apparently misread the requirement too—we need n_64, not n_16 in (b), which makes the math much nicer.