2
$\begingroup$

I've been studying recently some variations of the CORDIC, and it seems to me that the logic behind at least the basic cordic or the redundant CORDIC is very similar to the logic used to design digit recurrence for the division (using redundant digit set). Is this statement correct?

At each iteration they have to guess a digit to construct the final answer, I would also say that the division is a particular case of the CORDIC iteration, indeed the CORDIC used in linear mode actually allows to retrieve the iteration for the division.

Therefore I'd say that a good understanding of the division is essential for understanding CORDIC, because it teaches you how to use Robertson and PD diagrams, tools that are used also in the Redundant CORDIC for example.

Are these observations correct?

$\endgroup$
  • $\begingroup$ (1) Yes. (2, third paragraph) No. (3) Yes & No = No. I am not sure what exactly the answer should look like. Basically the yes/no answers are not good fit. In meanwhile you may be interested in division CORDIC. $\endgroup$ – Evil Jun 6 '17 at 5:36
  • $\begingroup$ Can you elaborate a bit more? $\endgroup$ – user8469759 Jun 6 '17 at 10:57
  • $\begingroup$ Yes, this is the same algorithm, very good insight. But there is a small division method (like Newton) that checks on which side of the true value is at given step and then modifies the next move. It is not "guess", just preparation for the next step. Volders algorithm was invented in 1958, Robertson in 1956. Since both use the same idea, it is hard to imply that understanding of CORDIC is essential to understand Robertson graph, which may be learned independently of CORDIC. Using PD in that context is a stretch, I do not see any insight from CORDIC that helps. $\endgroup$ – Evil Jun 6 '17 at 19:45
  • $\begingroup$ And the full PD is not built in CORDIC, it is not used, and if it is, I must say that your insight is greater than mine understanding of the algorithm. And now, back to the question, you know that these are connected (for sure it is true), but it seems that you seek confirmation that learning one boosts the understanding of the second. I disagree here, and I think that your question should be rephrased to be objectively answeable, without "y/no" or opinion about the insight. And for what its worth, I did not wanted to be impolite if this sounded that way, I honestly do not know what you expect $\endgroup$ – Evil Jun 6 '17 at 19:51
  • $\begingroup$ @Evil I was actually the other way around. Namely understanding the high radix division could make easier understanding high radix CORDIC. Digits choice in high radix divison are performed using the LUT derived by stuydying robertson and PD diagrams, I've some reference that use the Robertson diagrams in redundant cordic as well, but no one mention whether or not there's a relationship between the two algorithms and related design techniques. Therefore I'm asking if my observations are correct or not. $\endgroup$ – user8469759 Jun 6 '17 at 20:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.