Talk:Booth's multiplication algorithm: Difference between revisions

Revision as of 09:24, 5 September 2013

Latest revision as of 00:20, 12 February 2024

(7 intermediate revisions by 6 users not shown)

Line 1:

{{WikiProject Computing}}banner shell|class=C|

{{WikiProject Computing|importance=Low}}

{{WikiProject Computer science|importance=Mid}}

}}

==Reference to Booth's paper==

Here is a reference to Booth's paper:

Line 43 ⟶ 46:

:What is the additional bit initialized to? Reinderien 04:37, 2 December 2007 (UTC)

:I added some clarification to that part, based on what I believe is a correct interpretation: the typical implementation states, "A: Fill the most significant (leftmost) bits with the value of m. Fill the remaining (y + 1) bits with zeros." It seems to be the case that, should you add more bits to the front, you still fill all of them with the value of '''m'''--you're just filling more bits with the same value now, so the sign bit will be further out, and you'll have an extra '1' for each bit you add, to balance out the fact that the sign bit is now more significant. [[User:KoriganStone|KoriganStone]] ([[User talk:KoriganStone|talk]]) 20:19, 20 August 2014 (UTC)

==Number of times to perform the loop==

Line 70 ⟶ 75:

== Unsigned case ==

For unsigned case, it's simply increase that to 5 x 5 signed bit, so a 15 x 15 would work.

As the statement of this article, it's only work for signed 2's complement, I wonder if this could work for unsigned values.

But keep in mind, if you are work on radix-4 or high order of design, let say a 8 x 8 in radix-4 method.

You are actually working on a 9 x 9 case instead, and the last cycle will be radix-2 only, yes that's odd.

So for a 64 x 64 => 65 x 65 case, the chose of radix-?? will be difficult.

<span style="font-size: smaller;" class="autosigned">— Preceding [[Wikipedia:Signatures|unsigned]] comment added by [[Special:Contributions/192.165.214.193|192.165.214.193]] ([[User talk:192.165.214.193|talk]]) 09:24, 5 September 2013 (UTC)</span><!-- Template:Unsigned IP --> <!--Autosigned by SineBot-->

The examples in this article will NOT work for maximum unsigned values ie 15 x 15.

== ambiguous dots in formula ==

Is there modified version for this algorithm for unsigned?

The dots in the formula

Thanks.

https://wikimedia.org/api/rest_v1/media/math/render/svg/b66a7ba976185c7dc2ad64dd760ff420be473e8e

mean very different things depending on where they are. This could be clarified. "Whatever bits in this ńumber", "same bits as around the dots", "same bits as in the preceding binary number" etc.