Review Request: BigInteger patch for efficient multiplication and division (#4837946)
brian.burkhalter at oracle.com
Mon May 13 23:44:29 UTC 2013
On May 13, 2013, at 11:57 AM, Brian Burkhalter wrote:
> On May 11, 2013, at 8:35 PM, Alan Eliasen wrote:
>> On 05/09/2013 03:02 PM, Brian Burkhalter wrote:
>>>> First you have:
>>>> * The threshold value for using 3-way Toom-Cook multiplication.
>>>> * If the number of ints in both mag arrays are greater than this number,
>>>> * then Toom-Cook multiplication will be used. This value is found
>>>> * experimentally to work well.
>>>> private static final int TOOM_COOK_THRESHOLD = 75;
>> You're right that the actual code will use Toom-Cook if 1.) both of
>> the numbers are greater than the Karatsuba threshold *and* 2.) at least
>> one of the numbers is greater than the Toom-Cook threshold.
>> If you want to change the comment to something like my first sentence
>> in the first paragraph, that would be fine. Alternately, we could
>> change the logic to match the comment, but that would probably mean that
>> we should re-tune the thresholds.
> I would prefer simply to change the javadoc of the constant unless others have a strong preference otherwise.
* The threshold value for using 3-way Toom-Cook multiplication.
* If the number of ints in each mag array is greater than the
* Karatsuba threshold, and the number of ints in at least one of
* the mag arrays is greater than this threshold, then Toom-Cook
* multiplication will be used.
private static final int TOOM_COOK_THRESHOLD = 75;
More information about the core-libs-dev