RFR 7131192: BigInteger.doubleValue() is depressingly slow

[David Chase]

Not a reviewer, but depressingly much experience hacking Java FP.

It all looks right to me (i.e., exponent extraction and rounding modes including the round-up carry to the next higher power of two) and I like the comments explaining the little hacks, and I trust the testing to smoke out fencepost errors.

David

On 2013-06-18, at 8:45 PM, Brian Burkhalter <brian.burkhalter at oracle.com> wrote:

> The updated comments have been propagated to the online webrev
> 
> http://cr.openjdk.java.net/~bpb/7131192/
> 
> On Jun 17, 2013, at 11:39 AM, Brian Burkhalter wrote:
> 
>> I'll update this but without reposting the webrev until I can upload the HTML version.
>> 
>> On Jun 17, 2013, at 11:37 AM, Louis Wasserman wrote:
>> 
>>> You're right, it'd be WIDTH - 1, but since most of the comments refer to BITS + 1, that evens out nicely.
>>> 
>>> On Mon, Jun 17, 2013 at 11:34 AM, Brian Burkhalter<brian.burkhalter at oracle.com> wrote:
>>> Would it be WIDTH or WIDTH - 1, i.e., with or without the "implied" bit?
>>> 
>>> On Jun 17, 2013, at 11:32 AM, Louis Wasserman wrote:
>>> 
>>>> The comments mention SIGNIFICAND_BITS, which I think should probably be SIGNIFICAND_WIDTH?
>> 
>