On Feb 26, 2014, at 8:55 PM, Brian Burkhalter <[email protected]>
wrote:
>
> On Feb 26, 2014, at 5:38 AM, Paul Sandoz wrote:
>
>>>> Not sure the static powerCache field, in the original code, needs to be
>>>> volatile either:
>>>>
>>>> 1137 private static volatile BigInteger[][] powerCache;
>>>
>>> Is there consensus on whether "volatile" is necessary here?
>>>
>>
>> Looking back at the discussions i believe it was made volatile to ensure
>> threads don't observe a partially updated and published cache lines.
>>
>> Since we are already using Unsafe for deserialization I think it might be
>> possible to do the following instead (warning: not tested!):
>
> Thanks for the suggestion, Paul. Assuming it is correct, in what way would
> this be a better approach? (I apologize for being obtuse.)
>
IMHO it is a simpler solution. It's a cache line that requires updating not the
cache line*s*, as the current approach can potentially knock out cache lines
for other radixes.
> If it looks worth doing, I'll file another issue to track the idea. I already
> have it on my list anyway to follow up on Alan Eliasen's comment in the
> BigInteger code:
>
> * This could be changed to a more complicated caching method using
> * {@code Future}.
> */
> private static BigInteger getRadixConversionCache(int radix, int exponent)
> {
>
Not quite sure what that would entail.
Paul.
