RFR 8005311: Add Scalable Updatable Variables, DoubleAccumulator, DoubleAdder, LongAccumulator, LongAdder (original) (raw)

Peter Levart peter.levart at gmail.com
Fri Jan 11 16:35:30 UTC 2013

On 01/11/2013 05:18 PM, Chris Hegarty wrote:

Now with explicit disclaimer on DoubleA*

"The order of accumulation within or across threads is not guaranteed. Thus, this class may not be applicable if numerical stability is required when combining values of substantially different orders of magnitude."

It doesn't have to be substantially different order of magnitude.

For example:

     double a = 1d/3d;
     double b = 1d;

     double x = a + a + a + b;
     double y = b + a + a + a;

     System.out.println(x == y);

... prints false.

Regards, Peter

Updated spec: http://cr.openjdk.java.net/~chegar/8005311/ver.02/specdiff/java/util/concurrent/atomic/package-summary.html

Unless there are any objections, I'll finalize this spec and seek approval for its integration. -Chris. On 07/01/2013 19:40, Doug Lea wrote: On 01/07/13 14:07, Joe Darcy wrote: Hello,

I had a question about how the double accumulation logic was intended to be used. I've taken a quick look at the code and it uses straightforward "sum = sum + nextValue" code to compute the double sum. Summing doubles values with code numerical accuracy is surprisingly tricky and if the DoubleAccumulator code is meant for wide use, I'd recommend using instead some form of compensated summation: http://en.wikipedia.org/wiki/Kahansummationalgorithm

I'm sympathetic... Complete lack of control over arithmetic issues (here and for plain atomics) led me to resist offering Double versions for years. But many people are content with the scalability vs numerical stability tradeoffs intrinsic here (and I think unsolvable). I suppose it could stand an explicit disclaimer rather than the implicit one there now. How about: "The order of accumulation of sums across threads is uncontrolled. Numerical stability of results is not guaranteed when values of substantially different orders of magnitude are combined" Or something better? -Doug

More information about the core-libs-dev mailing list