On 04/12/2018 22:29, Gilles wrote:
I was in fact running Java 10 for the benchmarks: "BoxMuller" is indeed
50% faster on it than on Java 8 and "Marsaglia" 10% slower!
Interesting. Given the relative popularity of Java 10 vs 8 (Java 10 was
only out in March 2018 and usage statistics are unknown on various
summary surveys) I suggest either (a) stating the benchmark was done on
Java 10, or (b) changing back to the Java 8 benchmark.
Note that the detail at the top of the 'User Guide > 4. Performance'
section states OpenJDK 1.8 was used.
So, you'll change your vote?
Yes. +1 (non binding)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org
