> This PR improves both the performance of `nextExponential` and `nextGaussian`
> and the distribution of output at the tails. It fixes the following
> imperfections:
>
> * Repeatedly adding DoubleZigguratTables.exponentialX0 to extra causes a
> rounding error to accumulate at the tail of the d
> This PR improves both the performance of `nextExponential` and `nextGaussian`
> and the distribution of output at the tails. It fixes the following
> imperfections:
>
> * Repeatedly adding DoubleZigguratTables.exponentialX0 to extra causes a
> rounding error to accumulate at the tail of the d
> This PR improves both the performance of `nextExponential` and `nextGaussian`
> and the distribution of output at the tails. It fixes the following
> imperfections:
>
> * Repeatedly adding DoubleZigguratTables.exponentialX0 to extra causes a
> rounding error to accumulate at the tail of the d
> This PR improves both the performance of `nextExponential` and `nextGaussian`
> and the distribution of output at the tails. It fixes the following
> imperfections:
>
> * Repeatedly adding DoubleZigguratTables.exponentialX0 to extra causes a
> rounding error to accumulate at the tail of the d
> This PR improves both the performance of `nextExponential` and `nextGaussian`
> and the distribution of output at the tails. It fixes the following
> imperfections:
>
> * Repeatedly adding DoubleZigguratTables.exponentialX0 to extra causes a
> rounding error to accumulate at the tail of the d
On Fri, 17 Jun 2022 12:59:17 GMT, Claes Redestad wrote:
> - Refactor micro using a range of parameters to a simpler benchmark that
> tests different values in one pass
> - Reduce iterations and their runtimes, but add forks (important to be able
> to detect run-to-run variation anomalies)
>
>
> This PR improves both the performance of `nextExponential` and `nextGaussian`
> and the distribution of output at the tails. It fixes the following
> imperfections:
>
> * Repeatedly adding DoubleZigguratTables.exponentialX0 to extra causes a
> rounding error to accumulate at the tail of the d
> This PR improves both the performance of `nextExponential` and `nextGaussian`
> and the distribution of output at the tails. It fixes the following
> imperfections:
>
> * Repeatedly adding DoubleZigguratTables.exponentialX0 to extra causes a
> rounding error to accumulate at the tail of the d
Hello.
I found suspicious null check in java.lang.invoke.ClassSpecializer constructor:
S topSpecies = null;
if (tsk != null && topSpecies == null) {
// if there is a key, build the top species if needed:
topSpecies = findSpecies(tsk);
}
'null' is assigned to the 'topSpecies' variable and
On Wed, 15 Jun 2022 19:49:52 GMT, Andrey Turbanov wrote:
> If a thread-safe implementation is not needed, it is recommended to use
> HashMap in place of Hashtable.
> `ConditionalSpecialCasing.entryTable` is read-only Map which is modified only
> in `static` block. It means we can safely replace
10 matches
Mail list logo
