On Fri, 12 May 2023 20:41:32 GMT, Chris Hennick <d...@openjdk.org> wrote:
>> This PR improves both the worst-case 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 distribution (probably >> starting around `2*exponentialX0 == 0x1.e46eff20739afp3 ~ 15.1`); this PR >> fixes that by tracking the multiple of exponentialX0 as a long. (This >> distortion is worst when `x > 0x1.0p56` since in that case, a rounding error >> means `extra + x == extra`. >> * Reduces several equations using `Math.fma`. (This will almost certainly >> improve performance, and may or may not improve output distribution.) >> * Uses the newly-extracted `computeWinsorizedNextExponential` function to >> prevent `nextGaussian` from going into the `nextExponential` tail twice. > > Chris Hennick has updated the pull request incrementally with two additional > commits since the last revision: > > - Merge remote-tracking branch 'origin/patch-1' into patch-1 > - Optimize: move some code out of the fast path Latest change looks good. I noted the improvement to the comment beginning "We didn't use the upper part of U1 after all". ------------- PR Comment: https://git.openjdk.org/jdk/pull/8131#issuecomment-1546292204
