RFR: JDK-8316708: Augment WorstCaseTests with more cases [v7]

[Raffaello Giulietti]

On Tue, 27 Feb 2024 06:13:08 GMT, Joe Darcy <darcy at openjdk.org> wrote:

>> A new paper
>> 
>>  "Accuracy of Mathematical Functions in Single, Double, Double Extended, and Quadruple Precision"
>> by Brian Gladman, Vincenzo Innocente and Paul Zimmermann
>> https://members.loria.fr/PZimmermann/papers/accuracy.pdf
>> 
>> details the inputs with generate the worst-case observed errors in different math library implementations. The FDLIBM-related worst cases should be added to the test suite.
>
> Joe Darcy has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Appease jcheck.

test/jdk/java/lang/StrictMath/HypotTests.java line 711:

> 709:             // Empirical worst-case points in other libraries with
> 710:             // larger worst-case errors than FDLIBM
> 711:             {0x0.00000000039a2p-1022, 0x0.0000000000001p-1022, 0x0.00000000039a2p-1022},

I can't find this one on the paper.
Suggestion:

            {-0x0.fffffffffffffp-1022, 0x0.0000000000001p-1022, 0x0.fffffffffffffp-1022},

test/jdk/java/lang/StrictMath/Log1pTests.java line 220:

> 218:             {-0x1.2e496d25897ecp-2,     -0x1.663d81cb08f56p-2},
> 219:             {-0x1.ffffffbaefe27p-2,     -0x1.62e42faa93817p-1},
> 220:         };

I think there's another case
Suggestion:

            {-0x1.5efad5491a79bp-1022,     -0x1.5efad5491a79bp-1022},
        };

test/jdk/java/lang/StrictMath/LogTests.java line 74:

> 72:             {0x1.0ffea3878db6bp+0,   0x1.f07a0cca521fp-5},
> 73:             {0x1.490af72a25a81p-1,  -0x1.c4bf7ae48f078p-2},
> 74:             {0x1.69e7aa6da2df5p-1,  -0x1.634508c9adfp-2},

There are no libraries that have worse errors than OpenLibm, so I'm wondering what these values are good for?

test/jdk/java/lang/StrictMath/TrigTests.java line 171:

> 169:             {-0x1.0e16eb809a35dp+944, 0x1.b5e361ed01dadp-2},
> 170:             {-0x1.842d8ec8f752fp+21, -0x1.6ce864edeaffep-1},
> 171:             {-0x1.1c49ad613ff3bp+19, -0x1.fffe203cfabe1p-2},

Some of these cases are for libraries that have _better_ worst case errors than OpenLibm.
Are they needed here?

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/15879#discussion_r1504067869
PR Review Comment: https://git.openjdk.org/jdk/pull/15879#discussion_r1504067977
PR Review Comment: https://git.openjdk.org/jdk/pull/15879#discussion_r1504068053
PR Review Comment: https://git.openjdk.org/jdk/pull/15879#discussion_r1504068092