[libc-commits] [libc] [llvm] [libc][math] Implement double-precision acosh (PR #199953)

[via libc-commits]

zimmermann6 wrote:

I get the following failures with the CORE-MATH test suite:
```
Running worst cases check in --rndn mode...
FAIL x=0x1.0001cd7360138p+0 ref=0x1.e6114b40aa763p-8 z=0x1.e6114b40aa764p-8
FAIL x=0x1.0006a6c8e5b84p+0 ref=0x1.d2daa7159c86fp-7 z=0x1.d2daa7159c87p-7
FAIL x=0x1.297de35d02e9p+13 ref=0x1.3b562d2651a5dp+3 z=0x1.3b562d2651a5ep+3
Running worst cases check in --rndz mode...
FAIL x=0x1.0381e68cac923p+0 ref=0x1.52a11832e847cp-3 z=0x1.52a11832e847dp-3
FAIL x=0x1.061f4c39e16f2p+0 ref=0x1.bf0305e2c6c36p-3 z=0x1.bf0305e2c6c37p-3
FAIL x=0x1.02125da072322p+0 ref=0x1.0460265bb6594p-3 z=0x1.0460265bb6595p-3
Running worst cases check in --rndu mode...
FAIL x=0x1.05fb3d7a14c4ep+0 ref=0x1.b9dc255b60a03p-3 z=0x1.b9dc255b60a04p-3
FAIL x=0x1.061f4c39e16f2p+0 ref=0x1.bf0305e2c6c37p-3 z=0x1.bf0305e2c6c38p-3
FAIL x=0x1.04ba7bebbed46p+0 ref=0x1.8905f76a6135ep-3 z=0x1.8905f76a6135fp-3
FAIL x=0x1.a5bf3acfde4b2p+0 ref=0x1.1585720f35cd9p+0 z=0x1.1585720f35cdap+0
FAIL x=0x1.0000000000048p+0 ref=0x1.7fffffffffff8p-23 z=0x1.800096a830b71p-23
Running worst cases check in --rndd mode...
FAIL x=0x1.0381e68cac923p+0 ref=0x1.52a11832e847cp-3 z=0x1.52a11832e847dp-3
FAIL x=0x1.061f4c39e16f2p+0 ref=0x1.bf0305e2c6c36p-3 z=0x1.bf0305e2c6c37p-3
FAIL x=0x1.02125da072322p+0 ref=0x1.0460265bb6594p-3 z=0x1.0460265bb6595p-3
```

https://github.com/llvm/llvm-project/pull/199953