| Re: [hatari-devel] FPU update |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/hatari-devel Archives
]
I have an update regarding SoftFloat FPU implementation:
After lots of testing with Toni and quite some work with FPSP we now have support for packed decimal to binary float conversion. It is exact on the bit level with correct OPERR, INEX1 and INEX2 exceptions.
This means we have bit level precise SoftFloat functions for everything except FACOS, FASIN, FATAN, FATANH, FCOS, FCOSH, FETOX, FETOXM1, FLOG10, FLOG2, FLOGN, FLOGNP1, FSIN, FSINCOS, FSINH, FTAN, FTANH, FTENTOX and FTWOTOX. After successfully translating FPSP bindec (maybe the most complicated part of FPSP) to SoftFloat i am quit optimistic that this can also be done for all functions still missing.
Maybe someone wants to jump into this. Once we have translated those, all FPU emulation will be fully independent from host floating point formats (SoftFloat does not use any floats).