Grounded0

i486 Microprocessor Programmer's Reference Manual Precision Control in the 80486 FPU:

If we're being hit with: `00` => hardware `10` => hardware `11` => hardware (x86-64) or software (everything else) I call this Floating Point Precision Predecoder Engine or FPPPE. 🙂

@dancek Intel Math Coprocessor Advanced Diagnostics: https://anttipeltola.eu/misc/files/MCPDIAG.ZIP

If i run Comprehensive Floating Point Conformance Test on macOS/x86-64 i get 0 errors currently on `main` on all the tests because its done on the x87 with up to...

Performance sucked so bad on the x87 until Pentium that programmers probably were wary of using `11` extended precision unless absolutely necessary. That's probably the reason we're seeing so few...

The `binary128` support would be currently aimed for `ppc64le` platform (POWER9+) which we also have as an CI runner. It's probably better to concentrate all resources to 80-bit extended precision...

Should we roll back to SDL 2.26.5 on Windows CI runner? Can someone check this on macOS too since i'm short on time due to other stuff. Someone should file...

> Quake and maybe a handful of simulators, that's it. Just keep that in mind before coming up with something massively overengineered 😄 This is Finland. 😁

@kklobe That's wonderful. That's without the Floating Point Precision Predecoder Engine (FPPPE) we're planning with @dancek to optimise performance. 👍

Start logging incoming `FPU_FADD`, `FPU_FSUB`, `FPU_FMUL`, and `FPU_FDIV` as pure count and: `FINIT` / `FNINIT` `FLDCW` / `FLDENV` `FRSTOR` / `FXRSTOR` for incoming control words for precision control so we...