Marcel
Marcel
Simple fix proposal that works, since everything else looked tedious to check also w. r. t. compiler settings and compilation times on RPi: 1. Make `NormalizeDeltas` a template function. 2....
I'm seeing `float_t`-users in `Map.cpp` and `Scriptable.cpp`.
When I discovered that, I wrote a test on Linux on both x86_64 and RPi-ARM that report `sizeof(float_t)` and `FLT_EVAL_METHOD` with `4, 0`, with `-O3 -ffast-math` GCC v12/13. I assume...
What did I do: rotate vector (2, 2) by 1° up to 360 times and then put x, y every time into `NormalizeDeltas`. One calculation is done over float types,...
* `-ffast-math -frounding-math`: ARM and x86_64 show different results in the float calculation * `-frounding-math`: ARM and x86_64 show the same differences, and on each platform there is the same...
`std::hypot` generates some diff noise when using it, in the less significant parts of course but I better don't try to analyze the numerical impact here now.
Looking at godbolt.org assembly output for such a check, that does not seem to make any difference (using `-ffast-math` or not).
I guess so.
Sounds somewhat counter-intuitive to fix a problem of small errors with even bigger errors? Also, not sure if `NormalizeDeltas` is even the root cause here, given that the numbers we...
> I don't know why you're insisting the problem is one of precision. just because the problem "goes away" when you use a double means nothing. In 2nd paragraph I...