This allows most demos to run without any ubsan or asan errors. There
are still some things in thirdpart/ and some things in AudioServer that
needs a look but this fixes a lot of issues. This should help debug less
obvious issues, hopefully.
This fixes#25217 and fixes#25218
We fork off twice once with and once without DIR_PRIME=1 set. We
then use the vendor string to determine what GPU to use.
We prefer (in order)
1) AMDGPU/AMDGPU-PRO/NVidia non-free driver
2) Intel driver
3) Nouveau
4) Software rendering
If a driver can't be detected it will default to DRI_PRIME=0
While looking into a different issue, I've noticed that Visual Studio Intellisense does not work well for Godot project when using Windows Vista+ APIs (e.g. CreateThreadpool), i.e. it does not recognise the APIs because they are defined in Windows header files for Vista+ only.
This is because the WINVER and _WIN32_WINNT symbols don't have their values set in the generated Godot project file. This fixes the problem by setting the values when generating the project file.
Make the sanitizer names more explicit (use_ubsan, use_asan, use_lsan).
Comment has been adjusted to include GCC as supported compiler for these
and exclude -fno-omit-frame-pointer option (should not cause any
problems).
From August 1, 2019, Google Play requires that all new apps and app updates
include 64-bit versions, so we enable ARM64 by default.
IINM support for x86 and x86_64 is still be optional, so not enabling them
out of the box.
Part of #25030.
Godot supports many different compilers and for production releases we
have to support 3 currently: GCC8, Clang6, and MSVC2017. These compilers
all do slightly different things with -ffast-math and it is causing
issues now. See #24841, #24540, #10758, #10070. And probably other
complaints about physics differences between release and release_debug
builds.
I've done some performance comparisons on Linux x86_64. All tests are
ran 20 times.
Bunnymark: (higher is better)
(bunnies) min max stdev average
fast-math 7332 7597 71 7432
this pr 7379 7779 108 7621 (102%)
FPBench (gdscript port http://fpbench.org/) (lower is better)
(ms)
fast-math 15441 16127 192 15764
this pr 15671 16855 326 16001 (99%)
Float_add (adding floats in a tight loop) (lower is better)
(sec)
fast-math 5.49 5.78 0.07 5.65
this pr 5.65 5.90 0.06 5.76 (98%)
Float_div (dividing floats in a tight loop) (lower is better)
(sec)
fast-math 11.70 12.36 0.18 11.99
this pr 11.92 12.32 0.12 12.12 (99%)
Float_mul (multiplying floats in a tight loop) (lower is better)
(sec)
fast-math 11.72 12.17 0.12 11.93
this pr 12.01 12.62 0.17 12.26 (97%)
I have also looked at FPS numbers for tps-demo, 3d platformer, 2d
platformer, and sponza and could not find any measurable difference.
I believe that given the issues and oft-reported (physics) glitches on
release builds I believe that the couple of percent of tight-loop
floating point performance regression is well worth it.
This fixes#24540 and fixes#24841
