]> git.lizzy.rs Git - minetest.git/commit
Inline triLinearInterpolationNoEase and triLinearInterpolation (#12421)
authorparadust7 <102263465+paradust7@users.noreply.github.com>
Sat, 11 Jun 2022 18:01:30 +0000 (11:01 -0700)
committerGitHub <noreply@github.com>
Sat, 11 Jun 2022 18:01:30 +0000 (20:01 +0200)
commit7ffc0268dfd78647187554d6248015329e9f5d2d
treee278298817298a549e227a1bd0e2bb3ff8ee1672
parente7d4ec6834282402c24fe432e0dcebe78b7fcd01
Inline triLinearInterpolationNoEase and triLinearInterpolation (#12421)

Performance profiling on Linux AMD64 showed this to be a significant bottleneck. The non-inlined functions are expensive due to XMM registers spilling onto the stack.
src/noise.cpp