Fast homogeneous and rotation matrix multiplications - #1845

Draft

SamFlt wants to merge 26 commits into

lagadic:masterfrom

SamFlt:fast_homogeneous_proj

Commits on Nov 26, 2025

SIMD homogeneous matrix multiplicaiton
SamFlt
committed
711fe77

Commits on Nov 27, 2025

implement first version of simd for rotaiton matrix
SamFlt
committed
196660c

Commits on Dec 2, 2025

AVX implem for rotation matmul
SamFlt
committed
729110a
Fix intrinsics usage, performance improvement
SamFlt
committed
e230a72

Commits on Dec 3, 2025

Commits on Dec 4, 2025

Fix test, improve vpRBDenseDepth
SamFlt
committed
4c88d39
Remove debug prints
SamFlt
committed
77adf9d

Commits on Dec 5, 2025

Move initVVS to cpp file, resize matrix there
SamFlt
committed
1f3c7c8
Add ENABLE_NATIVE_ARCH option for gcc
SamFlt
committed
cc729b6
Remove reference to MBT tukey estimator, disable prints
SamFlt
committed
a05c329

Commits on Dec 6, 2025

Fix SSE3 flag check
SamFlt
committed
d847bfb

Commits on Dec 7, 2025

Merge branch 'master' into fast_homogeneous_proj
fspindle
committed
c58271a
Update copyright headers
fspindle
committed
c4bc02a
Fix warning unused variable
fspindle
committed
49e8c83
Remove useless empty lines
fspindle
committed
ae5966b
Fix warning variable set but not used
fspindle
committed
3f3bac2
Remove vpSIMD namespace from doxygen doc
fspindle
committed
adc26ab
Fix bug when input vector is not transposed and AVX or SSE2 not available
fspindle
committed
c6ff19a
Fix bug when input vector is not transposed and AVX or SSE3 not available
fspindle
committed
51e89a3
Cleanup tests to help debugging
fspindle
committed
09fc8cd
Fix _mm_hadd_pd() usage that requires SSE3 on pixi windows CI
fspindle
committed
f9f59e2

Commits on Dec 12, 2025