Pavel Borisov 3950bc3dc6 Speed up ivfflat build: use float instead of double for dot product (#180) ...

calculation

On ARM this makes CPU using vector multiply-add instruction (fmadd)
instead of vector multiplication + conversion to double + addition
(fmul + fcvt + fadd) at each vector dimension.

Output of distance functions and calculations that are are done once
per vector pair are left double as this don't make speed difference
and for compatibility.

2023-07-18 12:55:38 -07:00

..

ivfbuild.c

Improved includes

2023-07-17 13:17:48 -07:00

ivfflat.c

Moved _PG_init

2023-07-16 20:03:34 -07:00

ivfflat.h

Improved includes

2023-07-17 13:15:10 -07:00

ivfinsert.c

Removed unneeded comments [skip ci]

2023-07-17 13:08:27 -07:00

ivfkmeans.c

Improved variable scoping

2023-07-15 18:53:41 -07:00

ivfscan.c

Improved includes

2023-07-17 13:17:48 -07:00

ivfutils.c

Removed unused variable from IvfflatUpdateList

2023-07-17 00:15:20 -07:00

ivfvacuum.c

Improved variable name [skip ci]

2023-07-17 00:18:14 -07:00

vector.c

Speed up ivfflat build: use float instead of double for dot product (#180)

2023-07-18 12:55:38 -07:00

vector.h

Moved _PG_init

2023-07-16 20:03:34 -07:00