Adds support for AVX2 for 32-bit types for quicksort and quickselect - #60

Merged

r-devulap merged 19 commits into

sterrettm2:avx2-32bit

Oct 24, 2023

Commits on Oct 20, 2023

Core AVX2 code logic

sterrettm2
authored and
Raghuveer Devulapalli
committed
f986043
Bug fix in runtime dispatch and variadic args
Raghuveer Devulapalli
committed
68938d4
Bugfixes in AVX2 code

sterrettm2
authored and
Raghuveer Devulapalli
committed
73b9ba5
Removed some unused code, changed how single compresstore works for AVX2

sterrettm2
authored and
Raghuveer Devulapalli
committed
0f3c138
Some small bug fixes

sterrettm2
authored and
Raghuveer Devulapalli
committed
55a6b51
Removes unnecessary forward includes

sterrettm2
authored and
Raghuveer Devulapalli
committed
b02fe0a
Fixed types in a few places

sterrettm2
authored and
Raghuveer Devulapalli
committed
dec2278
Make pivot selection simpler
Raghuveer Devulapalli
committed
0e41ebc
Fixes/changes many small things

sterrettm2
authored and
Raghuveer Devulapalli
committed
15d6025
Fix rebase merge bugs
Raghuveer Devulapalli
committed
4fa2255
Move avx2_vector declaration to common includes
Raghuveer Devulapalli
committed
3a3baae
Add avx2 to examples
Raghuveer Devulapalli
committed
2f397e4
Changed partition code
sterrettm2
committed
3561db3

Commits on Oct 23, 2023

Get rid of the avx2_mask_helper
Raghuveer Devulapalli
committed
9ad4432
format files
Raghuveer Devulapalli
committed
f8f611f
Wrapper function for qsort, qselect and partialsort
Raghuveer Devulapalli
committed
d032bc4
Header include directives
Raghuveer Devulapalli
committed
191ef61
Remove need for avx2-32bit-common.h
Raghuveer Devulapalli
committed
3dab097
Explicit instantiation of avx512_partial_qsort for _Float16
Raghuveer Devulapalli
committed
9810e05