Daniel Lemire

See blog post [Faster sorted array unions by reducing branches](https://lemire.me/blog/2021/07/14/faster-sorted-array-unions-by-reducing-branches/)

Parallelize

2

The Go version has been parallelized https://github.com/RoaringBitmap/roaring/pull/109

Replace ROARING_DISABLE_AVX and ROARING_DISABLE_NATIVE options

1

@andreigudkov [proposed we do the following](https://github.com/RoaringBitmap/CRoaring/pull/177#issuecomment-410769028): * Remove entirely ROARING_DISABLE_AVX and ROARING_DISABLE_NATIVE options * For conditional compilation, use #ifdef __AVX2__ * To select target architecture, use CMAKE_C_FLAGS I think that...

Modernize LTO support in CMake

1

Should be as simple as... ``` cmake_minimum_required(VERSION 3.9) # CMP0069 NEW include(CheckIPOSupported) check_ipo_supported(RESULT ltoresult) if(ltoresult) set(CMAKE_INTERPROCEDURAL_OPTIMIZATION TRUE) endif() ```

For dense bitsets, we should use SIMD instructions. See: Iterating over set bits quickly (SIMD edition) https://lemire.me/blog/2018/03/08/iterating-over-set-bits-quickly-simd-edition/ Note that it is not a good idea for sparse bitsets.

Gather some test coverage statistics

5

Maybe like so? http://stackoverflow.com/questions/13116488/detailed-guide-on-using-gcov-with-cmake-cdash

Microsoft launched a package manager that should be supported by this library... https://github.com/Microsoft/vcpkg

Stats are obviously wrong (C++)

2

I have many more than 1531 stars in my C++ projects... http://git-awards.com/users/lemire E.g., a single one of my projects exceeds 6000 starts... https://github.com/lemire/simdjson

We have a vectorized encoder: https://github.com/lemire/streamvbyte