We have ARM assembly for union2by2 as per #287
It should be said that there are vectorized algorithm (hint: ARM NEON) for union routines. It appears in the C code (for SSE only but it is portable).
https://github.com/RoaringBitmap/CRoaring/blob/master/src/array_util.c#L1569-L1647
It is described in section 4.3 of the following paper:
It is actually not difficult to implement. For someone that is either fluent in ARM NEON or wants to learn... it is a great project.
We have ARM assembly for union2by2 as per #287
It should be said that there are vectorized algorithm (hint: ARM NEON) for union routines. It appears in the C code (for SSE only but it is portable).
https://github.com/RoaringBitmap/CRoaring/blob/master/src/array_util.c#L1569-L1647
It is described in section 4.3 of the following paper:
It is actually not difficult to implement. For someone that is either fluent in ARM NEON or wants to learn... it is a great project.