Hello Nathan,
We tried auto-vectorization and observed no performance improvement.
The instructions in src/include/port/simd.h are based on older SIMD architectures like NEON, whereas the patch uses the newer SVE, so some of the instructions used in the patch may not have direct equivalents in NEON. We will check the feasibility of integrating SVE in "src/include/port/simd.h" and get back to you.
The actual encoding/decoding implementation takes less than 100 lines. The rest of the code is related to config and the "choose" logic. One option is to move the implementation to a new file, making src/backend/utils/adt/encode.c less bloated.
Thanks,
Chiranmoy
