The article wasn’t great at laying out the concepts at the start. As I understand it, the big idea is essentially a bloom filter as the first phase of a retrieval.
The SWAR (SIMD-within-a-register) numbers are strictly better than the SIMD versions as well as the standard library baseline. Why is that? SIMD should be strictly faster if the machine supports it, since the SWAR max bitwidth is 64, while SIMD starts at 128 bits.
The Java SIMD API used here must not result in using actual SIMD machine code.
The Java SIMD API used here must not result in using actual SIMD machine code.