Apache
/
lucene
mirror of https://github.com/apache/lucene.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Add comment to the SIMD intersection logic. (#14073) 

I did not know it when I checked in the code, but this is almost exactly the v1
intersection algorithm from the "SIMD compression and the intersection of
sorted integers" paper.
This commit is contained in:
Adrien Grand 2024-12-23 09:38:50 +01:00 committed by GitHub
parent 8c7050b428
commit 42c4c115d6
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 4 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
lucene/core/src/java21/org/apache/lucene/internal/vectorization/PanamaVectorUtilSupport.java
View File

@ -771,6 +771,10 @@ final class PanamaVectorUtilSupport implements VectorUtilSupport {
@Override
public int findNextGEQ(int[] buffer, int target, int from, int to) {
if (ENABLE_FIND_NEXT_GEQ_VECTOR_OPTO) {
// This effectively implements the V1 intersection algorithm from
// D. Lemire, L. Boytsov, N. Kurz SIMD Compression and the Intersection of Sorted Integers
// with T = INT_SPECIES.length(), ie. T=8 with AVX2 and T=16 with AVX-512
// https://arxiv.org/pdf/1401.6399
for (; from + INT_SPECIES.length() < to; from += INT_SPECIES.length() + 1) {
if (buffer[from + INT_SPECIES.length()] >= target) {
IntVector vector = IntVector.fromArray(INT_SPECIES, buffer, from);