Bring the optimizations from https://github.com/rapidsai/cuvs-lucene/issues/158 into cuVS-Lucene
Bring the optimizations from #158 into cuVS-Lucene