[Xapian-discuss] Benchmark (index, search) on 20 million docs Xapian/Linux vs. Lucene/Windows

[Kevin Duraj]

Benchmark performed on 2 Dual Core AMD
Opteron Processor with 16 GB of memory
***************************************************************
Documents used:
  - number of documents = 19999995
  - average document length = 8.18631
***************************************************************
Indexing 20 million docs
   - Xapian 0 hours 47 minutes
   - Lucene 3 hours 30 minutes

***************************************************************
Result: Xapian/Linux showed to be 77% faster
then Lucene/Windows on indexing.
***************************************************************
Searching 20 million docs:
  - Xapian average  56 ms
  - Lucene average 87 ms

Same 50,000 unique boolean queries were applied to both
Xapian and Lucene, between 100-1000 terms per search.
***************************************************************
Result: Xapian/Linux showed to be 35% faster
then Lucene/Windows on searching.
***************************************************************

Disclaimer: This benchmark results will not be exactly same for your
type of documents and definitely your should make your own research
before making any decision based on this benchmark.

Thanks,
- Kevin Duraj