mccullocht commented on PR #15742:
URL: https://github.com/apache/lucene/pull/15742#issuecomment-3956928661
1M cohere 1024d vectors, dot_product.
```
Baseline Results:
recall latency(ms) netCPU avgCpuCount nDoc topK fanout maxConn
beamWidth quantized visited index(s) index_docs/s force_merge(s)
num_segments index_size(MB) filterStrategy filterSelectivity overSample
vec_disk(MB) vec_RAM(MB) bp-reorder indexType
0.914 2.317 2.309 0.997 1000000 100 100 64
250 4 bits 8707 119.32 8380.54 213.79
1 4559.57 null N/A 1.000 4409.790
503.540 false HNSW
Experiment Results:
recall latency(ms) netCPU avgCpuCount nDoc topK fanout maxConn
beamWidth quantized visited index(s) index_docs/s force_merge(s)
num_segments index_size(MB) filterStrategy filterSelectivity overSample
vec_disk(MB) vec_RAM(MB) bp-reorder indexType
0.914 2.194 2.187 0.996 1000000 100 100 64
250 4 bits 8707 119.32 8380.54 213.79
1 4559.57 null N/A 1.000 4409.790
503.540 false HNSW
```
Looks like about 5%, roughly the same as on Graviton 3.
Targeting 10.5 SGTM.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
