Adding RTX 4090 benchmarks, architecture diagram

README.md

@@ -6,6 +6,8 @@ This is an integration for [CuVS](https://github.com/rapidsai/cuvs), GPU acceler
 
 As an initial integration, the CuVS library is plugged in as an IndexSearcher. This project has two layers: (1) Java/JNI layer in `lucene` dir, (2) CuVS/C++ layer in `cuda` dir.
 
+![Architecture](architecture.png "Lucene CuVS Architecture")
+
 By way of a working example, OpenAI's Wikipedia corpus (25k documents) can be indexed, each document having a content vector. A provided sample query (query.txt) can be executed after the indexing.
 
 > :warning: This is not production ready yet.
@@ -24,10 +26,11 @@ Set the correct path for Raft in `cuda/CMakeLists.txt` file. Then, proceed to ru
 
 Wikipedia (768 dimensions, 1M vectors):
 
-|                           | Indexing   | Improvement | Search | Improvement |
-| ------------------------- | ---------- | ----------- | ------ | ----------- |
-| CuVS (RTX 2080 Ti)        | 37.83 sec  |  **26x**    |  2 ms  |   **4x**    |
-| Lucene HNSW (Ryzen 7700X) | 992.37 sec |      -      |  8 ms  |      -      |
+|                                | Indexing   | Improvement | Search | Improvement |
+| ------------------------------ | ---------- | ----------- | ------ | ----------- |
+| CuVS (RTX 4090, NN_DESCENT)    | 38.80 sec  |  **25.6x**  |  2 ms  |   **4x**    |
+| CuVS (RTX 2080 Ti, NN_DESCENT) | 46.18 sec  |  **21.5x**  |  3 ms  |   **2.7x**  |
+| Lucene HNSW (Ryzen 7700X)      | 992.37 sec |       -     |  8 ms  |      -      |
 
 
 ## Next steps