filter_speed_benchmark

Qdrant memory benchmark

• Estimate speed with and without filters

Data

Randomly generated vectors

Usage

Install dependencies

pip install poetry
poetry install

Generate data

python -m benchmark.generate_data -n 1000000 -d 64 -q 1000 -p 5

Run Qdrant instance:

docker-compose up

Upload search data and build index for field a

python -m benchmark.upload_data -i a

Uploaded data available for search immediately, but building of HNSW index may take quite some time.

Run search benchmark

# Without filters
python -m benchmark.search
# With filters
python -m benchmark.search -q a

Example Results

Qdrant params:

• Num parallel searchers: 4
• Num parallel queries: 4

For GloVe Angular 100 dataset:

• num_vectors = 1.000.000
• dim = 64
• metric = cosine

Used HNSW index params:

• M = 16
• efConstruct = 100
• ef = 100

Speed without filters

total time = 2.240 sec
time per query = 0.0022 sec
query latency = 0.0089 sec

Speed with filter

total time = 4.666 sec
time per query = 0.0047 sec
query latency = 0.0184 sec