← all runs

Cohere-10M — 2026-06-12

r8i.16xlarge · peak recall 0.9957 @ 48 QPS

QPS at recall

Recallmkt-n12000-cctrue-fs16-soarmkt-n12000-cctrue-fs16-soar-cfsmkt-n16000-cctrue-fs16-soarmkt-n16000-cctrue-fs16-soar-cfsmkt-n20000-cctrue-fs16-soarmkt-n20000-cctrue-fs16-soar-cfsmkt-n24000-cctrue-fs16-soarmkt-n24000-cctrue-fs16-soar-cfs
0.90416452430487467544473533
0.95242250251264262288257290
0.997879808082888492

Latency at recall

mkt-n12000-cctrue-fs16-soarmkt-n12000-cctrue-fs16-soar-cfsmkt-n16000-cctrue-fs16-soarmkt-n16000-cctrue-fs16-soar-cfsmkt-n20000-cctrue-fs16-soarmkt-n20000-cctrue-fs16-soar-cfsmkt-n24000-cctrue-fs16-soarmkt-n24000-cctrue-fs16-soar-cfs
@ recall 0.90
p502.37 ms2.19 ms2.31 ms2.03 ms2.13 ms1.81 ms2.08 ms1.84 ms
p952.96 ms2.78 ms2.79 ms2.55 ms2.57 ms2.31 ms2.58 ms2.39 ms
p993.24 ms3.04 ms3.04 ms2.79 ms2.80 ms2.52 ms2.79 ms2.69 ms
@ recall 0.95
p504.10 ms3.97 ms3.96 ms3.75 ms3.81 ms3.43 ms3.84 ms3.42 ms
p955.00 ms5.04 ms4.72 ms4.56 ms4.48 ms4.28 ms4.70 ms4.32 ms
p995.56 ms5.53 ms5.13 ms5.00 ms4.88 ms4.80 ms5.32 ms4.80 ms
@ recall 0.99
p5012.74 ms12.53 ms12.41 ms12.47 ms12.37 ms11.34 ms11.83 ms10.84 ms
p9514.99 ms15.49 ms14.18 ms14.26 ms14.13 ms13.38 ms14.16 ms13.30 ms
p9916.10 ms16.59 ms14.93 ms15.14 ms15.17 ms14.65 ms15.74 ms14.21 ms

All values are taken at exactly the stated recall: each configuration's own sweep is log-interpolated between its two surrounding measured points (measured values are used verbatim on an exact hit) and the best configuration at that recall is shown — values are never interpolated across different configurations. “—” means no single sweep spans that recall; nothing is extrapolated. Recall is distance-based (ann-benchmarks, tie-aware). Insert/update/delete workloads are not measured.

Build comparison

mkt-n12000-cctrue-fs16-soarmkt-n12000-cctrue-fs16-soar-cfsmkt-n16000-cctrue-fs16-soarmkt-n16000-cctrue-fs16-soar-cfsmkt-n20000-cctrue-fs16-soarmkt-n20000-cctrue-fs16-soar-cfsmkt-n24000-cctrue-fs16-soarmkt-n24000-cctrue-fs16-soar-cfs
Build time614 s171 s215 s214 s295 s277 s314 s311 s
Index size2.2 GB2.2 GB2.3 GB2.3 GB2.3 GB2.3 GB2.3 GB2.3 GB

Plots

Pareto svg  ·  Distance-recall svg  ·  Per-config svg  ·  Build times svg  ·  Index sizes svg
paretodistance-recallbuild-timesbuild-sizes

Host

Instance typer8i.16xlarge
Hostnameip-172-31-73-29
CPUIntel(R) Xeon(R) 6975P-C (64 vCPU)
Memory496 GB
OSUbuntu 26.04 LTS
Kernel7.0.0-1006-aws

Versions

rekall0.1.0 (77145bf / main)
PostgreSQLPostgreSQL 18.4 (Ubuntu 18.4-1.pgdg26.04+1) on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu 15.2.0-16ubuntu1) 15.2.0, 64-bit
meerkat0.1.0 (8e85a2b)
pg_prewarm1.2
vector0.8.2
meerkat branchperf/fastscan-centroids

Search

k10
runs3
warmup500
num_queries10000
S3 archives3://rekall-bench/runs/cohere-10m/2026-06-12/r8i.16xlarge-152510/

Index builds

IndexTypeBuild timeSizeWorkers
mkt-n12000-cctrue-fs16-soarmeerkat613.7s2.2 GB32
mkt-n12000-cctrue-fs16-soar-cfsmeerkat171.0s2.2 GB32
mkt-n16000-cctrue-fs16-soarmeerkat215.2s2.3 GB32
mkt-n16000-cctrue-fs16-soar-cfsmeerkat213.8s2.3 GB32
mkt-n20000-cctrue-fs16-soarmeerkat295.0s2.3 GB32
mkt-n20000-cctrue-fs16-soar-cfsmeerkat276.5s2.3 GB32
mkt-n24000-cctrue-fs16-soarmeerkat314.2s2.3 GB32
mkt-n24000-cctrue-fs16-soar-cfsmeerkat311.3s2.3 GB32
Index parameters
mkt-n12000-cctrue-fs16-soar: {"boundary_epsilon":0.05,"centroid_compression":true,"centroid_fastscan":null,"fan_out":null,"fastscan":true,"fastscan_bits":null,"nlist":12000,"rerank_vectors":null,"soar_lambda":1.0}
mkt-n12000-cctrue-fs16-soar-cfs: {"boundary_epsilon":0.05,"centroid_compression":true,"centroid_fastscan":true,"fan_out":null,"fastscan":true,"fastscan_bits":null,"nlist":12000,"rerank_vectors":null,"soar_lambda":1.0}
mkt-n16000-cctrue-fs16-soar: {"boundary_epsilon":0.05,"centroid_compression":true,"centroid_fastscan":null,"fan_out":null,"fastscan":true,"fastscan_bits":null,"nlist":16000,"rerank_vectors":null,"soar_lambda":1.0}
mkt-n16000-cctrue-fs16-soar-cfs: {"boundary_epsilon":0.05,"centroid_compression":true,"centroid_fastscan":true,"fan_out":null,"fastscan":true,"fastscan_bits":null,"nlist":16000,"rerank_vectors":null,"soar_lambda":1.0}
mkt-n20000-cctrue-fs16-soar: {"boundary_epsilon":0.05,"centroid_compression":true,"centroid_fastscan":null,"fan_out":null,"fastscan":true,"fastscan_bits":null,"nlist":20000,"rerank_vectors":null,"soar_lambda":1.0}
mkt-n20000-cctrue-fs16-soar-cfs: {"boundary_epsilon":0.05,"centroid_compression":true,"centroid_fastscan":true,"fan_out":null,"fastscan":true,"fastscan_bits":null,"nlist":20000,"rerank_vectors":null,"soar_lambda":1.0}
mkt-n24000-cctrue-fs16-soar: {"boundary_epsilon":0.05,"centroid_compression":true,"centroid_fastscan":null,"fan_out":null,"fastscan":true,"fastscan_bits":null,"nlist":24000,"rerank_vectors":null,"soar_lambda":1.0}
mkt-n24000-cctrue-fs16-soar-cfs: {"boundary_epsilon":0.05,"centroid_compression":true,"centroid_fastscan":true,"fan_out":null,"fastscan":true,"fastscan_bits":null,"nlist":24000,"rerank_vectors":null,"soar_lambda":1.0}
PostgreSQL settings
effective_cache_size = 384 GB
effective_io_concurrency = 256
huge_pages = try
maintenance_work_mem = 204 GB
max_connections = 100
max_parallel_maintenance_workers = 32
max_parallel_workers = 64
random_page_cost = 1.1
shared_buffers = 128 GB
wal_level = replica
work_mem = 1 GB