Cohere-50M — 2026-06-23

r8i.16xlarge · peak recall 0.9813 @ 79 QPS

QPS at recall

Recall	Meerkat	pgvector HNSW	Δ vs best competitor
0.90	— (min r=0.901)	382	—
0.95	208	195	+7%

Latency at recall

	Meerkat	pgvector HNSW	Δ vs best competitor
@ recall 0.90
p50	—	2.63 ms	—
p95	—	3.50 ms	—
p99	—	4.13 ms	—
@ recall 0.95
p50	4.79 ms	5.29 ms	-9%
p95	5.72 ms	6.56 ms	-13%
p99	6.23 ms	7.15 ms	-13%

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.

Interpolated on config — pgvector HNSW @0.90: hnsw-m24 · Meerkat @0.95: mkt-n480000-cctrue-fs16-soar-cfs · pgvector HNSW @0.95: hnsw-m24

Plots

Pareto svg · Distance-recall svg · Per-config svg

Host

Instance type	r8i.16xlarge
Hostname	ip-172-31-79-47
CPU	Intel(R) Xeon(R) 6975P-C (64 vCPU)
Memory	496 GB
OS	Ubuntu 26.04 LTS
Kernel	7.0.0-1006-aws

Versions

rekall	0.1.0 (44cdebe / main)
PostgreSQL	PostgreSQL 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
meerkat	0.1.0 (8695684)
pg_prewarm	1.2
vector	0.8.2

Search

k	10
runs	3
warmup	500
num_queries	3000
S3 archive	s3://rekall-bench/runs/cohere-50m/2026-06-23/r8i.16xlarge-132642/

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