← all runs
Cohere-100M — 2026-07-07
r8i.4xlarge · peak recall 0.9878 @ 40 QPS
“smallest-practical probe: 100M on 128GB/8 workers”
QPS at recall
| Recall | mkt-n240000-cctrue-fs16-soar-cfs | mkt-n400000-cctrue-fs16-soar-cfs |
|---|
| 0.90 | 188 | — (min r=0.900) |
| 0.95 | 86 | 134 |
Latency at recall
| mkt-n240000-cctrue-fs16-soar-cfs | mkt-n400000-cctrue-fs16-soar-cfs |
|---|
| @ recall 0.90 | | |
| p50 | 5.22 ms | — |
| p95 | 6.87 ms | — |
| p99 | 7.67 ms | — |
| @ recall 0.95 | | |
| p50 | 11.47 ms | 7.36 ms |
| p95 | 14.74 ms | 9.62 ms |
| p99 | 16.11 ms | 10.53 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-n240000-cctrue-fs16-soar-cfs | mkt-n400000-cctrue-fs16-soar-cfs |
|---|
| Build time | 79695 s | 112173 s |
| Index size | 13.1 GB | 21.3 GB |
Plots
Pareto svg ·
Distance-recall svg ·
Per-config svg ·
Build times svg ·
Index sizes svgHost
| Instance type | r8i.4xlarge |
| Hostname | ip-172-31-67-239 |
| CPU | Intel(R) Xeon(R) 6975P-C (16 vCPU) |
| Memory | 124 GB |
| OS | Ubuntu 26.04 LTS |
| Kernel | 7.0.0-1008-aws |
Versions
| rekall | 0.1.0 (e2b6c7f / 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 (78c22ca) |
| pg_prewarm | 1.2 |
| vector | 0.8.4 |
Search
Index builds
| Index | Type | Build time | Size | Workers |
|---|
| mkt-n240000-cctrue-fs16-soar-cfs | meerkat | 79694.7s | 13.1 GB | 8 |
| mkt-n400000-cctrue-fs16-soar-cfs | meerkat | 112172.8s | 21.3 GB | 8 |
Index parameters
mkt-n240000-cctrue-fs16-soar-cfs: {"boundary_epsilon":0.05,"centroid_compression":true,"centroid_fastscan":true,"fan_out":null,"fastscan":true,"fastscan_bits":null,"nlist":240000,"rerank_vectors":null,"soar_lambda":1.0}
mkt-n400000-cctrue-fs16-soar-cfs: {"boundary_epsilon":0.05,"centroid_compression":true,"centroid_fastscan":true,"fan_out":null,"fastscan":true,"fastscan_bits":null,"nlist":400000,"rerank_vectors":null,"soar_lambda":1.0}
PostgreSQL settings
effective_cache_size = 96 GB
effective_io_concurrency = 256
huge_pages = try
maintenance_work_mem = 51 GB
max_connections = 100
max_parallel_maintenance_workers = 8
max_parallel_workers = 16
random_page_cost = 1.1
shared_buffers = 32 GB
wal_level = replica
work_mem = 1 GB
