Speed sweep

pg-01a026ab-613d-7f75-abd8-6e6733422f0c-sweep

pg-01a026ab-613d-7f75-abd8-6e6733422f0c-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-qwen3-6-35b-a3b-iq3s-iq3s-apple-m4-max-128gb-llama-cpp-d31b1e0d7e
Measured: 2026-08-21T23:32:40.873Z
Points: 4

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	80.8	—	observed
1	32,832	—	70.2	—	observed
1	262,080	—	16.5	—	observed
1	32,768	—	—	42,372.7	observed

Remaining fields

Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.

id: pg-01a026ab-613d-7f75-abd8-6e6733422f0c-sweep
measured at: 2026-08-21T23:32:40.873Z
recipe id: pg-qwen3-6-35b-a3b-iq3s-iq3s-apple-m4-max-128gb-llama-cpp-d31b1e0d7e
schema version: local-ai-registry/v1

metrics

base memory bytes: 19165528064
base memory context tokens: 192
concurrency: 1
decode32k context tokens: 32,832
decode32k tps: 70.216
decode8k context tokens: 8,256
decode8k tps: 80.822
decode max context tokens: 262,080
decode max context tps: 16.471
decode mode: non-mtp
inference engine version: 1 (32e41fa)
latest point at: 2026-08-21T23:32:40.873Z
max context tokens: 262,080
max prompt tokens: 262,016
memory8k bytes: 19678085120
memory8k context tokens: 8,256
memory max context bytes: 22566748160
memory max context tokens: 262,080
peak generation tps: 70.216
peak memory bytes: 22566748160
peak prompt tps: 773.328
point count: 224
ttft32k cached prompt tokens: 28,668
ttft32k context tokens: 32,768
ttft32k seconds: 42.373

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	80.822	80.822	Unknown	18.327	Unknown	1
1	32,832	70.216	70.216	Unknown	21.017	Unknown	1
1	262,080	16.471	16.471	Unknown	21.017	Unknown	1
1	32,768	Unknown	Unknown	Unknown	Unknown	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:01a026ab-613d-7f75-abd8-6e6733422f0c
repository: local.ai ↗