Speed sweep

pg-01a02c5f-edea-7eb2-8ade-3a6f71a380d4-sweep

pg-01a02c5f-edea-7eb2-8ade-3a6f71a380d4-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-qwen3-6-35b-a3b-iq3s-iq3s-apple-m4-max-36gb-llama-cpp-bf46efe043
Measured: 2026-08-23T02:07:59.451Z
Points: 4

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	67.8	—	observed
1	32,832	—	59.1	—	observed
1	262,080	—	13	—	observed
1	32,768	—	—	44,287.6	observed

Remaining fields

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

id: pg-01a02c5f-edea-7eb2-8ade-3a6f71a380d4-sweep
measured at: 2026-08-23T02:07:59.451Z
recipe id: pg-qwen3-6-35b-a3b-iq3s-iq3s-apple-m4-max-36gb-llama-cpp-bf46efe043
schema version: local-ai-registry/v1

metrics

base memory bytes: 19390971904
base memory context tokens: 192
concurrency: 1
decode32k context tokens: 32,832
decode32k tps: 59.143
decode8k context tokens: 8,256
decode8k tps: 67.842
decode max context tokens: 262,080
decode max context tps: 13.014
decode mode: non-mtp
inference engine version: unknown
latest point at: 2026-08-23T02:07:59.451Z
max context tokens: 262,080
max prompt tokens: 262,016
memory8k bytes: 19903168512
memory8k context tokens: 8,256
memory max context bytes: 22765993984
memory max context tokens: 262,080
peak generation tps: 59.143
peak memory bytes: 22765993984
peak prompt tps: 739.892
point count: 224
ttft32k cached prompt tokens: 28,668
ttft32k context tokens: 32,768
ttft32k seconds: 44.288

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	67.842	67.842	Unknown	18.536	Unknown	1
1	32,832	59.143	59.143	Unknown	21.202	Unknown	1
1	262,080	13.014	13.014	Unknown	21.202	Unknown	1
1	32,768	Unknown	Unknown	Unknown	Unknown	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:01a02c5f-edea-7eb2-8ade-3a6f71a380d4
repository: local.ai ↗