Speed sweep

pg-01a03076-e8e4-7199-883b-d7a11a9c860f-sweep

pg-01a03076-e8e4-7199-883b-d7a11a9c860f-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-laguna-xs-2-1-q4km-unknown-apple-m4-max-36gb-llama-cpp-0698f0a9be
Measured: 2026-08-23T21:11:34.362Z
Points: 4

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	97	—	observed
1	32,832	—	73.7	—	observed
1	32,960	—	73.5	—	observed
1	32,768	—	—	57,590.3	observed

Remaining fields

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

id: pg-01a03076-e8e4-7199-883b-d7a11a9c860f-sweep
measured at: 2026-08-23T21:11:34.362Z
recipe id: pg-laguna-xs-2-1-q4km-unknown-apple-m4-max-36gb-llama-cpp-0698f0a9be
schema version: local-ai-registry/v1

metrics

base memory bytes: 21938323456
base memory context tokens: 192
concurrency: 1
decode32k context tokens: 32,832
decode32k tps: 73.703
decode8k context tokens: 8,256
decode8k tps: 96.976
decode max context tokens: 32,960
decode max context tps: 73.466
decode mode: non-mtp
inference engine version: 10017 (7dc1bfead)
latest point at: 2026-08-23T21:11:34.362Z
max context tokens: 32,960
max prompt tokens: 32,896
memory8k bytes: 22185082880
memory8k context tokens: 8,256
memory max context bytes: 22426320896
memory max context tokens: 32,960
peak generation tps: 73.703
peak memory bytes: 22426320896
peak prompt tps: 568.985
point count: 29
ttft32k cached prompt tokens: 28,673
ttft32k context tokens: 32,768
ttft32k seconds: 57.59

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	96.976	96.976	Unknown	20.661	Unknown	1
1	32,832	73.703	73.703	Unknown	20.886	Unknown	1
1	32,960	73.466	73.466	Unknown	20.886	Unknown	1
1	32,768	Unknown	Unknown	Unknown	Unknown	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:01a03076-e8e4-7199-883b-d7a11a9c860f
repository: local.ai ↗