Speed sweep

pg-01a029ca-0e56-70f6-9bb3-abbfbf7b351c-sweep

pg-01a029ca-0e56-70f6-9bb3-abbfbf7b351c-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-qwen3-6-35b-a3b-q3km-unknown-apple-m4-pro-48gb-llama-cpp-3fcce8e85b
Measured: 2026-08-22T14:05:02.922Z
Points: 4

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	48.7	—	observed
1	32,832	—	40.3	—	observed
1	262,080	—	7.3	—	observed
1	32,768	—	—	81,581.5	observed

Remaining fields

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

id: pg-01a029ca-0e56-70f6-9bb3-abbfbf7b351c-sweep
measured at: 2026-08-22T14:05:02.922Z
recipe id: pg-qwen3-6-35b-a3b-q3km-unknown-apple-m4-pro-48gb-llama-cpp-3fcce8e85b
schema version: local-ai-registry/v1

metrics

base memory bytes: 22315843584
base memory context tokens: 192
concurrency: 1
decode32k context tokens: 32,832
decode32k tps: 40.31
decode8k context tokens: 8,256
decode8k tps: 48.736
decode max context tokens: 262,080
decode max context tps: 7.336
decode mode: non-mtp
inference engine version: unknown
latest point at: 2026-08-22T14:05:02.922Z
max context tokens: 262,080
max prompt tokens: 262,016
memory8k bytes: 22827466752
memory8k context tokens: 8,256
memory max context bytes: 25691258880
memory max context tokens: 262,080
peak generation tps: 40.31
peak memory bytes: 25691258880
peak prompt tps: 401.66
point count: 224
ttft32k cached prompt tokens: 28,668
ttft32k context tokens: 32,768
ttft32k seconds: 81.582

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	48.736	48.736	Unknown	21.26	Unknown	1
1	32,832	40.31	40.31	Unknown	23.927	Unknown	1
1	262,080	7.336	7.336	Unknown	23.927	Unknown	1
1	32,768	Unknown	Unknown	Unknown	Unknown	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:01a029ca-0e56-70f6-9bb3-abbfbf7b351c
repository: local.ai ↗