Speed sweep

pg-01a026ab-792b-759d-9dbc-c0f175599cf5-sweep

pg-01a026ab-792b-759d-9dbc-c0f175599cf5-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-qwen2-72b-instruct-q4km-unknown-apple-m5-pro-64gb-llama-cpp-a18f18770c
Measured: 2026-08-21T23:32:47.006Z
Points: 4

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	5.6	—	observed
1	32,704	—	4.5	—	observed
1	32,704	—	4.5	—	observed
1	32,640	—	—	724,218.1	observed

Remaining fields

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

id: pg-01a026ab-792b-759d-9dbc-c0f175599cf5-sweep
measured at: 2026-08-21T23:32:47.006Z
recipe id: pg-qwen2-72b-instruct-q4km-unknown-apple-m5-pro-64gb-llama-cpp-a18f18770c
schema version: local-ai-registry/v1

metrics

base memory bytes: 57669386240
base memory context tokens: 192
concurrency: 1
decode32k context tokens: 32,704
decode32k tps: 4.525
decode8k context tokens: 8,256
decode8k tps: 5.615
decode max context tokens: 32,704
decode max context tps: 4.525
decode mode: non-mtp
inference engine version: unknown
latest point at: 2026-08-21T23:32:47.006Z
max context tokens: 32,704
max prompt tokens: 32,640
memory8k bytes: 57765445632
memory8k context tokens: 8,256
memory max context bytes: 57868566528
memory max context tokens: 32,704
peak generation tps: 4.836
peak memory bytes: 57868566528
peak prompt tps: 55.812
point count: 28
ttft32k cached prompt tokens: 28,677
ttft32k context tokens: 32,640
ttft32k seconds: 724.218

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	5.615	5.615	Unknown	53.798	Unknown	1
1	32,704	4.525	4.525	Unknown	53.894	Unknown	1
1	32,704	4.525	4.525	Unknown	53.894	Unknown	1
1	32,640	Unknown	Unknown	Unknown	Unknown	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:01a026ab-792b-759d-9dbc-c0f175599cf5
repository: local.ai ↗