Speed sweep

pg-01a011ee-c397-74aa-8a5a-59418f9e5d6a-sweep

pg-01a011ee-c397-74aa-8a5a-59418f9e5d6a-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-antirez-deepseek-v4-gguf-deepseek-v4-flash-iq2xxs-w2q2k-apr-8b20120753de-apple-m3-ultra-96gb-80c-llama-cpp-962f62b7a7
Measured: 2026-08-17T22:54:15.443Z
Points: 3

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	8.1	—	observed
1	32,832	—	7.6	186,934.2	observed
1	258,112	—	4.9	—	observed

Remaining fields

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

id: pg-01a011ee-c397-74aa-8a5a-59418f9e5d6a-sweep
measured at: 2026-08-17T22:54:15.443Z
recipe id: pg-antirez-deepseek-v4-gguf-deepseek-v4-flash-iq2xxs-w2q2k-apr-8b20120753de-apple-m3-ultra-96gb-80c-llama-cpp-962f62b7a7
schema version: local-ai-registry/v1

metrics

base memory bytes: 89078857728
base memory context tokens: 0
concurrency: 1
decode32k context tokens: 32,832
decode32k tps: 7.643
decode8k context tokens: 8,256
decode8k tps: 8.148
decode max context tokens: 258,112
decode max context tps: 4.93
decode mode: non-mtp
inference engine version: 10017 (7dc1bfead)
latest point at: 2026-08-17T22:54:15.443Z
max context tokens: 258,112
max prompt tokens: 258,048
memory8k bytes: 87569186816
memory8k context tokens: 8,256
memory max context bytes: 89065979904
memory max context tokens: 258,112
peak generation tps: 7.643
peak memory bytes: 89078857728
peak prompt tps: 175.292
point count: 224
ttft32k cached prompt tokens: 28,735
ttft32k context tokens: 32,768
ttft32k seconds: 186.934

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	8.148	Unknown	Unknown	81.555	Unknown	1
1	32,832	7.643	Unknown	Unknown	82.961	Unknown	1
1	258,112	4.93	Unknown	Unknown	82.949	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:01a011ee-c397-74aa-8a5a-59418f9e5d6a
repository: local.ai ↗