Speed sweep

pg-01a0139f-f174-7448-9212-f90dc7d011a1-sweep

pg-01a0139f-f174-7448-9212-f90dc7d011a1-sweep

JSON API Back to speed sweep search

Record

Recipe: pg-unsloth-gemma-4-e2b-it-gguf-ud-q4-k-xl-apple-m3-ultra-96gb-60c-llama-cpp-f37f0c389a
Measured: 2026-08-18T06:47:24.268Z
Points: 4

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	8,256	—	110.7	—	observed
1	32,832	—	98.3	—	observed
1	131,008	—	67.7	—	observed
1	32,768	—	—	19,081.5	observed

Remaining fields

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

id: pg-01a0139f-f174-7448-9212-f90dc7d011a1-sweep
measured at: 2026-08-18T06:47:24.268Z
recipe id: pg-unsloth-gemma-4-e2b-it-gguf-ud-q4-k-xl-apple-m3-ultra-96gb-60c-llama-cpp-f37f0c389a
schema version: local-ai-registry/v1

metrics

base memory bytes: 5321734272
base memory context tokens: 192
concurrency: 1
decode32k context tokens: 32,832
decode32k tps: 98.257
decode8k context tokens: 8,256
decode8k tps: 110.746
decode max context tokens: 131,008
decode max context tps: 67.673
decode mode: non-mtp
inference engine version: unknown
latest point at: 2026-08-18T06:47:24.268Z
max context tokens: 131,008
max prompt tokens: 130,944
memory8k bytes: 5899761792
memory8k context tokens: 8,256
memory max context bytes: 14698625152
memory max context tokens: 131,008
peak generation tps: 91.29
peak memory bytes: 14698625152
peak prompt tps: 1,312.264
point count: 112
ttft32k cached prompt tokens: 28,676
ttft32k context tokens: 32,768
ttft32k seconds: 19.082

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	8,256	110.746	110.746	Unknown	5.495	Unknown	1
1	32,832	98.257	98.257	Unknown	13.689	Unknown	1
1	131,008	67.673	67.673	Unknown	13.689	Unknown	1
1	32,768	Unknown	Unknown	Unknown	Unknown	Unknown	1

Provenance & metadata (1)

source

paths: publication:pg-20260827T060320709Z, run:01a0139f-f174-7448-9212-f90dc7d011a1
repository: local.ai ↗