Speed sweep

pg-019f3295-e202-7201-bc3c-afc9adf736f4-sweep

pg-019f3295-e202-7201-bc3c-afc9adf736f4-sweep

Record

Recipe
pg-mlx-community-nvidia-nemotron-3-nano-30b-a3b-nvfp4-nvfp4-apple-m3-ultra-96gb-80c-mlx-9628aba580
Measured
2026-07-05T14:01:51.358Z
Points
4

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatus
18,256103.4observed
132,83294.5observed
1258,11154.5observed
132,76817,529.1observed

Remaining fields

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

id
pg-019f3295-e202-7201-bc3c-afc9adf736f4-sweep
measured at
2026-07-05T14:01:51.358Z
recipe id
pg-mlx-community-nvidia-nemotron-3-nano-30b-a3b-nvfp4-nvfp4-apple-m3-ultra-96gb-80c-mlx-9628aba580
schema version
local-ai-registry/v1

metrics

base memory bytes
19922041604
base memory context tokens
192
concurrency
1
decode32k context tokens
32,832
decode32k tps
94.491
decode8k context tokens
8,256
decode8k tps
103.388
decode max context tokens
258,111
decode max context tps
54.519
decode mode
non-mtp
inference engine version
unknown
latest point at
2026-07-05T14:01:51.358Z
max context tokens
258,111
max prompt tokens
258,047
memory8k bytes
22882278522
memory8k context tokens
8,256
memory max context bytes
26110432390
memory max context tokens
258,111
peak generation tps
94.491
peak memory bytes
26110432390
peak prompt tps
1,869.346
point count
223
ttft32k cached prompt tokens
28,736
ttft32k context tokens
32,768
ttft32k seconds
17.529

rows

concurrencycontext tokensdecode tok sdecode tok s per streamoutput tokenspeak vram gbprefill tok ssamples
18,256103.388103.388Unknown21.311Unknown1
132,83294.49194.491Unknown24.317Unknown1
1258,11154.51954.519Unknown24.317Unknown1
132,768UnknownUnknownUnknownUnknownUnknown1
Provenance & metadata (1)

source

paths
publication:pg-20260827T060320709Z, run:019f3295-e202-7201-bc3c-afc9adf736f4
repository
local.ai