Speed sweep

llama-3-2-3b-instruct-bf16-apple-m1-max-64gb-llama-cpp-tp1-sweep

llama-3-2-3b-instruct-bf16-apple-m1-max-64gb-llama-cpp-tp1-sweep

Record

Recipe
llama-3-2-3b-instruct-bf16-apple-m1-max-64gb-llama-cpp-tp1
Measured
2026-06-05T20:37:39.798Z
Points
1

Measured speed

ConcurrencyContextPrefillDecodeTTFT msStatus
16401,21950.9observed

Remaining fields

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

id
llama-3-2-3b-instruct-bf16-apple-m1-max-64gb-llama-cpp-tp1-sweep
measured at
2026-06-05T20:37:39.798Z
recipe id
llama-3-2-3b-instruct-bf16-apple-m1-max-64gb-llama-cpp-tp1
schema version
local-ai-registry/v1

metrics

concurrency
1
latest point at
2026-06-05T20:37:39.798Z
max context tokens
640
peak generation tps
50.92
peak prompt tps
1,218.95
point count
1

rows

concurrencycontext tokensdecode tok sdecode tok s per streamoutput tokenspeak vram gbprefill tok ssamples
164050.92Unknown128Unknown1,218.951
Provenance & metadata (1)