Speed sweep

nvidia-nemotron-3-5-lightning-30b-a3b-bf16-q4-0-intel-arc-pro-b70-32gb-llama-cpp-tp1-sweep

nvidia-nemotron-3-5-lightning-30b-a3b-bf16-q4-0-intel-arc-pro-b70-32gb-llama-cpp-tp1-sweep

Record

Recipe: nvidia-nemotron-3-5-lightning-30b-a3b-bf16-q4-0-intel-arc-pro-b70-32gb-llama-cpp-tp1
Measured: 2026-08-12T06:53:16.931Z
Points: 1

Measured speed

Concurrency	Context	Prefill	Decode	TTFT ms	Status
1	16,384	766	87.4	—	observed

Remaining fields

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

id: nvidia-nemotron-3-5-lightning-30b-a3b-bf16-q4-0-intel-arc-pro-b70-32gb-llama-cpp-tp1-sweep
measured at: 2026-08-12T06:53:16.931Z
recipe id: nvidia-nemotron-3-5-lightning-30b-a3b-bf16-q4-0-intel-arc-pro-b70-32gb-llama-cpp-tp1
schema version: local-ai-registry/v1

metrics

concurrency: 1
inference engine version: b306 SYCL (f785fc9ea) + custom Mamba2 kernel fusion + MMQ/XMX dispatch
latest point at: 2026-08-12T06:53:16.931Z
max context tokens: 16,384
peak generation tps: 87.4
peak prompt tps: 766
point count: 1

rows

concurrency	context tokens	decode tok s	decode tok s per stream	output tokens	peak vram gb	prefill tok s	samples
1	16,384	87.4	87.4	0	Unknown	766	1

Provenance & metadata (1)

source