Benchmarking VLM on Orin

[AastaLLL]

Hi,

You can try the one from GitHub.
The vLLM version is different and the GitHub one should be newer.

Throttling is expected when running GPU-heavy loading under performance mode.
The warning can be turned off. Please check below comment:

Offline cv0, cv1 and cv2

Hi, We test a custom nvpmodel with 2x CPU@1728, GPU@1020 and EMC@3199. Under the heavy load GPU tasks (ex. LLM inference), we can still observe the OC throttling behavior. But similar to the previous testing, the impact on the performance is limited as the clocks don’t decrease much. (1019->1003) 01-22-2025 06:53:47 RAM 7119/7620MB (lfb 4x4MB) SWAP 580/16384MB (cached 155MB) CPU [0%@1728,2%@1728,off,off,off,off] EMC_FREQ 43%@3199 GR3D_FREQ 99%@[1008] NVDEC off NVJPG off NVJPG1 off VIC off OF…

Thanks.