lightseq icon indicating copy to clipboard operation
lightseq copied to clipboard

Published 20 hours ago • verified publisher icon bytedance

Is it normal that A10 inference speed is lower than 2080ti?

Open qinbo23 opened this issue 1 year ago • 1 comments

hello？I tested the Transformer-base inference speed on different devices. It's weird that A10 speed is lower than 2080ti speed.

MODEL: Transformer-base DATA: fp16 SPEED: (number of src characters / second) 3090 7.5k/s 2080 4.5k/s A10 2.0K/s

qinbo23 avatar Nov 07 '23 06:11 qinbo23

me too. can anyone help?

Mr-wang2016 avatar Dec 12 '23 07:12 Mr-wang2016