Is anyone still follow this case now?
There’s a know issue in K5.10. Please refer to discussion in
Bad Jetpack5 nvme writing performance compared with Jetpack4
Thanks for sharing.
I look up in the answer, it explains the Buffering IO will affect the perfermance. But in my test case, I used FIO with configure “–direct=1” which means will not use buffer, but the performance still not ahcieve GEN4 * 4.