EMMC vs. NVMe as Storage Device

Hello NVIDIA Forum,
When flashing JetPack using SDKManager the option Storage Device can be selected.
Given that I have inserted a NVMe SSD into carrier board.
I have tried both, they both work.

What is best practice?
By best practice I mean factors like
Lifetime, R/W speed, resilience to bad power and abrupt power failures, industry standards in remote environments.
Hope you can provide some feedback pro/cons on this option

Supposed the storage size and cost is one major consideration of your product.
If the eMMC is enough for your application and operation, then why need to consider to add an extra NVMe?

Thanks for the reply.
I’m in a situation where I’d like to add an NVMe SSD anyway because I’d like to store files/data.
But the internal eMMC is large enough to host my root.
So, I’d like the opinion from this forum on which one of these configurations I should use:

  1. Flash to eMMC, have root and install everything (apt install) to the eMMC storage and mount my SSD in a folder and store raw data there.
  2. Flash to NVMe SSD and use it for everything. Not use the eMMC storage at all.

Again, my main reason for posting this is to get feedback on what is best practice, lifetime, R/W speed, resilience to bad power and abrupt power failures, industry standards in remote environments.

1 Like

Anyone?

If I’m going to mount an NVMe SSD anyways, should I flash root to it or just keep it as external storage?
What I am afraid of is if the SSD becomes corrupt or I get mounting problems then the computer is dead.
While mounting problems may not occur with the eMMC?

I’m not sure which solution is most robust.

Is this still an issue to support? Any result can be shared? Thanks

Yes, this is still an issue. It’s not easy to resolve solely by me since I’m asking about user experience and best practices. Hence eMMC vs NVMe must be tested multiple times for weeks to gather statistics.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.