[Title Corrected] DEVICEID is NOT unique - it is "random" !

[awneil]

64 bit unique device identifier.
DEVICEID[0] contains the least significant bits of the device identifier. DEVICEID[1] contains the most significant bits of the device identifier.

infocenter.nordicsemi.com/.../ficr.html

Client wants a unique id - but would also prefer that products in a batch do not have sequential IDs.

Is it likely that chips on a reel will have sequential DEVICEIDs ?

[Jørgen Holmefjord]

Hi,

Each number is randomly generated, see this post. There is an extremely low probability that the chips on a reel will be sequential.

Best regards,
Jørgen

[awneil]

So the documentation is incorrect - it is not actually unique at all?!

That needs to be clearly stated in the documentation!!

There have been many threads here which rely upon the DEVICEID being unique!

eg,

https://devzone.nordicsemi.com/f/nordic-q-a/48642/generation-of-a-serialized-id-code-at-load-time/193030#193030

Even Nordic Staff have got this wrong:

https://devzone.nordicsemi.com/f/nordic-q-a/22089/specific-unique-id-for-device-detection/198059#198059

[awneil]

awneil said:

the documentation is incorrect

This was noted three years ago:

https://devzone.nordicsemi.com/f/nordic-q-a/17569/is-nrf51-deviceid-unique/198062#198062

and it's still wrong!

[Jørgen Holmefjord]

No, we can't guarantee that it is unique. But with 18446744073709551616 possible combinations, you should be quite unlucky to get numbers that are not unique.

We have never gotten any complaints or seen issues with this solution, and we use the number to uniquely identify chips ourselves.

[awneil]

But the fact remains that the documentation is incorrect.

Please update it to clearly state that it is random, not unique.

You could include the information about how it is generated, and the (un)likelihood of collisions - but don't lie!

It would save you having to keep answering the same question on the forum!