L8 Upgrade

We are getting an upgrade package soon for one of our machines that have L6 in it now and we will get new HMI's

I am trying to learn some about L8 before we get at as we have messages and produce and consume from the L6 to other machines and controllers on the production line.

From what I am reading the L8 is 64 bit and the L6 is 32 bit and the older SLC and MicroLogix is 16 bit?

Does that mean my DINT's in the L8 will be 64 bits long instead of 32 bit long like they are now?

Also, what is the meaning of Atomic Data Type? I keep seeing that term but don't know what it means?

You should have no issues with produce/consume between older processors. DINTs are still 32 bits. But I think it does support 64 bit words too. Atomic data types are bool, integer, DINTs. They are the most basic data types from which other data types can be created, i.e. UDTs.

There are enough differences between the 1756-L8z hardware and operating system and the older controllers that there is a detailed "Replacement Guide" publication from Rockwell Automation:

https://literature.rockwellautomation.com/idc/groups/literature/documents/rm/1756-rm100_-en-p.pdf

While there's a lot of data there, from what I've read on this Forum and heard in the field, the usual issues have to do with very old motion systems (SERCOS and analog) that are migrated to the new controllers, and with MSG instructions and various buffer handling techniques and CIP paths.

The CIP paths have changed and that (with my experience, as little as it is) sometimes causes problems and sometimes doesn't. The other big change for me is the way the discrete I/O is structured (not sure about analog). With the L8's, each input/output channel has its own word rather than just one bit in a word. For me the biggest impact of that is I can't write/read a single word and get all 16 inputs or outputs.
Having said all of that, going into it knowing that there are differences is 90% of the battle. If something doesn't work on startup you'll know that there is a possibility that it has to do with the difference between processors rather than taking the approach "it can't be the program, it's the same PLC's" (which you shouldn't take anyway but I can't tell you the number of times that would be a response I'd get during a tech call).