It's eventually will be a production
We were purging with DBCC cachedataremove and bind /unbind and also purging statement cache to have the same for both tables but always got the same result.
cache is to test a single 4G table
8 Kb 491520 Kb 69132.00 Mb 69132.00 Mb 10
16 Kb 102400 Kb 500.00 Mb 500.00 Mb 10
64 Kb 491520 Kb 32768.00 Mb 32768.00 Mb 10
Also decreased to 1 partition. but the sysmon output keep the same apl/dol ratio.
The only thing it wasn't me running these tests, but i'm pretty sure the did everything as instructed.
Another "strange" thing (at least for me) that with these cache sizes we have "LRU Buffers gubbed" (as in all cases rather similar for apl and dol) but this I think still not explain this ap/dol difference,.