I have a DM aggregation block running on 14 million records x 38 fields. On 1 field doing unique values (has 5 million unique values) I am running various sums, 6 first non-empty values, 1 unique value count, 1 min & 1 max
It has been running for 2 days now yet is only using 6GB of my 64GB RAM (windows 10 pro, Xeon 8 core processor, Omniscope version 2.9 build 1765 (rc) [plus] x64)
Is there anything i can do to make aggregate use the available RAM more intelligently - i have other files use far more of the RAM, but always have this issue with aggregation.
we are also very interested in generally increasing performance both in batch and in interactive data processing. Can I also invite you to have a look at my last post about increasing available memory?