Assuming there is sufficient system memory (32g DDR4 in my case), should there be any adverse impact to processing time as the number of results retained is increased? I would assume not, but I am curious in the technical details here as I observed some unexpected behavior on one of my machines as I progressively increased that parameter (but the memory utilization was still well below the memory threshold utilized). The performance degradation happened on my Intel system which has an i7 1065G7 processor (8 logical cores) and 16g DDR4. The number of iterations retained was low (by relative measure) when this occurred, only 100.
For the heavier compute I'm utilizing my AMD system with a Ryzen 7 2700x processor (16 logical cores) and 32g DDR4, low-latency.
I just wanted to pose the question so that I could be aware of any architectural nuances that might exist.
Thx.
Comment