As usual, I couldn't help but poke my nose in and run a few tests for curiousity's sake. Though LCT didn't really have a direct option to import old logs to make the testing easier. My only recourse was to inject EQ2 logs into an already open log file and hope for the best.
Of course the best wasn't great, but I'll share anyhow. Just as a preface, LCT saw the 32MB of injected logs as one encounter, which I'm sure caused problems I cannot predict. Additionally it would draw a graph of the giant encounter every second as it was parsing, undoubtably slowing the parsing down. I couldn't fool ACT into doing that as the injected logs were in the past and not "new". So for the tests, ACT had a CPU advantage of an unknown amount.
I ran a few tests for ACT using different perfomance settings, and some settings to simulate one giant encounter like LCT was doing. So labels like (1) mean one giant encounter. A label with "
-" means ACT had memory saving options enabled.
Code:
Program/Test LCT(1) ACT(73) ACT(1) ACT(73)- ACT(1)-
Startup CPU 0.937 1.453 " " "
Startup RAM 25.6M 35.2M " " "
Startup Private 32.0M 29.2M " " "
Read 32MB CPU 1:25.953 1:09.656 1:01.375 49.703 51.828
Read 32MB RAM 95.6M 218M 218M 106M 102M
Read 32MB Private 96.5M 214M 214M 100M 95M