Did some FME Performance testing over the holiday break. First up: FME 2022.0 vs FME 2021.1. 6 million points overlaid on areas. All on 2020 Intel Macbook Pro. Summary -- 10% faster with no workspace changes, 20% if we flag "areas first" so the CSV multithreading really kicks in. 
Next up: Actual CPU usage. Slight increase in FME 2022 due to CSV reader multithreading
Last up: Let's see how running Intel compiled FMEs on M1 Max holds up. Elapsed time for emulated Intel only 16% more than on native Intel processor.
And what about Windows? Using Windows 11 ARM inside a Parallels VM on M1 Max running Windows Intel compiled FME takes roughly double the elapsed time of running Intel FME on MacOS M1 Max.
But hey, Intel Windows FME works just fine on Windows 11 ARM running in a VM on an M1 Max Mac. Just a bit slower.
Didn't foresee or imagine that Intel targeted Windows programs would work so well in this environment.
(Which means I can keep using my 1993 Quicken!)
🤯🤯🤯
Didn't foresee or imagine that Intel targeted Windows programs would work so well in this environment.
(Which means I can keep using my 1993 Quicken!)
🤯🤯🤯
• • •
Missing some Tweet in this thread? You can try to
force a refresh
