I’m building a desktop app. It has high CPU charge as it’s basically formed by heavy math calculations (iterations). I’m using multithread (trough shells) to ensure full CPU usage. I’m compiling the app for MacOS (64 bits) and for Windows (64 bits).
I’m running the app in the same hardware (a MacBook Pro with 6 Intel Core i9 & 32 GB RAM).
Starting the computer under MacOS (10.15.6), processing a file takes 5 minutes 22 seconds (I’m writing the calculation time on the output file). The process uses the CPU at 100%.
Starting the same computer under Windows (BootCamp, so not a virtual machine, really starting on Windows) (version 2004 OS Build 19041.450), processing exactly the same file takes 9 minutes 40 seconds. Again if I monitor the resources, the process uses the CPU at 100%.
Of course the code makes exactly the same under MacOS and under Windows.
Of course I’m testing this in a fresh & clean Windows & MacOS installation. I’m also taking care no background process to be running in both cases.
I’ve made tests with other programs (Adobe PhotoShop) and making the same operations takes more or less the same time under both “platforms” (differences of seconds between the time under MacOS & Windows).
The only explanation I’ve for this big difference is the optimisation of the program build from Xojo for MacOS or for PC.
I’m using Xojo 2020r1 with no additional plugins for this project.
I know this is not a bug or a problem, as in both platforms the software is running fine and correctly, but I’d like to know if you’re experiencing this high differences in calculation times if you run the code under MacOS & Windows.
I am not sure Xojo is to blame. The Windows platform is inherently slower. Heck, even the UI is slow in Windows. If you have more than a dozen icons on the desktop, it take a minute at least to refresh.
From what I gather, Xojo uses the very same compiler to produce 64 bit executable. Yet, the Windows executable is visibly slower.
The IDE itself is noticeably slower on PC. What takes perhaps less than a minute to compile on Mac will often take two minutes on Windows.
I avoid working on PC, yet I always me sure to test the program on a real PC, let alone to tune up the UI to avoid flicker.
My work machine is an iMac, not even the fastest. A late 2015 27". It is way snappier than the PC, which is the fastest I could find.
My IDE is empty. I’m just loading a file with data, making calculations over the values on the file and finally saving the results into a second file, so 100% working internally with math operations.
I referred to 2016R3 I am forced to use, so I get the same transparency as under Mac. I simply cannot refactor entirely my UI in existing apps, to accommodate pseudo transparency. If even possible at all.
It’s good to know the current IDE is faster, though. I may use it for projects that do not require elaborate UI.
It’s take 4 hours to compile… and the differences has not been significant… Thanks for the tip. I’ll compile in that way the final release, but of course, not the ones while developing…
Generations have changed.