Not true. This once again shows your utter ignorance.
Method calls don't start new threads you know...
That website seemed to imply that they do.
Something did. Maybe it was the system I/O calls.
We wrote a C+ program to do a simple sequence for an automated experiment:
1. Turn on the lights, so the subject knows to begin.
2. Measure exactly 4 seconds of subject activity at 1/1000 second intervals, using position and force sensors. Put it all in a 2d array.
3. Save the entire array to a disk file after the trial is over.
I wrote the code so that each of these things were done in a given order.
But that's not what we got.
- Our data were collected not once every 1/1000 second, but once every 1/18 second. So the experiment code ran for 3 minutes and 42 seconds if left to itself.
- Each time a "one millisecond" block of data was collected, the data were immediately saved to disk, without waiting for the entire array to be filled.
- After all of the disk I/O was finished, it turned on the lights to tell the subject to begin.
Somehow the compiler or operating system decided which I/O processes were "more important" from its own viewpoint, and changed the order of events to be "more efficient".
We never got to fully troubleshoot it. The administrator said we had wasted too …