determinism in the host computer

[IñigoP]

Hi there,

I want to determine if my programe on the host is deterministic. The thing is that when I use the benchmarking technique and execute the code, the time is variable. How can I fix this?

Thanks in advance.

Cheers, Iñigo.

[billko]

---

@IñigoP wrote:

Hi there,

I want to determine if my programe on the host is deterministic. The thing is that when I use the benchmarking technique and execute the code, the time is variable. How can I fix this?

Thanks in advance.

Cheers, Iñigo.

---

I suspect that the variance should already tell you something about determinism on your computer?

[Bob_Schor]

Well, if you posted the code (meaning something we could examine and, optionally, execute), we might be able to help you answer that question.  What do you mean by "deterministic"?  I presume you are running it on a Windows PC, which "allows" code to run according to its scheduling algorithm (oops, time for a Virus scan, oops, check if Adobe software needs to be updated, etc.).  Your code might also have "non-determinism" built into it -- does it do I/O, for example?

Bob Schor

[IñigoP]

Hello,

Is it possible to achieve determinism on a windows PC? I have a programme, but when I'm trying to benchmark it, I obtain an overflow on the counter (the maximum value for an U32). What can it be?

Thank you.

Cheers, Iñigo.

[LennartM]

Determinism is also relative. Loop cycle times of a few seconds with a simple loop can be sort of deterministic in Windows.

Best way to measure the determinism is to use another device to time it. If you have a daq connected you can for instance trigger a digital output and measure it with an oscilloscope.

[Artem.SPb]

First of all better to show your code/example of your problem.
Now It is very difficult (impossible?) to say, why your counter is overflow

[IñigoP]

Here you have the program. You can see that the 2 ways I use to benchmark show different times, maybe I'm doing it wrong... I don't know...

[Artem.SPb · Accepted solution]

yes, both wrong. In both cases, you subtract a larger value from the smaller

see correct

[IñigoP]

ooh..... I see it know!! Thank you so much!