Sub-routine

how to used subroutine in cqm1?, because i want to make my program short so that plc will lessen the time of scanning my program. THANKS!!!

Subroutine itself don't do program more fast.
Subroutine call itself need stackin etc. for the processor so program will came more slowly, if there still have same functions.

What kind of program-developer you have ?

OMRON JUMPS

In OMRON, its best to use JMP & JME this will allow you to jump to certain parts of a program or skip others alltogether.

Although if its a CQM1H you have, I have written some huge programs and never needed to use JUMPS, if you use 'DIFU' & 'DIFD' as much as possible, you will have a program which will run quickly anyway.

Chris

Huh?...

CMF said:

I have written some huge programs and never needed to use JUMPS, if you use 'DIFU' & 'DIFD' as much as possible, you will have a program which will run quickly anyway.

Click to expand...

Please explain how using DIFU and DIFD "as much as possible" has ANY effect on how 'fast' the program will run...

beerchug

-Eric

Re: Huh?...

Eric Nelson said:

Please explain how using DIFU and DIFD "as much as possible" has ANY effect on how 'fast' the program will run...

Click to expand...

They are one shot evaluations as OSR (DIFU) and OSL (DIFD) what I never need.

Yeah Seppo, I know what they are, I use 'em all the time...

I just want to know how using them 'as much as possible' will magically decrease scan time...

Chris, using them where not needed will actually INCREASE scan time. Check the execution time of a DIFU/D vs. a regular OUT, the DIFU/D is considerably longer. They also require more memory!...

Or, did I misunderstand something?...

beerchug

-Eric

The question was how can subroutines make programs run faster.

rblopena, you are right, you can save CPU time by using subroutines as long as they are not called contuinously. When subroutines are not called then subroutines use no CPU time. The main scan of a well organinzed program may consist mainly of conditional calls to subroutines.

I routinely put the inintialization code in a subroutine that gets called only when the first scan is active. The rest of the time this initialization subroutine has no affect on the program or the scan time. This is much better than putting the initialization code in the main loop where is get scanned and consumes CPU time.

Seppo is correct that there is extra overhead in calling a subroutine. The program will execute more slowly if the subroutine is called every scan. I would be faster to put the code in-line in theis case. However, it may make the main program to long a look disorgranized.

CMF, is correct that jumps over sections of code will save time, but I like the concept of conditionally call subroutines better.

Eric, the DIFU/D is just a way to keep from continously call a subroutine. This is one way to make sure a subroutine gets called only when it required. Another trick I have used to to have the subroutine clear the bit that activated it. Seppo must use this trick if he doesn't use one shots.

Re: The question was how can subroutines make programs run faster.

Peter Nachtwey said:

Eric, the DIFU/D is just a way to keep from continously call a subroutine. This is one way to make sure a subroutine gets called only when it required. Another trick I have used to to have the subroutine clear the bit that activated it. Seppo must use this trick if he doesn't use one shots.

Click to expand...

Aha! Maybe THAT'S what CMF meant to say?

That idea could come in handy Peter. I've never needed to call a subroutine for only one scan, but that's definately a great way to speed up a program when necessary. Wait a minute... As you mentioned, I call my 'initialization' subroutine(s) for 1 scan, so I guess I HAVE done that...

For the most part, my 'main' program is usually just a bunch of calls. I do that primarily for organizational purposes, so I don't see where I could benefit with one-shotting a call. Most of my calls are conditional, so there's usually just a handful of subroutine running. Rarely do I run into scan time issues, as today's processors are plenty quick. "Back in the day", you had to get pretty clever sometimes in order to get the scan down to a 'usable' speed...

beerchug

-Eric

Peter Nachtwey said:

Seppo must use this trick if he doesn't use one shots.

Click to expand...

In the following circuit 7215Step5 is one shot step and it allmost 'free' without separate evolution. 7215Step5 is there needlessly 2 times, 1 is enough (as well as 7215Step8). This is not Step5 advertise !!