Found solution to - Having trouble determining elapsed time with pasm

Marc Gebauer

*** Found the soulton to my problem looking at how FullDuplexSerial determined bittime. ***

I call settimelimit from my top object·which·calculates the number of ticks for·my duration in·milliseconds and stores it in the time_limit parameter:

PUB settimelimit(duration)
· time_limit := clkfreq/1000*duration+cnt

Then·the main loop of my·pasm·routine does the following:

loop

· <do this all the time>

······· rdlong timelimit, par
······· sub timelimit,cnt
······· cmps··timelimit,#0·· wc
if_nc··jmp· #loop·· 'if time has not elapsed jump to·loop

· <do this if time has elapsed>

······· jmp· #loop




·

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Post Edited (Marc Gebauer) : 3/5/2009 6:58:41 PM GMT

Cluso99

CNT does not start from 0. It is 32 bits continually incrementing counter. You would be better to do

rdlong tl,par
add tl,cnt
waitcnt tl,0

Think the above is correct (hope you get the idea).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:

· Home of the MultiBladeProps:· SixBladeProp, TriBladeProp
· Prop Tools under Development or Completed (Index)
· Emulators (Micros eg Altair, and Terminals eg VT100) - index
· Search the Propeller forums (via Google)

My cruising website is: ·www.bluemagic.biz·· MultiBladeProp is: www.bluemagic.biz/cluso.htm

Marc Gebauer

I don't want to stop the cog. I need to keep processing until the time limit is exceeded. Plus I want to set the time limit by calling a spin routine from·top object.


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Post Edited (Marc Gebauer) : 3/5/2009 5:08:55 AM GMT

Kye

Okay, don't use cmps, infact, don't use signed operations unless you really need them in the operation. In the situation you're using it in counter value is not signed.
Use cmp, just by itself.

Also, note that the cnt just loops arround over and over again. So the comparison you're using will fail·when cnt is less than your time limit... That being said look what hapens when the counter wraps arround.

For a fix, first have some flag register for the processor in asm to start the loop. Once its started the processor should get the current value of cnt, and continaully compare that older value with the newest current until the difference is greater than the time limit. This method should always work.

You will need to add in some ability to correct abs differences addtionally.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nyamekye,

Post Edited (Kye) : 3/5/2009 5:21:54 AM GMT

Marc Gebauer

I tried cmp and that did not work either.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

Kye

I edited my post.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nyamekye,

Marc Gebauer

Studied how bit time was determined in Fullduplex serial and came up with the solution to my problem. See original post.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔