Thread: help - timer in C?

Hi guys. Sorry if this is a regularly posted question, but here goes:

We're currently working on a project for a class based on C programming. Last term, we did the same class, but with MatLab. (I'm not sure if you guys are familiar with the software)

The project that we're assigned this term is to power two 7-segment LED's to reflect elapsed time in seconds. We've searched around, and found "time.h" and the "clock_t" function, but it seems that it only keeps track of time in full seconds? Also, it doesn't seem to have the functionality that we are looking for - we can calculate the difference in time via "difftime(t2-t1)", but in order for this to work effectively, we must calculate the difference constantly.

Right now, we've got a good idea of what the code should look like, but as you can see, we keep running into a wall in regards to measuring time.

The reason I brought up MatLab earlier in the post is because there's a function called tic toc. "tic" starts a timer in MatLab, while "toc" returns a time value since the intial "tic". Is there something in C that is similar to this?

for example, for our last term's project, we did the following:

Code:

<headers, blah blah blah blah etc>
tic

while (toc < [set time values])
{
do stuff
}
etc

Thanks for the help!

edit: I am reading time.h now and there is a method called clock() .. it returns clock_ticks since launch of program , there is also a variable called CLOCKS_PER_SEC .. dividing your ticks by this variable gives you seconds.. (using a double will allow for larger precision than seconds)
do this

Code:

#include <time.h>

main()
{
clock_t launch = clock();
//do work
clock_t done = clock();
double diff = (done - launch) / CLOCKS_PER_SEC;
}

If you are working on Linux there are kernel timers that will give you better granularity in that but unless the kernel is compiled in RTOS mode, any resolution < 20ms (if I recall correctly) will be iffy.

Originally Posted by jeffcobb

If you are working on Linux there are kernel timers that will give you better granularity in that but unless the kernel is compiled in RTOS mode, any resolution < 20ms (if I recall correctly) will be iffy.

It's 10ms/multiples of 10 (I've tested this*). That means the userspace function "nanosleep" should not be called with a granularity finer than 100000ns (1/100th sec).

If you are using *nix I wrote some examples of how to use the userspace timers here:

SourceForge.net: POSIX timers - cpwiki

I would presume that all OS's have some equivalent to nanosleep and/or ntp_gettime.

* "You can test this yourself by calling nanosleep with a 10000 nanosecond (1 millisecond) delay 10000 times -- it will work out to much more than 10 seconds. However, if you ask for 100000 nsecs (1/100th second) 1000 times, you will get exactly 10 seconds."

actually no, clock() does not start nor stop a clock .. it only tells you the time since you started your program

heres a program that prints hello every 1 second ( i think, ive actually never had much use for time.h , i might be wrong )

Code:

#include <time.h>
#include <stdio.h>
main()
{
   clock_t last_check = clock() + CLOCKS_PER_SEC;
   printf("%ld\n",CLOCKS_PER_SEC);
   while( 1)
   {
      if( clock() < last_check  )
           continue;
      printf("HELLO!!"); //change display here
      last_check = clock() + CLOCKS_PER_SEC;
   }
}

edited code, still doesnt work lol