Post by thread:compact sort program

Hi! I have written a program for a PIC16F84 that emulate a motorized 2 cam programmable daily timer. I keep in PIC's eeprom the setpoint (20 max) in the following 24h BCD format 1)byte HH 2)byte mm 3)byte which bit I want to set/reset . Every minute I compare the actual time with the setpoint and if time is >= setpoint I copy the third byte in a buffer then repaet for next setpoint. At the end of eeprom , I copy the buffer to the out port . This works properly only if setpoint (HH:mm) are in ascending order. Unfortunately I have only 27 bytes of code free and 25 of ram. Does anyone have any code to do this sort in few bytes of code ; or some suggestion to solve my problem ? (I have already sqeezed the program as possible). Thanks Gianni

On Thu, 31 Dec 1998 18:08:30 +0100 Gianni <spam_OUTgia.ma.fraTakeThisOuTIOL.IT> writes: >Hi! >I have written a program for a PIC16F84 that emulate a motorized 2 cam >programmable daily timer. >I keep in PIC's eeprom the setpoint (20 max) in the following >24h BCD format 1)byte HH 2)byte mm 3)byte which bit I want to >set/reset . >Every minute I compare the actual time with the setpoint and if time >is >= >setpoint I copy the third byte in a buffer then repaet for next >setpoint. >At the end of eeprom , I copy the buffer to the out port . >This works properly only if setpoint (HH:mm) are in ascending order. It seems like it would work no matter what the order is, if you check all the setpoints every minute. The only conflict would be if two or more setpoints are telling the same output to go opposite ways the same minute. In that case, the last one in the table would dominate. >Unfortunately I have only 27 bytes of code free and 25 of ram. If I were writing this (in assembler), I can't imagine the time-keeping and setpoint-checking parts of the program needing more than 100 or so code locations. The rest of the space could be used by the likely much more complicated code to provide a user interface to set the setpoints. It may be time to review the situation and re-write your program entirely. Most of my projects go through at least one such total rewrite. ___________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com/getjuno.html or call Juno at (800) 654-JUNO [654-5866]

On Thu, 31 Dec 1998 18:08:30 +0100, you wrote: >Hi! >I have written a program for a PIC16F84 that emulate a motorized 2 cam >programmable daily timer. >I keep in PIC's eeprom the setpoint (20 max) in the following >24h BCD format 1)byte HH 2)byte mm 3)byte which bit I want to set/reset . >Every minute I compare the actual time with the setpoint and if time is >= >setpoint I copy the third byte in a buffer then repaet for next setpoint. >At the end of eeprom , I copy the buffer to the out port . >This works properly only if setpoint (HH:mm) are in ascending order. >Unfortunately I have only 27 bytes of code free and 25 of ram. >Does anyone have any code to do this sort in few bytes of code ; or some >suggestion to solve my problem ? (I have already sqeezed the program as >possible). > Thanks > Gianni How are the setpoints generated - could you ensure ordering by either 1) if all points are loaded into the PIC from an external source,. sort them before sending to the PIC 2) If they're added one-at-a-time, insert in the correct place in the list as each one is created

| That's exactly what I need ! But it's annoyng overwrite manually |a lot of setpoints every time I add /change a new value. |Perhaps is suitable change comparison mode rather than sort the table; |but I haven't idea how to plan this. What about searching through the list for the largest setpoint time which is less-than-or-equal-to the current time? Something like: best_time = 0; best_index = 0; For each setpoint If time[this_one] <= current_time and time[this_one] > best_time best_index = this_one best_time = time[this_one] endif end loop That should work no matter what order things are in.