Repeating negative numbers.

Comments

• Rayman Posts: 15,248

  2011-03-15 18:42 edited 2011-03-15 18:42

  I'd guess 2^32 - 1 ...
• kuroneko Posts: 3,623

  2011-03-15 18:53 edited 2011-03-15 18:53

  I'd guess 2^32 - 1 ...

  Yup, think tjz/djnz.
• Rayman Posts: 15,248

  2011-03-15 19:02 edited 2011-03-15 19:02

  Hard to tell with Spin repeat though...
  A long time ago I tried repeat... step -1 and got completely unexpected behavior....
• Phil Pilgrim (PhiPi) Posts: 23,514

  2011-03-15 19:11 edited 2011-03-15 19:11

  I looked at the interpreter source, but didn't see any comments related to non-variable finite loops. But I suspect they're the same as variable-based finite loops, but with an internal variable. So it's probably necessary to examine the byte code produced for a definitive answer. Either that, or set up a counter and wait a few hours for the result.
  
  In any event, I suspect the short answer is, "One more time than repeat -2."
  
  -Phil
• Kye Posts: 2,200

  2011-03-15 19:22 edited 2011-03-15 19:22

  Mmm, no quick answer for me...
• Heater. Posts: 21,230

  2011-03-16 00:09 edited 2011-03-16 00:09

  "repeat -n" causes the statements in the following indented block to be executed n times in reverse order.
  
  As far as I know this is a feature that is unique to Spin:-)
• kwinn Posts: 8,697

  2011-03-16 00:22 edited 2011-03-16 00:22

  I would expect this to depend on how the spin interpreter handles the negative number. Totally dependent on the interpreter program.
• Martin_H Posts: 4,051

  2011-03-16 07:05 edited 2011-03-16 07:05

  Nyamekye, I never would have thought to try a negative repeat. I'd bet you would be good at computer security if you have an interest in it.
• Kye Posts: 2,200

  2011-03-16 07:57 edited 2011-03-16 07:57

  Yeah, its just a computer security problem. I want to know if I have to limit the input value or not.
  
  @heater - your joking right?
• Heater. Posts: 21,230

  2011-03-16 08:14 edited 2011-03-16 08:14

  Kye,

  @heater - your joking right?

  
  Might be:)
• lonesock Posts: 917

  2011-03-16 09:19 edited 2011-03-16 09:19

  Yep, it behaves exactly like a DJNZ...check out the attached source.
  
  * An empty repeat loop takes about 224 clocks per pass
  * So, at 80MHz, the counter variable would be changing at a rate of ~357000 per second
  * Monitoring Hub RAM for values that changed at that rate found the counter
  * watching the counter showed that it starts at whatever value you specify, so
  * starting at -1 will keep going negative at that specified rate, terminating one the counter reaches 0
  * at 80MHz, it will take about 3.34 hours to complete an empty "repeat (-1)"
  
  Jonathan

  spin

  spin

  test_repeat.spin

  1K

  test_repeat.spin 1.1K
• Kye Posts: 2,200

  2011-03-16 09:29 edited 2011-03-16 09:29

  Thank you very much lonesock! This thread will add to the forum's knowledge.
  
  (Why doesn't parallax document this...)
• Phil Pilgrim (PhiPi) Posts: 23,514

  2011-03-16 09:50 edited 2011-03-16 09:50

  Kye wrote:

  Why doesn't parallax document this...

  Probably for the same reason it's taken one of us five years to discover it. They didn't think of it either.
  
  -Phil
• Duane Degn Posts: 10,588

  2011-03-16 10:02 edited 2011-03-16 10:02

  I think I had assumed the loop would be skipped if the repeat amount was negative. This is good to know.
  
  So far I know the repeat occurs more than 100,000,000 times (I left a program runing while reading the forums). I think lonesocks answer is better.
• Dave Hein Posts: 6,347

  2011-03-16 10:09 edited 2011-03-16 10:09

  I ran this under SpinSim, and the debug listing is shown below. It loads a -1 on the stack, and first tests for zero. It then executes a djnz until it reaches zero. The stack pointer and the value at the top of the stack are shown in columns 2 and 3. The columns after the "-" are the program counter, bytecodes and the instruction mnemonic. The number after the tjz and djnz instructions is a jump offset.
  
  Dave
  

  Cog 0: 002c 00000000 - 0018 34              ldlim1
  Cog 0: 0030 ffffffff - 0019 08 02           tjz 2
  Cog 0: 0030 ffffffff - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffffe - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffffd - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffffc - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffffb - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffffa - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffff9 - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffff8 - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffff7 - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffff6 - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffff5 - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffff4 - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffff3 - 001b 09 7e           djnz -2
  Cog 0: 0030 fffffff2 - 001b 09 7e           djnz -2