Page 2 - [DNO] Scoring System Benchmark

Codewalkers Forums Sponsor:

#16

March 7th, 2003, 04:43 PM

-vertigo-

Join Date: Apr 2007

Location: Louth, Lincolnshire

Posts: 314

Time spent in forums: 3 m 24 sec
Reputation Power: 2

[DNO] Food for thought

This contest reminds me of the infamous discs contest.

The duplicates caused major trouble for most people, but the winner realised that there was far too little time to use logic to handle the duplicates when large stacks were used.

He handled the duplicates in a totally simple way, and optimised for large stacks.

Similarly, the winner of this one will likely use simple logic and optimise the 50x50x8 cases.

It surely won't be possible to analyse all potential removals to determine the most promising one on each move, it seems too complex.

Lets look at this grid.

123
211
332

The most promising starting moves seem to be (1, 2), (2, 1) and (3, 1).

The first two produce scores of 21, and the last produces a score of 23. I don't see any logical way of determining that the third possibility is better than the other two.

Of course, thinking like this is a bad idea, because the large grids will determine the winner, not the small ones.

#17

March 7th, 2003, 04:44 PM

mcoder

Join Date: Apr 2007

Posts: 130

Time spent in forums: < 1 sec
Reputation Power: 2

RE: [DNO] Scoring System Benchmark

>$start=time()+microtime();

This timing routine is false !
Suppose you start to use it at 22s 999999.
time() will return 22, and microtime() will return 0.
So you have one second of difference !

Personally, I use:
$time=explode(" ",microtime());
return $time[1]+$time[0];

JC

#18

March 7th, 2003, 05:04 PM

Anonymous

Registered User

Codewalkers God 35th Plane (22000 - 22499 posts)

Join Date: Apr 2007

Posts: 22,309

Time spent in forums: < 1 sec
Reputation Power: 24

RE: [DNO] Food for thought

Quote:

Of course, thinking like this is a bad idea, because the large grids will determine the winner, not the small ones.

Ahhh..but if everyone uses the same dumb algorithm to get through the big ones then the contest will be one with the small ones ;)

#19

March 7th, 2003, 10:26 PM

pickleman78

Join Date: Apr 2007

Location: Dallas,TX,USA

Posts: 582

Time spent in forums: < 1 sec
Reputation Power: 2

RE: [DNO] Scoring System Benchmark

lol
101 seconds on my 566 celeron POS
You and your nefangled lotsd of ghz. Too bad new computers cost money, I spent all of 50 bucks on this piece of crap, but hey, it runs

#20

March 8th, 2003, 07:51 AM

tkarkkainen

Moderator

Join Date: Apr 2007

Location: Finland

Posts: 2,326

Folding Points: 15677 Folding Title: Novice Folder

Time spent in forums: 6 Days 9 h 32 m 40 sec
Reputation Power: 4

RE: [DNO] Scoring System Benchmark

55,10490500927 seconds

Maybe this Pentium II 450MHz isn't so slow after all.

#21

March 8th, 2003, 01:14 PM

Anonymous

Registered User

Join Date: Apr 2007

Posts: 22,309

Time spent in forums: < 1 sec
Reputation Power: 24

RE: [DNO] Scoring System Benchmark

How about giving some kind of "codewalker rating" when comparing individual progress on contest scripts? Like "my script finishes in 10.3 seconds at a cwr of 1.42854" or a shorter "10.3 s @ 1.42854 cwr". Or you could use your cwr to calculate an estimated time and post that if you prefer.

I edited the original script by Matt by adding the new timelimit setting and cw_rating calculation plus it's echoing:

php Code:

Original - php Code
<?php set_time_limit(300); // 5 mins should be enough function getmicrotime(){ list($usec, $sec) = explode(" ",microtime()); return ((float)$usec + (float)$sec); } $time_start = getmicrotime(); for ($i=0; $i < 26000000; $i++); $time_end = getmicrotime(); $time = $time_end - $time_start; $cw_rating = 30 / $time; echo "Script ran $time seconds<br>Your Codewalkers-Ratig is $cw_rating"; ?>

 
<?php
set_time_limit(300); // 5 mins should be enough
 
function getmicrotime(){ 
    list($usec, $sec) = explode(" ",microtime()); 
    return ((float)$usec + (float)$sec); 
    } 
 
$time_start = getmicrotime();
    
for ($i=0; $i < 26000000; $i++);
 
$time_end = getmicrotime();
$time = $time_end - $time_start;
$cw_rating = 30 / $time;
 
echo "Script ran $time seconds<br>Your Codewalkers-Ratig is $cw_rating";
?>

BTW: my PIII 450 MHz needs 60 secs to complete. On my P1 166 MHz I wasn't able to get a result in any browser, they all timed out before the script was through ;)
Maybe I should give the command line a try...

#22

March 8th, 2003, 01:35 PM

mcoder

Join Date: Apr 2007

Posts: 130

Time spent in forums: < 1 sec
Reputation Power: 2

RE: [DNO] Scoring System Benchmark

Good idea !
But, I get a cwr varying from 2.40 to 2.82.

JC

#23

March 9th, 2003, 08:28 PM

fractalbit

$fractalbit is offline$

Join Date: Apr 2007

Posts: 108 $fractalbit User rank is Just a Lowly Private (1 - 20 Reputation Level)$
Time spent in forums: < 1 sec
Reputation Power: 2

RE: [DNO] Scoring System Benchmark

Lets have more realistic results than just a for loop that does nothing. Matt please modify the following code to run at about 30 seconds on the testing machine. As it currently is, it runs on 13.7secs on mine(AthlonXP 1800+)

php Code:

Original - php Code
<?php function getmicrotime(){ list($usec, $sec) = explode(" ",microtime()); return ((float)$usec + (float)$sec); } $time_start = getmicrotime(); $max = 0; for ($i=0; $i < 3050000; $i++){ $x = rand(40, 100); $z = rand(50, 60); $y = $x + $z; if($y>$max) $max = $y; } $time_end = getmicrotime(); $time = $time_end - $time_start; echo "MAX = $max - $time seconds"; ?>

 
<?php
function getmicrotime(){ 
    list($usec, $sec) = explode(" ",microtime()); 
    return ((float)$usec + (float)$sec); 
    } 
 
$time_start = getmicrotime();
 
$max = 0;    
for ($i=0; $i < 3050000; $i++){
   $x = rand(40, 100);
   $z = rand(50, 60);
   $y = $x + $z;
   if($y>$max) $max = $y;    
}
 
$time_end = getmicrotime();
$time = $time_end - $time_start;
 
echo "MAX = $max - $time seconds";
?>

#24

March 11th, 2003, 11:05 AM

nodream

Join Date: Apr 2007

Location: Norway

Posts: 22

Time spent in forums: 5 m 59 sec
Reputation Power: 0

RE: [DNO] Scoring System Benchmark

On my PII-350 I get 75sec.

How reliable is this time?
Is it safe to work whit a timelimit of 75 sec on my computer, and it will roughly (+/- 1sec) get a runtime of 30 sec on the scoring-macine.

Thread Tools	Search this Thread
Email this Page	Search this Thread: Advanced Search
Display Modes	Rate This Thread
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode	Rate This Thread: