Curious about how loops work - GameCreators Forum

Author

Message

sneaky smith12

20

Years of Service

User Offline

Joined: 30th Apr 2006

Location: Like id tell you, oh wait i just did

Posted: 27th Oct 2010 22:51 Edited at: 27th Oct 2010 22:54

Link

Now, I'm sure someone is going to laugh at that title.

I went ahead and made some code just to have a loop count to 100,000 and sync in each loop. My results I found out where that Do/Loops were on average were the slowest. Goto statements were unpredictable, and the others were pretty consistent with each other.

What causes the differences? Even though there is only about 2/5ths of a second in differences I would think they should be the exact same.

sync on
sync rate 0

// Do... Loop
    loopcount = 0
    ElapsedTime = hitimer()
   
    do
        inc loopcount
        if loopcount = 100000 then exit
        sync
    loop
    
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for  Do... Loop to do 100,000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    sync
    Wait key
    CLS

// Goto...
    loopcount = 0
    ElapsedTime = hitimer()
   
    GotoMe:
        inc loopcount
        if loopcount = 100000 then Goto ExitMe
        Sync
    Goto GotoMe
    ExitMe:

ElapsedTime = hitimer() - ElapsedTime
    Print "Time for GoTo... to do 100,000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    sync
    Wait key
    CLS

//While... EndWhile
    loopcount = 0
    ElapsedTime = hitimer()
    
    While loopcount < 100000
        inc loopcount
        Sync
    EndWhile
   
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for While... EndWhile to do 100,000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    sync
    Wait key
    CLS
    
//Repeat... Until
    loopcount = 0
    ElapsedTime = hitimer()
  
    Repeat 
        inc loopcount
        Sync
    Until loopcount = 100000

ElapsedTime = hitimer() - ElapsedTime
    Print "Time for Repeat... Until to do 100,000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    sync
    Wait key
    CLS

//For... Next
    loopcount = 0
    ElapsedTime = hitimer()

For loopcount = 1 to 100000
        Sync
    Next loopcount
   
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for For... Next to do 100,000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    sync
    Wait key
    
End

+ Code Snippet

sync on
sync rate 0

// Do... Loop
    loopcount = 0
    ElapsedTime = hitimer()
   
    do
        inc loopcount
        if loopcount = 100000 then exit
        sync
    loop
    
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for  Do... Loop to do 100,000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    sync
    Wait key
    CLS

// Goto...
    loopcount = 0
    ElapsedTime = hitimer()
   
    GotoMe:
        inc loopcount
        if loopcount = 100000 then Goto ExitMe
        Sync
    Goto GotoMe
    ExitMe:

    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for GoTo... to do 100,000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    sync
    Wait key
    CLS

//While... EndWhile
    loopcount = 0
    ElapsedTime = hitimer()
    
    While loopcount < 100000
        inc loopcount
        Sync
    EndWhile
   
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for While... EndWhile to do 100,000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    sync
    Wait key
    CLS
    
//Repeat... Until
    loopcount = 0
    ElapsedTime = hitimer()
  
    Repeat 
        inc loopcount
        Sync
    Until loopcount = 100000

    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for Repeat... Until to do 100,000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    sync
    Wait key
    CLS

//For... Next
    loopcount = 0
    ElapsedTime = hitimer()

    For loopcount = 1 to 100000
        Sync
    Next loopcount
   
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for For... Next to do 100,000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    sync
    Wait key
    
End

Typical results I get:

+ Code Snippet

Do... Loop: 7535 ms
Goto: 7434 ms
While... EndWhile: 7122 ms
Repeat... Until: 7299 ms
For... Next: 7448 ms

If at first you dont succeed, LOWER YOUR STANDARDS.

Back to top

Profile PM Email

Kevin Picone

23

Years of Service

User Offline

Joined: 27th Aug 2002

Location: Australia

Posted: 27th Oct 2010 22:57

Link

The test is not counting the speed of the loops, but the page swapping speed.

New PlayBasic Learning Edition Released 24th April 2010

Back to top

Profile PM Website

BatVink

Moderator

23

Years of Service

User Offline

Joined: 4th Apr 2003

Location: Gods own County, UK

Posted: 27th Oct 2010 23:01

Link

with those results, you can't really tell. Any number of Windows functions could have interfered. Even the order in which you run the tests can have an impact, for example it's very normal for a first test to take longer as it grabs resources and possibly the last to take longer as resources get fragmented.

Back to top

Profile PM Email

sneaky smith12

20

Years of Service

User Offline

Joined: 30th Apr 2006

Location: Like id tell you, oh wait i just did

Posted: 28th Oct 2010 00:17

Link

Quote: "with those results, you can't really tell. Any number of Windows functions could have interfered. Even the order in which you run the tests can have an impact, for example it's very normal for a first test to take longer as it grabs resources and possibly the last to take longer as resources get fragmented.
"

so now if I ran each test say 1,000 times and took the averages... wouldn't that fix the problem?

Quote: "The test is not counting the speed of the loops, but the page swapping speed."

Shouldn't the page swapping speed be the same regardless of the type of loop since there is nothing changing in the screen during the loop?

If at first you dont succeed, LOWER YOUR STANDARDS.

Back to top

Profile PM Email

Daniel TGC

Retired Moderator

19

Years of Service

User Offline

Joined: 19th Feb 2007

Location: TGC

Posted: 28th Oct 2010 20:03 Edited at: 28th Oct 2010 20:04

Link

Running the tests on a freshly formatted and installed machine Windows (not vendor supplied version which often comes with alot of their installed software). This mean no virus checkers, or any other 3rd party installed software.

This means you'll get a base comparison. However even eliments suchs as background indexing, can prevent a consistant score each time. The simple fact of the matter is, Windows operating systems are extremely busy places with alot going on the user doesn't see. To get an idea of this just do a control all delete and look at all the processes and services going on.

Back to top

Profile PM Email Website

Kevin Picone

23

Years of Service

User Offline

Joined: 27th Aug 2002

Location: Australia

Posted: 28th Oct 2010 20:20

Link

Quote: "Shouldn't the page swapping speed be the same regardless of the type of loop since there is nothing changing in the screen during the loop?"

In theory, but we're not the only thing running on the system. Time is being eaten not purely by our program, but other applications in the background.

New PlayBasic Learning Edition Released 24th April 2010

Back to top

Profile PM Website

sneaky smith12

20

Years of Service

User Offline

Joined: 30th Apr 2006

Location: Like id tell you, oh wait i just did

Posted: 28th Oct 2010 20:55 Edited at: 28th Oct 2010 21:13

Link

Quote: " Windows operating systems are extremely busy places with alot going on the user doesn't see"

True, but I'm not exactly sure how that effects the speed of the test if the RAM is sufficient to have excess RAM left over for the program and System Idle Process is 99% CPU usage.

Quote: "In theory, but we're not the only thing running on the system. Time is being eaten not purely by our program, but other applications in the background."

I guess I could switch it to something other than sync... maybe just add an integer, but I might need to make a bunch more loops to see if theirs a time difference.

EDIT: Actually, I just changed it to perform no calculations/commands that don't relate to the loop.

Results after 1 million loops taken once:

+ Code Snippet

Baseline(taken after program loaded and got to first wait key):
     CPU Usage: 27%
     Mem Usage: 23%
Peak usages(during program):
     CPU Usage: 29% (No more than 75% on one core)
     Mem Usage: 23%

Do/Loop = 77315 ms
Goto = 78209 ms
While/EndWhile = 78227 ms
Repeat/Until = 78411 ms
For/Next = 2 ms

Source code:

// Do... Loop
    loopcount = 0
    ElapsedTime = hitimer()
   
    do
        inc loopcount
        if loopcount = 1000000 then exit
    loop
    
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for  Do... Loop to do 1000000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    Wait key
    CLS

// Goto...
    loopcount = 0
    ElapsedTime = hitimer()
   
    GotoMe:
        inc loopcount
        if loopcount = 1000000 then Goto ExitMe
    Goto GotoMe
    ExitMe:

ElapsedTime = hitimer() - ElapsedTime
    Print "Time for GoTo... to do 1000000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    Wait key
    CLS

//While... EndWhile
    loopcount = 0
    ElapsedTime = hitimer()
    
    While loopcount < 1000000
        inc loopcount
    EndWhile
   
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for While... EndWhile to do 1000000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    Wait key
    CLS
    
//Repeat... Until
    loopcount = 0
    ElapsedTime = hitimer()
  
    Repeat 
        inc loopcount
    Until loopcount = 1000000

ElapsedTime = hitimer() - ElapsedTime
    Print "Time for Repeat... Until to do 1000000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    Wait key
    CLS

//For... Next
    loopcount = 0
    ElapsedTime = hitimer()

For loopcount = 1 to 1000000
    Next loopcount
   
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for For... Next to do 1000000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    Wait key
    
End

+ Code Snippet

// Do... Loop
    loopcount = 0
    ElapsedTime = hitimer()
   
    do
        inc loopcount
        if loopcount = 1000000 then exit
    loop
    
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for  Do... Loop to do 1000000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    Wait key
    CLS

// Goto...
    loopcount = 0
    ElapsedTime = hitimer()
   
    GotoMe:
        inc loopcount
        if loopcount = 1000000 then Goto ExitMe
    Goto GotoMe
    ExitMe:

    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for GoTo... to do 1000000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    Wait key
    CLS

//While... EndWhile
    loopcount = 0
    ElapsedTime = hitimer()
    
    While loopcount < 1000000
        inc loopcount
    EndWhile
   
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for While... EndWhile to do 1000000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    Wait key
    CLS
    
//Repeat... Until
    loopcount = 0
    ElapsedTime = hitimer()
  
    Repeat 
        inc loopcount
    Until loopcount = 1000000

    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for Repeat... Until to do 1000000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    Wait key
    CLS

//For... Next
    loopcount = 0
    ElapsedTime = hitimer()

    For loopcount = 1 to 1000000
    Next loopcount
   
    ElapsedTime = hitimer() - ElapsedTime
    Print "Time for For... Next to do 1000000 loops = " + str$(ElapsedTime)
    Print "Press any key to continue."
    Wait key
    
End

If at first you dont succeed, LOWER YOUR STANDARDS.

Back to top

Profile PM Email

Kevin Picone

23

Years of Service

User Offline

Joined: 27th Aug 2002

Location: Australia

Posted: 28th Oct 2010 21:11

Link

The longer the period, the less likely you're timing the loop on it's own. A smaller period will display a bias just as well, providing you average the results over time.

New PlayBasic Learning Edition Released 24th April 2010

Back to top

Profile PM Website

Indicium

18

Years of Service

User Offline

Joined: 26th May 2008

Location:

Posted: 28th Oct 2010 21:43

Link

In my experience, the inc command is slow, replace it with loopcount = loopcount + 1

Back to top

Profile PM

baxslash

Valued Member

Bronze Codemaster

19

Years of Service

User Offline

Joined: 26th Dec 2006

Location: Duffield

Posted: 28th Oct 2010 22:08

Link

Quote: "In my experience, the inc command is slow, replace it with loopcount = loopcount + 1"

Really, I didn't know that! I always assumed it was quicker otherwise... why have it?

...OK so I did a test using this:

+ Code Snippet

st1=timer()
for x=1 to 100000
	cls
	print count
	inc count
next
ft1=timer()

count=0

st2=timer()
for x=1 to 100000
	cls
	print count
	count=count+1
next
ft2=timer()

do
	cls
	set cursor 0,0
	print "results"
	print "using inc took ",ft1-st1," milliseconds"
	print "using + took ",ft2-st2," milliseconds"
loop

...and there was nothing in it (I ran it about 20 times)... I guess "inc" wins cause it's quicker to type?

Back to top

Profile PM Email

Indicium

18

Years of Service

User Offline

Joined: 26th May 2008

Location:

Posted: 29th Oct 2010 00:02

Link

I'll run my own test, to see if it's specific to my computer, but I'm sure it was slower...

But yeah, you're right... I don't know where I got that from lol.

Back to top

Profile PM

sneaky smith12

20

Years of Service

User Offline

Joined: 30th Apr 2006

Location: Like id tell you, oh wait i just did

Posted: 29th Oct 2010 08:33

Link

After seven trials with baxslash's code I got about 3 miliseconds slower with the Inc command.

If I changed it to inc count, 1 though it seemed to slow down even more. Like 10 ms slower.

If at first you dont succeed, LOWER YOUR STANDARDS.

Back to top

Profile PM Email

Sorry your browser is not supported!

DarkBASIC Professional Discussion / Curious about how loops work