Code optimization tips and stuff.

[Agamemnus]

Mod and \ are not the same operation.

Oh, well in that case, mod is still faster! So ignore Blitz and use mod!

Ffix doesn't apply to integers... does it?

[Glenn]

MOD and "\" are the same thing? (Are my glasses *that* bad?)

[Agamemnus]

he didn't. I thought he did.


but he did say:

The mod instruction uses division, so you should also avoid it.

[Blitz]

Christ, is my english taht bad?
I said MOD USES division, i never said replace mod with division. I said replace mod which are powers of to with AND. Otherwise avoid it in time critical parts.

I didn't even have to write code to know that it's correct, but since you seem keen on me being wrong here.

MODs per second: 37577982
ANDs per second: 390095238
AND is percent faster: 1038

Code:

defint a-z

dim a as integer
dim i as integer, j as integer
dim tmrAIni as double, tmrBIni as double
dim tmrAEnd as double, tmrBEnd as double
dim modspersec as double, andspersec as double

tmrAIni = timer
for  i = 0 to 199
    for  j = 0 to 31999
        a = j mod 2
        a = j mod 4
        a = j mod 8
        a = j mod 16
        a = j mod 32
        a = j mod 64
        a = j mod 128
        a = j mod 256
        a = j mod 512
        a = j mod 1024
    next j
next i
tmrAEnd = timer

tmrBIni = timer
for  i = 0 to 199
    for  j = 0 to 31999
        a = j and 1
        a = j and 3
        a = j and 7
        a = j and 15
        a = j and 31
        a = j and 63
        a = j and 127
        a = j and 255
        a = j and 511
        a = j and 1023
    next j
next i
tmrBEnd = timer


modspersec = (32000.0#*200.0#*10.0#) / (tmrAEnd-tmrAIni)
andspersec = (32000.0#*200.0#*10.0#) / (tmrBEnd-tmrBIni)

print "MODs per second:" + str$( clng( modspersec ) )
print "ANDs per second:" + str$( clng( andspersec ) )
print "AND is percent faster:" + str$( cint( andspersec*100.0#/modspersec ) )

[DrV]

You said use / for integer division and \ for float division. That's backwards. \ chops off the fractional part of a division.

[Glenn]

I'm calling my eye doctor tomorrow!

[Agamemnus]

anyways, you still don't have a point. What if I want mod 1000, or say mod 1515?

[Blitz]

Oh and the obvious things, which i way to to often see people who're not aware of.

Do not call a routine to move say just one object, make a routine that moves all of the objects in that same call. QB subs/functions are horribly slow, it amazes me that back then when qb was made cpu power was a something you had very little of. Yet they made qb which produces such bad code and does so stupid things. Ok, compiler technology wasn't very advanced back then. But it does so many stupid things that just goes against common sense. What were they thinking?

And Agamemnus, just give it up ok? If you have to use it, then you have to use it. Thus me saying "avoid it" and not "don't use it, i'll kill you if you do!!¤%". But there's always a better way, programming is inventing. And if an inventor can't invent, then he/she is pretty screwed aren't they ?

[Agamemnus]

Fine, I will, but AND 2^n is not always a substitute for MOD...... and you offer no alternative.

So you ought to say "you should avoid MOD if your number is represented in binary form as a one and a bunch of zeroes."

[Glenn]

And Blitz didn't say "always". He said "when". (That word, conceptually, is sort of like the word "if".)