Is it preferred to have geometry figures (E.g. walls) by the power of two?
75 replies, posted
This may sound a bit... noobish, but I have a question about mapping.
My friend said that walls e.t.c. is preferably by the power or two (2x2, 4x4, 8x8 and so on, or 2x8, 128x512) because they run faster. I don't know what the logic may be in that, but I don't have enough knowledge about the source engine to be certain.
Hope you can reply quick,
Dan.
[QUOTE=ZemogT;16016968]This may sound a bit... noobish, but I have a question about mapping.
My friend said that walls e.t.c. is preferably by the power or two (2x2, 4x4, 8x8 and so on, or 2x8, 128x512) because they run faster. I don't know what the logic may be in that, but I don't have enough knowledge about the source engine to be certain.
Hope you can reply quick,
Dan.[/QUOTE]
I believe it's a fact because of the FPS and the speed of the compiling when making / testing the map. I have no idea why tho, plus it looks a bit better.
I always make things in powers of 2 when I can. Makes scaling easier and satisfies my OCD.
[QUOTE=GoatR;16017046]I believe it's a fact because of the FPS and the speed of the compiling when making / testing the map. I have no idea why tho, plus it looks a bit better.[/QUOTE]
Alright. But since you have no idea why I'd like to hear the fact behind it, if anyone knows why?
I think the compiler and engine interpret it better in powers of two, I've always done it as that is what I've always known.
Textures in the Source engine are [B]always[/B] powers of two. Therefore, if your brush dimensions are powers of two, textures should be easier to align because they fit better on the brush faces.
Yea. Powers of 2 walls make texture aligning really easy... plus, 8 + 16 are perfect wall thicknesses
The Quake engine (or more specifically, Source) runs best on powers of 2 because it builds the levels on a grid. This means that predicting visibility and lightmaps becomes faster.
What Yobdren means; to put it simply, things that are to the power of 2 (or any larger multiple) helps decrees compile times, and makes sure nothing randomly disappears because of bad visibility rendering, and help light look better.
The amount of time your saving by making all your walls a power of two is probably negated by the loss of detail and such. This isn't 1985.
[QUOTE=Lord Ned;16020582]The amount of time your saving by making all your walls a power of two is probably negated by the loss of detail and such. This isn't 1985.[/QUOTE]
Good year by the way
no, all source textures are 93.372 units. dont be fooled!!11
[QUOTE=funnymuffin;16021208]no, all source textures are 93.372 units. dont be fooled!!11[/QUOTE]
In contrary to your name, you're really not funny at all.
Okay this is a really stupid question, but what exactly do you mean with "power of 2"? Thanks!
Everything = times by 2. For example:
1 x 2 = [b]2[/b]
2 x 2 = [b]4[/b]
4 x 2 = [b]8[/b]
8 x 2 = [b]16[/b]
16 x 2 = [b]32[/b]
32 x 2 = [b]64[/b]
64 x blah = blah
blah x blah = blah
Etc ect, you get the point.
All those bolded numbers are numbers by the power of 2. When you choose what scale you want to work with in the hammer grid (By using the [ and ] keys), it will always go with numbers by the power of 2.
[QUOTE=lew06;16021731]Everything = times by 2. For example:
1 x 2 = [b]2[/b]
2 x 2 = [b]4[/b]
4 x 2 = [b]8[/b]
6 x 2 = [b]16[/b]
16 x 2 = [b]32[/b]
32 x 2 = [b]64[/b]
All those bolded numbers are number by the power of 2, hence easier to work with. When you scale what scale you want the grid to be in hammer (Using [ and ]), it will always use numbers by the power of 2.
And so on.[/QUOTE]
1 x 2 = [b]2[/b]
2 x 2 = [b]4[/b]
4 x 2 = [b]8[/b]
8 x 2 = [b]16[/b]
16 x 2 = [b]32[/b]
32 x 2 = [b]64[/b]
64 x 2 = [b]128[/b]
128 x 2 = [b]256[/b]
256 x 2 = [b]512[/b]
512 x 2 = [b]1024[/b]
1024 x 2 = [b]2048[/b]
2048 x 2 = [b]4096[/b]
4096 x 2 = [b]8192[/b]
8192 x 2 = [b]16384[/b]
16384 x 2 = [b]32768[/b]
32768 x 2 = [b]65536[/b]
65536 x 2 = [b]131072[/b]
131072 x 2 = [b]262144[/b]
262144 x 2 = [b]524288[/b]
524288 x 2 = [b]1048576[/b]
1048576 x 2 = [b]2097152[/b]
2097152 x 2 = [b]4194304[/b]
4194304 x 2 = [b]8388608[/b]
8388608 x 2 = [b]16777216[/b]
16777216 x 2 = [b]33554432[/b]
33554432 x 2 = [b]67108864[/b]
67108864 x 2 = [b]134217728[/b]
134217728 x 2 = [b]268435456[/b]
268435456 x 2 = [b]536870912[/b]
536870912 x 2 = [b]1073741824[/b]
1073741824 x 2 = [b]2147483648[/b]
2147483648 x 2 = [b]4294967296[/b]
4294967296 x 2 = [b]8589934592[/b]
8589934592 x 2 = [b]17179869184[/b]
17179869184 x 2 = [b]34359738368[/b]
34359738368 x 2 = [b]68719476736[/b]
68719476736 x 2 = [b]137438953472[/b]
137438953472 x 2 = [b]274877906944[/b]
274877906944 x 2 = [b]549755813888[/b]
549755813888 x 2 = [b]1099511627776[/b]
1099511627776 x 2 = [b]2199023255552[/b]
2199023255552 x 2 = [b]4398046511104[/b]
4398046511104 x 2 = [b]8796093022208[/b]
8796093022208 x 2 = [b]17592186044416[/b]
17592186044416 x 2 = [b]35184372088832[/b]
35184372088832 x 2 = [b]70368744177664[/b]
70368744177664 x 2 = [b]140737488355328[/b]
140737488355328 x 2 = [b]281474976710656[/b]
281474976710656 x 2 = [b]562949953421312[/b]
562949953421312 x 2 = [b]1125899906842624[/b]
1125899906842624 x 2 = [b]2251799813685248[/b]
2251799813685248 x 2 = [b]4503599627370496[/b]
4503599627370496 x 2 = [b]9007199254740992[/b]
9007199254740992 x 2 = [b]18014398509481984[/b]
18014398509481984 x 2 = [b]36028797018963968[/b]
36028797018963968 x 2 = [b]72057594037927936[/b]
72057594037927936 x 2 = [b]144115188075855872[/b]
144115188075855872 x 2 = [b]288230376151711744[/b]
288230376151711744 x 2 = [b]576460752303423488[/b]
576460752303423488 x 2 = [b]1152921504606846976[/b]
1152921504606846976 x 2 = [b]2305843009213693952[/b]
2305843009213693952 x 2 = [b]4611686018427387904[/b]
4611686018427387904 x 2 = [b]9223372036854775808[/b]
9223372036854775808 x 2 = [b]????[/b]
18446744073709551616
-snip- ninjad
Though all large brushes are divided into 1024 units.
I once used a million digit calculator I wrote in C++ to calculate the highest power of two I could get. I won't ever find it again, because it almost crashed my computer (Rig: [url]http://www.xfire.com/profile/zemogt/[/url] ) but I can tell you this much:
Large as fu-
[QUOTE=Chili Banan;16025879]What the fuck? Honestly?[/QUOTE]
If you think that's a lot, [url=http://www.spikad.se/files/powers_of_two.txt]this is what it made in 20 minutes[/url]
That's some pretty huge numbers. Computers...
[QUOTE=Sippeangelo;16026402]If you think that's a lot, [url=http://www.spikad.se/files/powers_of_two.txt]this is what it made in 20 minutes[/url]
That's some pretty huge numbers. Computers...[/QUOTE]
By the sounds of it, you guys are using some sort of basic program that simply multiplies the previous result by 2. It would be quicker to just write an equation for it then just figure out the billionth result in seconds.
[QUOTE=HiddenMyst;16027032]By the sounds of it, you guys are using some sort of basic program that simply multiplies the previous result by 2. It would be quicker to just write an equation for it then just figure out the billionth result in seconds.[/QUOTE]
Defeats the whole point.
[QUOTE=HiddenMyst;16027032]By the sounds of it, you guys are using some sort of basic program that simply multiplies the previous result by 2. It would be quicker to just write an equation for it then just figure out the billionth result in seconds.[/QUOTE]
The problem is that multiplying the previous result takes too long, and performing the math necessary to calculate the highest possible power of 2 would take the same time (if not a few minutes shorter.)
[QUOTE=Sippeangelo;16027116]Defeats the whole point.[/QUOTE]
How so? Why waste time with regression?
[QUOTE=HiddenMyst;16027572]How so? Why waste time with regression?[/QUOTE]
Why would I want the billionth result when the whole purpose was making a list.
Also, the billionth result would take ages to calculate.
[QUOTE=Sippeangelo;16027659]Also, the billionth result would take ages to calculate.[/QUOTE]
Do you know what an equation is? It wouldn't take any more than a second on a calculator.
Sorry, you need to Log In to post a reply to this thread.
