SilverCreator VS HTML5 GameMaker Speed Test

[Gan]

Hey guys, did some hard core speed tests between SilverCreator and the HTML5 GameMaker.

I think you will like the results. Here's the screen recording:
SC VS HTML5 GM Speed Test Video


Benchmarks:

SilverCreator
300,000 - 1.55 Seconds
3,000,000 - 16.82 Seconds

HTML5 GameMaker
300,000 - 0.001 seconds
3,000,000 - 0.011 seconds
30,000,000 - 0.109 seconds
300,000,000 - 1.342 seconds

Impressed? I sure am.

[Charlo]

Well, in-browser JavaScript engines have been in the process of being optimized for several years by large groups of extremely talented people.  And JavaScript compiles directly into machine code in most browsers (at least in Safari).  SilverCreator is the work of one dude who built a scripting language on top of a programming language (RealBasic) which does not seem to be focused on delivering insanely high speeds.   I'm not surprised by these results at all.   

[Gan]

Way to ruin a moment...

Haha, just kiddin. Yeah, Javascript is fast, though the point of this is to show the speed between the two GMs.
It is an impressive gap, considering with just a small change the user could upgrade from Sc to HTML5 GM and experience an insanely large speed increase in their code.

[Zoo]

Lol, nice sig, gan. Could we do that too, or is it just one of your magical powers?

[Gan]

Magical powers.

If you guys had these powers... it'd be too risky.

But hey, I can insert an HTML5 game into your siggy if you pm the game to me.

[Gan]

Magical powers.

It's too risky for users to have these powers but I can inject HTML5 games into your sigs for ya. Just PM me the game.

[Connors]

I figure almost anything is going to draw sprites faster than SC.

[x]

All you are doing is comparing Javascript to SC. One is highly optimized and often JIT compiled (or AOT compiled), the other is a scripting language made by one dude on top of an already slow compiled language. What a pointless comparison, the only thing more pointless would be comparing C to Ruby, finding Ruby is slower, then saying C is better in every way because its faster--I have actually seen a person do this.

Also, if I am not mistaken, all your "speed-test" does is update a variable hundreds of thousands of times. So all you are testing is variable allocation. The single most unfair thing to test between compiled and interpreted languages. Not to mention completely worthless in an real life scenario. Do something intense, like a program that calculates the n-body problem, or constructs a big binary search tree, then searches it. In addition a more meaningful and interesting comparison would be HTML5 GM vs raw javascript, so you can see how much, or little, performance is lost as an overhead for using the GM.

[Gan]

Well actually... in the video I did to a test using the Sin function. Anywho, this is just to show that the HTML5 GM is fast. Even if it was already obvious...

Edit: And to also give some real numbers as to how fast it is.

[Zoo]

Wow. Thank you. I like this signature.

[EqwanoX]

well, so much for SC. im very impressed, plus, we know its cross platform. and no sound glitch bullshit. now we can make the action games that we couldnt make in sc. someone port over SpaceMarine!

(woah! 87 wins in elements!)

[Zoo]

I couldn't program space marine in regular old SC.

[Circuit]

I kinda want to do it, but I'm busy with school and stuff.  This is probably something that we could all work on together, since it will need to be re-designed somewhat in order to work in HTML5 GM.

[Gan]

I gotta finish some homework right now but later tonight and tomorrow I can start the recode.

If I make progress I'll post a try-it link and the source.

[Zoo]

Wow, an unfinished game, and we're already talking about remaking it. It must be good!