Love performance comparison/question.

General discussion about LÖVE, Lua, game development, puns, and unicorns.
Post Reply
Pupnik-
Prole
Posts: 3
Joined: Mon Nov 19, 2012 6:04 am

Love performance comparison/question.

Post by Pupnik- » Mon Nov 19, 2012 6:36 am

Hi, I'm currently making games in as3 with flashpunk, but have been wanting to expand my horizons a little into a new language (and a system that is more network friendly), and love looks like a good option. Lots of people I know constantly rave about how great lua is, so I decided to do some investigating and came across this benchmark:

http://www.sparkrift.com/2012/1/love2d- ... na-vs-sfml

It shows (for this specific benchmark) poor performance for love compared to the other engines. It was indeed the reason that the developer (who had praise for lov's ease of use and development speed) chose not to use it. Is this a fair assessment of love's general performance compared to the others? The code is provided here.

I'm planning to develop a 2d multiplayer shooter/platformer similar to soldat for ~16 players (more would be nice) so performance is a concern for me.

User avatar
kikito
Inner party member
Posts: 3153
Joined: Sat Oct 03, 2009 5:22 pm
Location: Madrid, Spain
Contact:

Re: Love performance comparison.

Post by kikito » Mon Nov 19, 2012 8:52 am

When you buy a car, max speed is not the only thing you look. You also consider the price. How many repair shops are there. The reputation of its makers. If you have children, then you make sure there's space for them in the backseats. And so on.

The same goes for engines and frameworks (I would consider LÖVE a framework, not a game engine).

LÖVE is not the fastest solution out there. But that has never been its objective anyway. Ease of use, for example, comes first.

It's right there in the article:
Porting the benchmark to Love2D was really quick. I think it may have been the easiest to use after XNA.
Take into account that the poster was already quite familiar with XNA. If he had not, LÖVE would have been the easiest one.

But the article doesn't have a graph for ease of use, or platform independence.
When I write def I mean function.

User avatar
substitute541
Party member
Posts: 484
Joined: Fri Aug 24, 2012 9:04 am
Location: Southern Leyte, Visayas, Philippines
Contact:

Re: Love performance comparison/question.

Post by substitute541 » Mon Nov 19, 2012 9:06 am

I prefer Flash (Adobe/Macromedia) than Flashpunk.

And uhh, Love is just an easy-to-use framework, not necessarily a fast one (although, since there is no limit for the FPS, other than your computer, it can be fast for relatively powerful computers).
Currently designing themes for WordPress.

Sometimes lurks around the forum.

User avatar
bartbes
Sex machine
Posts: 4946
Joined: Fri Aug 29, 2008 10:35 am
Location: The Netherlands
Contact:

Re: Love performance comparison/question.

Post by bartbes » Mon Nov 19, 2012 9:30 am

For what it's worth, I ran the test and got way more performance than that. Unfortunately, since he only published the windows ones, I couldn't really test the others myself.

Pupnik-
Prole
Posts: 3
Joined: Mon Nov 19, 2012 6:04 am

Re: Love performance comparison/question.

Post by Pupnik- » Mon Nov 19, 2012 10:21 am

bartbes wrote:For what it's worth, I ran the test and got way more performance than that. Unfortunately, since he only published the windows ones, I couldn't really test the others myself.
Your numbers will depend on how strong your computer is. I checked both love and XNA which were able to hold 60fps at 5000 and 25000 sprites respectively, consistent with the original authors results (under windows 7 64-bit).

User avatar
dreadkillz
Party member
Posts: 223
Joined: Sun Mar 04, 2012 2:04 pm
Location: USA

Re: Love performance comparison/question.

Post by dreadkillz » Mon Nov 19, 2012 2:16 pm

Was this benchmark done with LuaJIT? It probably would help boost the score by a lot but I'd imagine we still end up last.

User avatar
monsieur_h
Citizen
Posts: 65
Joined: Tue Oct 30, 2012 4:43 pm

Re: Love performance comparison/question.

Post by monsieur_h » Mon Nov 19, 2012 3:13 pm

LÖVE also ends last because he didn't bench pygame. :awesome:

Pupnik-
Prole
Posts: 3
Joined: Mon Nov 19, 2012 6:04 am

Re: Love performance comparison/question.

Post by Pupnik- » Mon Nov 19, 2012 3:15 pm

dreadkillz wrote:Was this benchmark done with LuaJIT? It probably would help boost the score by a lot but I'd imagine we still end up last.
I performed it again with slime's LuaJIT Love, and it attained ~9000 sprites, a 1.8x improvement over regular love and closes the gap on XNA significantly.

User avatar
Nsmurf
Party member
Posts: 191
Joined: Fri Jul 27, 2012 1:58 am
Location: West coast.

Re: Love performance comparison/question.

Post by Nsmurf » Mon Nov 19, 2012 5:21 pm

I would also note that for multiplayer, having it be platform independent is AWESOME. That way, you don't need to worry about, say, macs having a different version then windows, or something like that, because they all run the same code.
OBEY!!!
My Blog
UE0gbWUgd2l0aCB0aGUgd29yZCAnSE1TRycgYXMgdGhlIHN1YmplY3Q=

User avatar
Dattorz
Citizen
Posts: 66
Joined: Sat Oct 27, 2012 12:32 am
Contact:

Re: Love performance comparison/question.

Post by Dattorz » Mon Nov 19, 2012 6:50 pm

Given the large number of calls from Lua into C++ routines each frame, we might see another performance improvement once(?) LOVE gets hooked in with LuaJIT's FFI.

Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests