Scratch

I can do it at an avg speed of 1-2.5 fps lol

Last edited by Lucario621 (2009-01-04 19:13:05)

What do you mean stats?

Does my computer have Attack, Defence, Hitpoints and stuff?

Anyway here's my "Stats":
-Microsoft Windows XP Professional Version 2002 Service Pack 3
-CPU (AI intelligence lol) 2.40 GHz
-512 MB of RAM

Helps?

archmage wrote:

Lucario621 wrote:

What do you mean stats?

Does my computer have Attack, Defence, Hitpoints and stuff?

Anyway here's my "Stats":
-Microsoft Windows XP Professional Version 2002 Service Pack 3
-CPU (AI intelligence lol) 2.40 GHz
-512 MB of RAM

Helps?

Hmmm, I have a computer that is about the same as it doesn't go as slow as you describe. It goes about 1.5 - 2 times slower online than offline but it is was faster than 2 frames a second.

Lol yeah I was just kidding. Honestly I'm not sure cuz when I watch I feel a 30min TV show it feels like 10 min, though when my mom makes me go outside for 10 min (which I hate lol), it feels like 30 mins, so I don't have a good sence of time.

Mines pretty rubbish. Takes like half a second for a costume change.

2.8 GHz
448 Ram

Dan01 wrote:

Btw, what I said about the game looking a bit poor quality, that was online. I've just downloaded it and it is great.

Yeah I know, its a bummer. I think that one big issue may be ram consumption. People with low ram are reporting the biggest problems despite their moderately powerful processors. Absolutely nothing I can do about that though, that is the fault of the scratch player.

Ok I made a very slight "optimization" (it seems to make it a bit faster online but logically it shouldn't)

Please, test this and tell me if there is any improvement with the lag issue
http://scratch.mit.edu/projects/archtest/373155

Last edited by archmage (2009-01-06 18:51:11)

Zelda123 wrote:

archmage wrote:

Ok I made a very slight "optimization" (it seems to make it a bit faster online but logically it shouldn't)

Please, test this and tell me if there is any improvement with the lag issue
http://scratch.mit.edu/projects/archtest/373155

It works much better for me than the other one. It can actually be played nicely online.

This is actually kinda ridiculous. I will tell you how exactly I "fixed" it and I am sure you will also find it ridiculous. I did some testing and figured that the set x/y blocks is much faster than the change x/y blocks when played online.

So this is what I did.

First I made 2 variables: xposition and yposition

This is my original code basically

change x by xvelocity
change y by yvelocity


This is the new code

Change yposition by yvelocity
Change xposition by xvelocity

Another script was also added

when flag clicked
forever
go to x:xposition y:yposition


So if you don't understand what I did I will explain. Instead of actually moving the sprite, I changed 2 variables that represent the sprite's x and y position. Then the new script I added set the sprite's x and y position to the variables.

This is ALL I did. Does it make any sense that this made it run faster online? The answer is no, in fact it should have made it slightly slower.

Ridiculous right?

That is rediculous. It still baffles me now!  Next thing you know it, "switch to costume (costume + 1)" will be faster than "next costume" !

Dan01 wrote:

The new version is defentily a lot better.

I don't know whether to feel happy or sad about that. :S

On one hand I finally got it to run faster, however, the change I made should have made it slightly slower instead of faster. It is very hard to optimize scratch scripts because sometimes you don't know what will work, or to optimize it you have to use many repetitive scripts. It promotes very bad coding practices. If people learn how to program this way, their scripts in other languages will be very messed up 

Ok, I am doing a large fix for this which involves splitting up the scripts.
http://scratch.mit.edu/projects/archtest/374231

Please let me know how it preforms compared to this one
http://scratch.mit.edu/projects/archtest/373155

archmage wrote:

Here is another version you should take a look at.
http://scratch.mit.edu/projects/archtest/371623

It should be faster than the other one I posted but it sometimes misses key presses.

EDIT: It doesn't detect key presses in the initial jump animation only from what I can see.

A "SUPERFIX" is in the works. It combines both methods of lag reduction. I think this may work well.

By the way, all testing projects will be uploaded to my archtest account.

Faster, however sometimes Mario gets glitched. When he runs from one side of the stage to the other, the run animation starts well but then he ends up gliding without any costume change.

archmage wrote:

Dan01 wrote:

archmage wrote:

Here is another version you should take a look at.
http://scratch.mit.edu/projects/archtest/371623

It should be faster than the other one I posted but it sometimes misses key presses.

EDIT: It doesn't detect key presses in the initial jump animation only from what I can see.

A "SUPERFIX" is in the works. It combines both methods of lag reduction. I think this may work well.

By the way, all testing projects will be uploaded to my archtest account.

Faster, however sometimes Mario gets glitched. When he runs from one side of the stage to the other, the run animation starts well but then he ends up gliding without any costume change.

That was fixed I think.

That version is outdated. Please try this one
http://scratch.mit.edu/projects/archtest/374231

Works well. Best one yet for me.