[Development] 4267 conquer server.

Quote:

Originally Posted by ChingChong23 @Viscount S: i understand that you guys wish to keep memory usage low, but you wouldn't even sacrifice a couple of mb's at max (and not even that much if done correctly) of memory for a much better structure? Memory in this situation should not even be considered. For example you saying Millions of Dialog instances can be created, while that is possible if someone decided to spam a million packets to talk to an NPC, and that's even if you have not handled it to even have a delay or check if he's already interacting with an NPC, then yes. But that's not going to happen? @CptSky: sorry, i don't know much about .net and/or it's collections but you get the point.

switch(VAL)
{
goto case<VAL>;
 case 0:
 // code gets executed.
 goto END;
 case 1:
 //code gets executed
 goto END;
}
END:

Quote:

Originally Posted by ChingChong23 @Viscount S: i understand that you guys wish to keep memory usage low, but you wouldn't even sacrifice a couple of mb's at max (and not even that much if done correctly) of memory for a much better structure? Memory in this situation should not even be considered. For example you saying Millions of Dialog instances can be created, while that is possible if someone decided to spam a million packets to talk to an NPC, and that's even if you have not handled it to even have a delay or check if he's already interacting with an NPC, then yes. But that's not going to happen? @CptSky: sorry, i don't know much about .net and/or it's collections but you get the point.

Quote:

Originally Posted by ChingChong23 @Viscount S: i understand that you guys wish to keep memory usage low, but you wouldn't even sacrifice a couple of mb's at max (and not even that much if done correctly) of memory for a much better structure? Memory in this situation should not even be considered. For example you saying Millions of Dialog instances can be created, while that is possible if someone decided to spam a million packets to talk to an NPC, and that's even if you have not handled it to even have a delay or check if he's already interacting with an NPC, then yes. But that's not going to happen?

// Hideously slow program! Can you spot the object creation?
public static void main(String[] args) {
    Long sum = 0L;
    for (long i = 0; i <= Integer.MAX_VALUE; i++) {
        sum += i;
    }
    System.out.println(sum);
}

Quote:

Originally Posted by Nullable I get fussy about memory issues. Just because the GC handles memory compaction, that means that we should not care about memory and performance. Also a switch statement isn't a series of if elses, it is more of a goto table: Code: switch(VAL) { goto case<VAL>; case 0: // code gets executed. goto END; case 1: //code gets executed goto END; } END:

Quote:

Originally Posted by ImmuneOne Clear progress and plans on CPX; Is the source ready yet? We're basicly far enough to launch a beta, we're just fixing up little things. How long will the beta last? Till the summer. What will happen then? We will switch to native C++ instead. How long will that take? We have already done the dataserver and the defender. I'm currently working on the webserver with own scripting language; Current progress: [Script]  Spoiler string Text = ""; string ImagePath = "C:\CPW\Root\img\Logo.png"; // You always have to seperate the definitions and functions with a line of eachother. Begin: { RenderImage(ImagePath, 4, 1, 12); //Target, Left,Right,Top RenderText(Text, 4, 1, 12); //Target, Left,Right,Top //The text will be on top of the image (They go in order). } End: Only thing left is the source.

Quote:

Originally Posted by Viscount S I agree, there are better ways to approach the handling of NPCs, but the memory footprint of each class and its counterparts can play a key role in a server's stability. I will cede that memory usage shouldn't be an issue at the moment (because we all know premature optimization is the root of all evil), and sometimes object creation is more appropriate to create objects to enhance clarity or for simplicity. But it can also be expensive, which is the point I am trying to convey. My point is not memory usage or the garbage collector, but it can effect speed drastically. I assume you have a background in Java, therefore I'll provide this example from Effective Java by Joshua Bloch: Item 5: Avoid creating unnecessary objects Code: // Hideously slow program! Can you spot the object creation? public static void main(String[] args) { Long sum = 0L; for (long i = 0; i <= Integer.MAX_VALUE; i++) { sum += i; } System.out.println(sum); } "Changing the declaration of sum from Long to long reduces the runtime from 43 seconds to 6.8 seconds on my machine." It is considered good practice to treat your code as if it were a public API, especially now that you have released it. You don't know how that code will be used, anything could happen, even if it's accidental. The code above could be chalked up to a typographical error.

long curTime = System.currentTimeMillis();
	    Long sum = 0L;
	    for (long i = 0; i <= 5000; i++) {
	        sum += i;
	    }
	    System.out.println("Finished: " + (System.currentTimeMillis() - curTime));

Quote:

Originally Posted by ChingChong23 I completely understand where your coming from, since I've proven memory is not an issue ill go on speed. I would not suggest using this if it was going to get called several thousands of times a second, but lets estimate how many of these objects would be created on a server been online for a week. ill estimate 1000 NPC interactions (and yes this is more than people would do) lets say they click 3 options, theres 3k dialog instances created. On my computer running this code Code: long curTime = System.currentTimeMillis(); Long sum = 0L; for (long i = 0; i <= 5000; i++) { sum += i; } System.out.println("Finished: " + (System.currentTimeMillis() - curTime)); Running this code takes 1millisecond, running it using long takes 0 milliseconds. Your not willing to drop a few milliseconds over your entire server time for a better structure? I also like the amount of knowledge in this topic, don't see much of it on epvp.

Quote:

Originally Posted by ImmuneOne Clear progress and plans on CPX; Is the source ready yet? We're basicly far enough to launch a beta, we're just fixing up little things. How long will the beta last? Till the summer. What will happen then? We will switch to native C++ instead. How long will that take? We have already done the dataserver and the defender. I'm currently working on the webserver with own scripting language; Current progress: [Script]  Spoiler string Text = ""; string ImagePath = "C:\CPW\Root\img\Logo.png"; // You always have to seperate the definitions and functions with a line of eachother. Begin: { RenderImage(ImagePath, 4, 1, 12); //Target, Left,Right,Top RenderText(Text, 4, 1, 12); //Target, Left,Right,Top //The text will be on top of the image (They go in order). } End: Only thing left is the source.

Quote:

Running this code takes 1millisecond, running it using long takes 0 milliseconds. Your not willing to drop a few milliseconds over your entire server time for a better structure? I also like the amount of knowledge in this topic, don't see much of it on epvp.

Quote:

Originally Posted by ImFlamedCOD ...