You know, I wanted to write a really long anwser to this, but then I remember what always happens in a "C++ vs C#" thread. I guess it's the same in any language vs another language.Quote:
Hold on, so VB6 isn't just a synonym for VB.NET and VB6 to VB.NET is pretty much like C++ to C++/CLI? Then I indeed mixed something up and just failed epically.
Can pax make a real sro private server?
08/11/2010 15:35
Windrius#61
08/11/2010 15:49
lesderid#62
I have a solution for the non-existing page bug:
[Only registered and activated users can see links. Click Here To Register...]
[Only registered and activated users can see links. Click Here To Register...]
08/11/2010 16:05
Mikawa#63
Oh yeah that will help us so much.Quote:
[Only registered and activated users can see links. Click Here To Register...]
08/11/2010 16:10
jM3#64
No, VB6 is a deprecated unsupported language. Visual Basic - Wikipedia, the free encyclopedia. It's almost never used anymore except in legacy applications.Quote:
C++ is a highly supported, very much widely used language that's still undergoing improvements C++ Technical Report 1 - Wikipedia, the free encyclopedia.
In that respect, your comparison is inherently flawed. Also, C++ was not designed to be managed, and this is where its power comes from, combining the capabilities of OOP while simultaneously being as close to the hardware as you care to get. This is why both C++/CLI and C# are nowhere near C++ in popularity.
No, performance is definitely a big factor there as well. Not even Java is as portable as C++, but the huge performance increase with what you can do is a definite plus.Quote:
If I may criticize you as well, I wouldn't write an emulator in C#. Rather I would use a native programming language like C++. Not because the performance is better, but because it would be possible to portable the code to other platforms.
08/11/2010 16:48
ms#65
Yes, I know that VB is certainly not as popular (if not dead) as C++ due to the reasons you already mentioned. What I wanted to say with my comparison is that C++/CLI is Microsoft's attempt of trying to make the C++ programming language ".NET-ready".Quote:
C++ is a highly supported, very much widely used language that's still undergoing improvements C++ Technical Report 1 - Wikipedia, the free encyclopedia.
In that respect, your comparison is inherently flawed. Also, C++ was not designed to be managed, and this is where its power comes from, combining the capabilities of OOP while simultaneously being as close to the hardware as you care to get. This is why both C++/CLI and C# are nowhere near C++ in popularity.
Certainly, a proper written C++ program will always overtake a proper written C# programs in terms of performance. Just the fact that managed code is being "translated" into machine code at runtime underlines that.Quote:
The question is not if managed code uses more resources, but if the performance difference is as significant as you mentioned. Managed code also has advantages like the eliminating of memory leaks.
08/11/2010 17:21
jM3#66
Even if your managed code is compiled into machine code. While things like language-enforced GC make traditional memory leaks go away, it does introduce new memory leaks because it's a leaky abstraction. I've seen many a C# and Java application that had huge memory leaks -- because not calling a very specific API in many cases results in the GC keeping objects marked as in-use even though the code that used them has long since renounced any need to do so. It's not a perfect system but it does let people program without worrying about issues like that so development is easier.Quote:
The question is not if managed code uses more resources, but if the performance difference is as significant as you mentioned. Managed code also has advantages like the eliminating of memory leaks.
But when you want performance, you don't want to go to a heap manager for everything, and you do not want your objects collected by a garbage collector. Even as raw machine code, needing to jump through hoops like this results in severe performance loss. An excellent C# or Java compiler may be smart enough to optimize a locally used object to be allocated on the stack, but in almost all cases they don't since this inherently breaks the ability to guarantee stack overflows and stack corruption don't happen, and it makes it impossible to do accounting on the objects. Heap allocations are SLOW, if you have 20,000 people connected to your server and you need to do 10 calls per each message sent/received per client, and each is sending about 1 thing each second, you're talking about 200,000 calls to a heap manager every second. That's a significant amount of CPU time.
Consider this program:
Code:
#include <stdio.h>
#include <Windows.h>
void foo(){
int* p = new int;
*p = 5;
delete p;
}
void bar(){
int p = 5;
}
int main(int argc, char** argv){
DWORD x = GetTickCount();
for(int i=0;i<200000;i++){
foo();
}
DWORD y = GetTickCount();
printf("new/delete took %d msec!\n",y-x);
x = GetTickCount();
for(int i=0;i<200000;i++){
bar();
}
y = GetTickCount();
printf("stack took %d msec!\n",y-x);
return 0;
}
Code:
new/delete took 203 msec! stack took 0 msec!
You can do things in C/C++ easily like pooling allocated memory into an allocator that can use a constant-time return and leak checks, even a ptr table and ref counts to fix leaks automatically, at the cost of performance, so you get as close to stack allocation as possible while still having managed memory objects, with the same new/delete style interface. If you can do that at C# or Java at all, it requires being very clever to force the language/compiler to do what you want. We also have common idioms to fix things like this, called ref counted smart pointers which let you wrap a persistent object, likely allocated from some pool, in a stack-based object that respects scope and exception safety which will automatically call your pool-deallocator on the object when scope is up or an exception is dispatched, interrupting the flow of execution, based on a reference count. This is a beautiful solution and is applicable in most situations. As always, if it's not threadsafe, it's VERY fast, but with thread safety comes mutual exclusion or live waiting for lock-free designs, in any case you lose performance in general.
It's inherent in how your language works that makes it less efficient, it's not about the code you write. Algorithmic efficiency is a very important factor, but assuming everyone uses the same efficient algorithms, it comes down to language artifacts and inefficacies like this.
08/11/2010 19:38
2F@st#67
its a threesome fight now vb6 vs c++ vs C# :D :awesome:Quote:
08/11/2010 19:58
ms#68
No, it's a fight about how much managed code sucks when it comes to raw performance. Everyone agrees that C++ is faster than C#. And if VB6 really doesn't support OOP (I have no idea) it's impossible to write larger pieces of code without losing track of it.Quote:
08/11/2010 20:23
npcdoom#69
just my 2cents in simple terms C++, C put the burden on the programmer capabilities, C# in the other hand on the VM =P and VB6 is just a waste of everyone time :D
Still using algorithm and good programming practices are more important than the language itself.
Still using algorithm and good programming practices are more important than the language itself.
08/12/2010 01:51
ninov#70
Couldn't understand a thing. Trying to be good in englisch is one thing, rambling nonsense is another.Quote:
but,the main wired,unusual situation is if it is very Perparatory Code Lines.
Those Perparatory Code Lines Done this Big goodwill reputation,So what if those codes are Fully Coded and developed By same Lang educated person Or whom have the knowledgeand a brain Skills !! .
Is doubts are reason To Force a project To be Shut Down ! Nop just some one needed a reason To stop what he have just started.
The Only Lost is at Our community,we lost it based On doubts and Wrong conclusions Based On Wrong reasons wich gained more Value than it deserve.
Why Can not back Online and Ignore those reason that has been tooked as a reason to Force yourself To STOP !!! <<donn comment at it just posted those thoughts Cuz i have been Burned deeply from inside by this complicated tragedy.
after all again we will not have next generation Of something without a seed to develop it.
it's about what if i like Win.Vista Or Win 7, we wont have win 7 without the fail Of Win Vista (some Ver.)
after all fail is just a big start for some matter's if belived in finding true reasons that keep those bad result as failed,,all time long change the reason see results,,,
08/12/2010 02:38
rEqueriUm#71
Usually i just reed stuff like this and make no posts but everytime i go read popx's post gives me a major headache.
/ontopic
I suggest clever guys that posted here useful stuff just ignore stupid threads like this,you said it once why write it all over again?
In my opinion you guys should team up (pax excluded,cuz so far every good programmer on this forum says that u just took of someone else code and upped it a bit)
and make a 99% working emu in months as they did for WoW,Aion etc...
it's all about teamwork guys,don't be greedy :P
/ontopic
I suggest clever guys that posted here useful stuff just ignore stupid threads like this,you said it once why write it all over again?
In my opinion you guys should team up (pax excluded,cuz so far every good programmer on this forum says that u just took of someone else code and upped it a bit)
and make a 99% working emu in months as they did for WoW,Aion etc...
it's all about teamwork guys,don't be greedy :P
08/12/2010 02:48
popx#72
post is removedQuote:
no more headache.Quote:
/ontopic
I suggest clever guys that posted here useful stuff just ignore stupid threads like this,you said it once why write it all over again?
In my opinion you guys should team up (pax excluded,cuz so far every good programmer on this forum says that u just took of someone else code and upped it a bit)
and make a 99% working emu in months as they did for WoW,Aion etc...
it's all about teamwork guys,don't be greedy :P
Note :- ignorance is bad side,try not to stay there too much
08/12/2010 04:18
-Chernobyl#73
It's already been done. How do you think people like Pax get the source code to sell?Quote:
/ontopic
I suggest clever guys that posted here useful stuff just ignore stupid threads like this,you said it once why write it all over again?
In my opinion you guys should team up (pax excluded,cuz so far every good programmer on this forum says that u just took of someone else code and upped it a bit)
and make a 99% working emu in months as they did for WoW,Aion etc...
it's all about teamwork guys,don't be greedy :P
SREMU was an open source project for anybody to download, mess with the source code, and add more functionality. Anybody that's been around for the last 2 or more years already knows this. What killed it was 2 sets of people.
First was all the noobs that couldn't figure out how to compile the code. They constantly flooded the SREMU page with idiotic questions (sort of like the ones constantly being asked here). Second were the people that started popping up claiming to have their own servers based on the SREMU code. Some of them (like Pax for instance) tried to sell these unfinished emulators for profit, claiming them as their own work. There were always some fools that would buy these because they thought they were the best there was available (just look at the join dates of all the noobs that defend Pax).
I've been around long enough and have messed with every emulator that's been released to know that this emulator IS NOT the best that's been released. All the little Pax boot kissers need to spend a little time researching Silkroad emulation before defending a scammer.
You are right about the useless topics. The Pax debate needs to die just like the DB7 debate died some time ago
08/12/2010 14:32
rEqueriUm#74
Well a lot of talented guys on this forum are doing solo project,or atleast they are saying so.Quote:
SREMU was an open source project for anybody to download, mess with the source code, and add more functionality. Anybody that's been around for the last 2 or more years already knows this. What killed it was 2 sets of people.
First was all the noobs that couldn't figure out how to compile the code. They constantly flooded the SREMU page with idiotic questions (sort of like the ones constantly being asked here). Second were the people that started popping up claiming to have their own servers based on the SREMU code. Some of them (like Pax for instance) tried to sell these unfinished emulators for profit, claiming them as their own work. There were always some fools that would buy these because they thought they were the best there was available (just look at the join dates of all the noobs that defend Pax).
I've been around long enough and have messed with every emulator that's been released to know that this emulator IS NOT the best that's been released. All the little Pax boot kissers need to spend a little time researching Silkroad emulation before defending a scammer.
You are right about the useless topics. The Pax debate needs to die just like the DB7 debate died some time ago
Only team-project i was aware was eSro but it "stopped" when jMerlin left(?) now they are doing again a closed beta and it's a good idea.
From the videos i saw eSro so far is best known emu outhere.
(My opinion i don't want now any pax-lovers to flame me pls)
On their official site they said they'll upload new videos of closed beta,i am sure that when they do that will motivate a lot of players including me to help them with some donations.
My hopes lie that someday soon someone will make a working emulator and eSro is the first thing that poops my mind :D
That's all from me.I'll continue to my silent 'stalking' now.
cya Strider
08/15/2010 13:46
pwner96#75
eSRO isnt an emu, its a server based on an emu, so they wont make an emu for public