The exe doesnt work. I tried it at home and I tried it at the library. Its broke or something. It gives me an error. Thats the point. I somehow expected that to be a global thing but perhaps its just here. Hm...

Have you tried to install the other exe that comes with it, oalinst.exe? It will crash if OpenAL is not installed (well, even if it has no sound).
Also, Intel integrated card? I found that it loves to crash when it meets one of those, and it's strange as it runs OpenGL2.0 without shaders and anything remotely modern.

Ive been working on a game for a little while now but Ive been thinking about giving up. not because I don't have the skills to develop it, but because my friends no longer want to hang with me because I talk about my game all the time.

Not only that but They enjoy playing games, they just play them and don't think about it. Because I design games, I automatically analyze the game, and try to understand the design choices made while playing it.

So my Friends wont hang with me because they fear that ill bore them with technical stuff they don't care about.

Anyone else know what I mean?

I did this "back then"... it was really stupid. Forcing some topic you like on someone who is completely uninterested is selfish, boring, and will make people avoid you.
Not because they don't care about games, but because YOU don't care about them, always speaking of whatever crosses your mind like in a monologue with yourself.

So, don't stop making games, but try to undestand what others like to hear... and get to know someone who has your same interests!

On a side note, I sometimes like to force something technical on people, mostly when they are like "wow you make games, why don't you work instead".

Meth(I meant to say that)...

I'd rather like to see hexels or tryxels or oxels(Ya know, the li'l animals) or sperels.

Nah, just telling the an unlie. I think this engine has a huuuuge potential...Wait, this is a cavestory ripoff isnt it!? 

(Remember, don't feed the trolls.)

ok, actually, if it were an exe, Id try it but... sorry.

Now that's an insightful post 

@Flops Thanks! I know, I got it to run once on other Macs... it isn't just as easy because it looks like OSX stores absolute paths INSIDE dynamic libraries (wtf) so I can't just copy my libraries in the bundle and expect them to work, it needs a bit more massaging.
It's a simple demo, so I didn't bother...

@Tanner AMD? I have a pretty un-decent Nvidia 8600 and it always runs capped at 60 fps, but I found that it has some issues with amd or older windows versions.

In the "game" you fly around and you shoot things, simple as that.
The "sun" in the demo procedurally creates voxel blobs and enemies, (the voxel part being much like minecraft), and spawns them around itself. They are streamed, so a bit of hiccups should be expected... while 5 fps is totally uncool.

It's sure crowded around here 

Well, actually it isn't even a game, it is a tech demo for the new 3D capabilities of my game engine.
But I'm quite satisfied with it, so I'm here, should it become something more.

After all Voxel games and artsy games are all the rage today, so that would be the perfect mix 

Move: WASD
Turn and lock enemies: left click and drag
Blow up things: Space
Download: here (for windows)

Actually it was developed for Mac OS X but XCode refuses to ship my frameworks, so, no executable 

Bye 

Naumachia used to be a wildly awesome space arena shooter, but then it went to the great plains where the great games killed by real life gather.

Dynamic resource Loading & Linking:

Code:

//purge preload group
removeSubgroup( preloadGroup );
delete preloadGroup;
preloadGroup = NULL;

addSubgroup( lateGroup );

Some duplicate resources in preloadGroup are loaded atomically (but they are few) and used for an intro; meanwhile lategroup loads all the level resources in a background thread.
Then when the user exits the intro, the bindings are swapped and the game starts to use the new resources.

For an iPhone game  

Code:

queueSemaphore.wait();

//awoken to be killed (oh the irony)
if( !running )
break;

I lolled SO hard  

tl;dr: Gaming is only fun if I'm in the right state of mind.

I'm lucky then that when I play, I play just as before  

I won't touch a game for a month, but then if something attracts my attention, I can play it as if it's the most important thing out there... just like Minecraft, or Halo Reach, or Plants Vs Zombies, or Fallout 3, or Forza 3, or whatever random games get me "in".

As a rule of thumb, rarely something that got me hyped also got me hooked. Hype is SO like a moral obligation to like the game, and then I just don't like it. If you don't need to play a game, don't just buy it because it's new.
I did it for some time leaving around tons of unplayed games... then I realized that I just needed to play what I wanted.
For example I really had a blast when finishing Mario 64 again (for the 30th time?) last week. It was just good as always.

imo a bit of this can be traced back to the insane (and unique to games) certainty that only what is new is worth to be played.
Any other media has countless age-old-but-still-actual masterpieces, but we gamers are tricked to think that we don't.

Already in, it's cool... and pretty deserted too.
I would say that it's more interesting as a Twitter replacement than a Facebook replacement.
it just copies Facebook adding the circles quirk, that's at best non-essential for most facebook users.
On the other hand, it features a follower relationship model, rich media features, and integration with google services, that are most likely to appeal more to those who use Twitter  

There still are some bad design choices though.
a) Being spammed by replies of followed persons' posts is rather annoyong as honestly I don't give a fuck of what 2236 guys think about Minecraft 1.8, and twitter got this right by hiding unfollowed persons' replies.

b) a reply puts the original post at the top of my stream, so posts from celebrities always clog my stream being constantly at the top pushing down friend's posts... and because of the irrelevance (to me) of the replies, it is really annoying down to the unusable.

I hope they fix this anytime soon, or I will just stick to Twitter+ FB.

Unfortunately I've just stopped playing games.
Which is not because I don't have time, really I have no schedules at all - but just because the other things keep my mind busy enough.

Before you give in to your assumptions you should give a test exe to someone to confirm whether that's true or not.

Well, I did. The other PC still runs at 30 FPS, and my own just passed from 4.5 ms to 0.7 for no actual reason.
I guess I will have to test some AMD cards and see how they behave...

EDIT: as I expected, this review puts the HD5470 at the same performance levels of the 8600 GT...
so a 28 times worst performance has to be caused by a bug somewhere.

Uhm, I tried on another PC after those optimizations, and it went first at 4.5 ms, then 0.7 like on OS X.

I was ready to swear that an HD5470 (even if Mobile) is better than an Nvidia 8600 being both low end and the latter being much older... instead I can't really get the HD5470 to run decently.

Neither I can update my December 2009 drivers, it won't recognize reference ones... damn you Acer.
I guess I'll be going to blame it on the driver  

Thanks that was really useful, while not resolutive

I did:
-set the color format to 32 bit
-disable the alpha buffer
-set 16 bit depth
-set PFD_SWAP_EXCHANGE
-made sure that VSync is disabled
-made sure that Swapbuffer is only called once

Anyway, if I look at empty space the rendering time drops to 3 ms... this is still quite much, but substantially less than 28 ms, as one would expect.
So I guess that the real culprit isn't DrvPresentBuffers, that just locks the command queue and waits for the render to end, but the slow render itself.

Then, why the render is slow on Windows and fast on OS X 
I think I will have to do as mcc suggests, just turning off things one by one... when I will dare to accept such a challenge 

The data sizes are whatever you compile as.  This is one reason why 32 bit programs typically need some kind of compatibility layer to run on 64 bit operating systems.

Back when Apple started using the 64 bit G5 chips, they introduced a 'fat binary' format that actually contained both 32 and 64 bit binaries, and the OS would select the correct one.  I think this still existed in some form after the Intel switch, but since all Macs are 64 bit now, it's probably going to be obsolete soon.

Actually it isn't at all, every single IOS library is a fat binary linked for i386 and ARM.
In fact, the iPhone Simulator is at his core an huge custom /usr/lib directory full of i386 IOS libs.
This could also come handy if they are going to switch from Intel to ARM.

Why is there a delay function getting 22% of runtime?

That's... interesting. At least interesting as those D3D calls.
I even disabled VSync, and I know it worked because reenabling it causes it to go even slower.

EDIT: to test for CPU fallback, I have max 25% CPU use when running the game, which is expected as there is heavy tesselation; it is just like OSX use.
Also I controlled the driver Vendor, and I'm correctly using the Nvidia/Ati driver on both the computers where I tried.
 

Uhm, I grabbed VerySleepy from the Interwebs and here are the top time wasters:



The heaviest method belonging to the game accounts for a mere 4%  
I'm afraid that the buffer swap is so slow because it runs in a window... time to go fullscreen?

I suspected something like this, as the frame time is "only" of 4 ms if I look at empty space.
I am not using shaders, I'm just using VBOs with multitexturing... maybe the multitexturing is the culprit?

Maybe some profiling will help here.

Hi all,
I've made a quite complex demo with my framework (for a course project) and on Mac OS X it ran like a charm, with a max frame time of 1 ms, meaning a lot of FPS, somewhere around 1000.

Then I reboot my machine on Windows and the frame time jumps to 28 ms with a worst case of 98 ms, meaning an average of 30 FPS with dips to 10 or less.
Yeah, the OSX is actually an hackintosh with a tacked on driver, still IT IS 28 TIMES FASTER.

I have the last drivers, Windows 7 x64, an NVidia 8600 GT.
I don't really know how to pin this one on something, some general directions?  

I'm curious to see how other people on this board handle game entities with subclasses.  For example, I have an Entity class that contains the subclasses Item and Actor.  I started out with a single linked list, "entities", and iterated over it once during each update.  When I needed actors and items alone, I would call a function to pull them out of the "entities" array.  Since then, I have found it easier to maintain separate linked lists, "actors" and "items".  I iterate over each separately, and dump them into one master list whenever I need to treat them as entities.

Obviously, these are just two sides of the same coin, but I'd be interested to know which method you guys use more often.

Uhm... usually when you make a class that owns a Entity* array, you DON'T want that class to know the final type of the managed entities.
It is a good form of encapsulation.
If you need to have different behaviours, let the entities themselves to manage it, and use polymorphism.

If polymorphism is not going to cut it because you need too much detail and there isn't a common interface between Item and Actor... just store two lists. And I would ask myself which is the reason behind an Entity class, if Item and Actor are so different.