Jump to content

  • Log In with Google      Sign In   
  • Create Account

chbfiv

Member Since 10 Apr 2001
Offline Last Active Apr 07 2012 01:33 PM

Topics I've Started

Fast GPU OpenGL ES 2 VBO update for physics soft bodies on mobile?

31 March 2012 - 04:08 PM

I would like to use soft bodies for some rope simulation on the mobile platforms, but I'm trying to grasp the best method to update my soft bodies every frame. Do I have options outside of unloading and loading a new vertex array/indices/faces each frame?

Different results when viewing on laptop vs desktop (w/ detials)

21 April 2011 - 12:42 PM

We are using XNA (custom rendering solution, not the one provided by default w/ XNA).

Both computers are Windows 7, w/ DirectX 9.0c, Shader 3.0
Desktop is x64 and Nvidia Geforce GTX 570
Laptop is x32 and Nvidia Quadro FX 1800M

The main problem is that we export our models w/ channel 2 AO passes; they arn't showing up on the laptop but they are (and have been) on the desktop. After talking with the team who developed the render engine they seem to think it's something in the shader that isn't supported by the laptop. From what they explained, the blending for channel 2 was happening at the shader level, not using TextureBlendStages at the api level.

The first thing I looked into was the MaxTextureBlendStages and MaxSimultaneousTextures in the DXSDK Caps viewer and I found that both the desktop and the laptop have the same setting of 8 for both. W/ that I'm out of obvious options as to understand why the shader wouldnt blend the same on both adapters at the shader level w/ the same shader version.

Also, a second more minor issue we found was that the laptop was exaggerating the specularity significantly more than on the desktop. It was easy enough to dial down for our purposes.

c Programming - Pass int * to func, return next int *?

20 July 2010 - 08:26 AM

I was doing something as simple as

int * nextToken(int * currentToken) {
return currentToken++;
}


it however doesn't change the int array pointer outside the func scope. It seems like I should be passing a pointer to the pointer for the current token, deref the current token, incrementing the curr token, then returning the deref token?

Which is more correct? C# implementation

09 July 2008 - 08:27 AM

I'm just looking for some feedback. Alternatives would also be interesting! I have a base class with a Size and Location. I also have a class that inherits from that base, but it doesn't care about the Size and Location being logged to Console. Of the two options below, which is more correct way to prevent the Console.WriteLine? Size and Location var's must remain on the base class, and the Console.WriteLine must happen for all classes except SomeFoo if you choose to write a 3rd option=) .NET 2.0 1:
    public class FooBase
    {
        Size mySize = new Size();
        Point myLocation = new Point();

        
        protected virtual void CallMeSomeTimeAfterMyClassIsConstructed()
        {   
            if (myLocation.IsEmpty && mySize.IsEmpty)
            {
                Console.WriteLine("Size and Location are empty.");
            }
        }
    }

    public class SomeFoo : FooBase
    {
        protected override void CallMeSomeTimeAfterMyClassIsConstructed()
        {
            
        }
    }






2:
    public class FooBase
    {
        protected bool hasSizeAndLocation = true;
        Size mySize = new Size();
        Point myLocation = new Point();

        private void CallMeSomeTimeAfterMyClassIsConstructed()
        {
            if (hasSizeAndLocation && myLocation.IsEmpty && mySize.IsEmpty)
            {
                Console.WriteLine("Size and Location are empty.");
            }
        }
    }

    public class SomeFoo : FooBase
    {
        public SomeFoo()
        {
            this.hasSizeAndLocation = false;
        }

    }








.NET Dataset.Table.Select() unreliable?

20 November 2007 - 02:01 AM

So I have a Dataset, and use it allot during runtime. Randomly somewhere in the process of selecting DataRows it'll decide that the content isn't there this time, or that the Table itself doesn't exist this time(Exceptioning out), but will work fine on the next Select happening ms' later, 500+/- or more times before or after that problematic Select. I can only assume that the data is corrupted during a read? or the GC is adjusting memory locations? I really am clueless atm as to why. So any ideas on what would cause this are welcome! Bourke Edit: Also as a quick example. I'll select a certain Column with a message value MyTable.Tables["Foo"].Select("Message='TEST'"); works fine, then I'll select more random Messages and then Select the first one again MyTable.Tables["Foo"].Select("Message='TEST'"); but the Select will return a empty DataRow object or, Exception out that the Table isnt ref to an obj. then it'll work again a ton more times, randomly doing the same as above for other messages....

PARTNERS