Finishing Asynchronous Texture Downloads — Handmade Hero — Episode Guide

0:55Build and run the game and note that our textures don't get downloaded

1:32Set the stage for the day with some thoughts on writing software from scratch and the occasional need to use systems which are black boxes

5:15Debugger: Take a stroll through our new OpenGL code

9:08win32_handmade.cpp: Make Win32InitOpenGL return an HGLRC

12:20Debugger: Step into Win32InitOpenGL and consider creating the OpenGL context outside of the queue

13:46win32_handmade.cpp: Claim the Window as a CS_OWNDC

14:25Debugger: Step into Win32CreateOpenGLContextForWorkerThread

15:20Explain what is happening in the debugger while two threads are initialising OpenGL

19:14win32_handmade.cpp: Assert if wglMakeCurrent fails

19:52Run the game and note that we never hit that assertion

20:33Debugger: Step through Win32AllocateTexture

21:48handmade_asset.cpp: Set FinalizeOperation to FinalizeAsset_Bitmap

22:32win32_handmade.cpp: Put in the pointers to AllocateTexture and DeallocateTexture

23:05Debugger: Step through Win32AllocateTexture and see the game correctly rendering in hardware

26:06"I don't know how that ended up working"^α

26:38Mention the fact that there's no synchronisation between the texture downloading and the rendering, and describe our goal with regard to failure cases

33:16handmade_opengl.cpp: Explain the importance of gamma correction

36:13handmade_opengl.cpp: Explain why there are two separate enables for sRGB

38:20Internet: Consult the OpenGL docs¹^,2

45:04win32_handmade.cpp: typedef these WINAPI wgl_ functions and enable the extensions in Win32InitOpenGL

1:08:48Q&A

🗩

1:08:48Q&A

🗩

1:08:48Q&A

🗩

1:09:10yoonfan Don't you have a race condition according to the C11 and C++11 memory models when you set the texture handle from the worker thread and read it on the rendering thread without any synchronization? I think you need to use an atomic operation or the compiler could mess up the generated code

🗪

1:09:10yoonfan Don't you have a race condition according to the C11 and C++11 memory models when you set the texture handle from the worker thread and read it on the rendering thread without any synchronization? I think you need to use an atomic operation or the compiler could mess up the generated code

🗪

1:09:10yoonfan Don't you have a race condition according to the C11 and C++11 memory models when you set the texture handle from the worker thread and read it on the rendering thread without any synchronization? I think you need to use an atomic operation or the compiler could mess up the generated code

🗪

1:10:43cubercaleb Why don't you render directly to the backbuffer at Molly?

🗪

1:10:43cubercaleb Why don't you render directly to the backbuffer at Molly?

🗪

1:10:43cubercaleb Why don't you render directly to the backbuffer at Molly?

🗪

1:14:20brokenopcode FYI, no you can't change pixel format once it's been set. Need to create a temp window & context, get the wgl*arb function pointers, then destroy context & window and create new window and context with wgl*arb. Very annoying but that's how it works on windows as far as I know

🗪

1:14:20brokenopcode FYI, no you can't change pixel format once it's been set. Need to create a temp window & context, get the wgl*arb function pointers, then destroy context & window and create new window and context with wgl*arb. Very annoying but that's how it works on windows as far as I know

🗪

1:14:20brokenopcode FYI, no you can't change pixel format once it's been set. Need to create a temp window & context, get the wgl*arb function pointers, then destroy context & window and create new window and context with wgl*arb. Very annoying but that's how it works on windows as far as I know

🗪

1:15:09Zamar037 Will we be back to the debug code and finish it up after this, or have we finished it now?

🗪

1:15:09Zamar037 Will we be back to the debug code and finish it up after this, or have we finished it now?

🗪

1:15:09Zamar037 Will we be back to the debug code and finish it up after this, or have we finished it now?

🗪

1:15:23cubercaleb Thanks for the in depth explanation. Will you be going over modern GL later?

🗪

1:15:23cubercaleb Thanks for the in depth explanation. Will you be going over modern GL later?

🗪

1:15:23cubercaleb Thanks for the in depth explanation. Will you be going over modern GL later?

🗪

1:17:15garryjohanson Apparently in DirectX 12 you can select a separate context for different GPUs

🗪

1:17:15garryjohanson Apparently in DirectX 12 you can select a separate context for different GPUs

🗪

1:17:15garryjohanson Apparently in DirectX 12 you can select a separate context for different GPUs

🗪

1:18:10gasto5 Where did you upload the 4coder videos?

🗪

1:18:10gasto5 Where did you upload the 4coder videos?

🗪

1:18:10gasto5 Where did you upload the 4coder videos?

🗪

1:18:33d7samurai Yes, D3D11, too

🗪

1:18:33d7samurai Yes, D3D11, too

🗪

1:18:33d7samurai Yes, D3D11, too

🗪

1:18:58cubercaleb Couldn't you also use a PBO for streaming in textures?

🗪

1:18:58cubercaleb Couldn't you also use a PBO for streaming in textures?

🗪

1:18:58cubercaleb Couldn't you also use a PBO for streaming in textures?

🗪

1:20:35Wrap it up

🗩

1:20:35Wrap it up

🗩

1:20:35Wrap it up

🗩

Keyboard Navigation

Global Keys

Menu toggling

In-Menu and Index Controls

Quotes and References Menus and Index

Quotes, References and Credits Menus

Filter Menu

Filter and Link Menus

Credits Menu