Adding a Particle System Cache — Handmade Hero — Episode Guide

0:05Set the stage for working on the particle system

3:15Blackboard: Sim Regions and a Persistent Particle System

6:43Blackboard: Deterministic f(t) Particle System

9:50Blackboard: f(delta*t) Particle System

10:57Blackboard: The expense of a persistent f(t) particle system

14:50Blackboard: Cached particles

19:57Set the stage for tomorrow

22:38Run the game and see our current situation, before taking a look at the sim_region stuff

25:12Create handmade_particles.h and handmade_particles.cpp, and talk a little about writing the usage code first

27:14handmade_sim_region.cpp: Make BeginSim() assert that we have a Source

28:41handmade_sim_region.cpp: Start to add the calling site code for the particle system to BeginSim()

32:49Blackboard: Shutting down the particle code

36:31handmade_sim_region.cpp: Make BeginSim() activate the particle system upon unpacking

37:56handmade_entity.cpp: Make UpdateAndRenderEntities() initialise the particle_cache and call UpdateAndRenderParticleSystem()

40:54handmade_particles.cpp: Introduce UpdateAndRenderParticleSystem(), GetOrCreateParticleSystem() and TouchParticleSystem()

43:57handmade_world_mode.cpp: Make UpdateAndRenderWorld() rather than UpdateAndRenderEntities() call UpdateAndRenderParticleSystem()

44:48handmade_particles.cpp: Pluralise UpdateAndRenderParticleSystems()

46:07handmade_particles.h: Introduce particle_system and particle_cache structs

48:09handmade_particles.cpp: Implement TouchParticleSystem()

50:14Calculate the number of cache lines required by our scheme

51:51handmade_particles.cpp: Start to implement GetOrCreateParticleSystem()

55:06handmade_particles.cpp and *.h: Introduce InitParticleSystem() and particle_spec struct

56:03handmade_sim_region.cpp: Make it optional to pass a ParticleCache to BeginSim() and do general clean-up

58:13handmade_entity.h: Introduce a version of IsEqual which takes entity_id

58:46handmade_particles.cpp: Slightly flesh out TouchParticleSystem() and UpdateAndRenderParticleSystems()

1:00:36handmade_entity.h: Add particle data to the entity struct and finish bookkeeping

1:02:54handmade_world_mode.cpp: Make PlayWorld() initialise the ParticleCache

1:05:08Run the game and crash and tie up some loose ends

1:07:25handmade_particles.cpp: Introduce InitParticleCache()

1:08:53Run the game, crash in EndSim() and investigate why

1:11:03handmade_particles.cpp: Make UpdateAndRenderParticleSystems() assert that the Cache is the correct size

1:13:58Investigate whether or not we're reclaiming the ParticleCache memory

1:20:03Spot the bug in GameUpdateAndRender() with our needed memory arena being closed

1:23:51todo.txt: Add "Arena upgrade!"

1:25:05handmade_world_mode.cpp: Make PlayWorld() allocate the ParticleCache from the ModeArena for now

1:26:20Run the game and see that it's fine now

1:26:48Q&A

🗩

1:26:48Q&A

🗩

1:26:48Q&A

🗩

1:27:34greenish__ I think you should be searching for MaxFramesSinceTouched instead of MinFramesSinceTouched

🗪

1:27:34greenish__ I think you should be searching for MaxFramesSinceTouched instead of MinFramesSinceTouched

🗪

1:27:34greenish__ I think you should be searching for MaxFramesSinceTouched instead of MinFramesSinceTouched

🗪

1:27:49handmade_particles.cpp: Make GetOrCreateParticleSystem() test on the MaxFramesSinceTouched

1:28:21gene_victor So was the problem that the transient memory vaporized per frame, while the particle cache needs to stay across frame?

🗪

1:28:21gene_victor So was the problem that the transient memory vaporized per frame, while the particle cache needs to stay across frame?

🗪

1:28:21gene_victor So was the problem that the transient memory vaporized per frame, while the particle cache needs to stay across frame?

🗪

1:31:15snoringtortoise Not sure if this has been answered on a previous stream (still trying to catch up). When you introduced the threading model, you mentioned that we would want one thread per core, but since the OS creates threads for us (so helpful) for things like audio, does it matter then that we end up with more threads than cores?

🗪

1:31:15snoringtortoise Not sure if this has been answered on a previous stream (still trying to catch up). When you introduced the threading model, you mentioned that we would want one thread per core, but since the OS creates threads for us (so helpful) for things like audio, does it matter then that we end up with more threads than cores?

🗪

1:31:15snoringtortoise Not sure if this has been answered on a previous stream (still trying to catch up). When you introduced the threading model, you mentioned that we would want one thread per core, but since the OS creates threads for us (so helpful) for things like audio, does it matter then that we end up with more threads than cores?

🗪

1:33:36chibin Do you know if he fully swapped out of emacs?

🗪

1:33:36chibin Do you know if he fully swapped out of emacs?

🗪

1:33:36chibin Do you know if he fully swapped out of emacs?

🗪

1:34:06Demo and plug (in-development) 4coder features

1:41:17Wrap it up

🗩

1:41:17Wrap it up

🗩

1:41:17Wrap 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