cinera_handmade.network/cmuratori/hero/code/code111.hmml

43 lines
3.7 KiB
Plaintext

[video member=cmuratori stream_platform=twitch stream_username=handmade_hero project=code title="Resolution-Independent Ground Chunks" vod_platform=youtube id=qDE7_6K2pyY annotator=Miblo annotator=dspecht]
[0:54][Make sure Ignored are unique names for Miblo]
[2:17][Pull the debug camera back a little further]
[4:09][Up our tree count]
[4:32][Start making the game very slow][quote 82]
[5:41][Check out all the trees and forecast today's work]
[9:10][Look at the ground]
[11:20][Re-enable this GroundBuffer routine]
[12:26][Give PushBitmap a scale, take a look and note that the Basis isn't being cleared]
[13:33][Change the Basis back to the Default]
[15:05][#if 0 the GroundBuffer routine and add visualisation for the places where GroundChunks are being rendered]
[17:07][See those places and note that it's never updating where the GroundChunks are]
[17:55][Fix that update and take a walk through the world]
[19:28][Figure out why we're not getting anything in the bitmap]
[21:36][Get the outlines drawing in the right place]
[23:38][Get FillGroundChunk working]
[26:59][Give DEBUGLoadBMP the ability to let us ask for the default alignment]
[29:38][Continue working on FillGroundChunk]
[31:15][See if that produces anything more usable for us]
[33:11][Verify that the problem is in the Bitmap]
[33:47][Try and step into FillGroundChunk]
[34:58][Moment of realisation: Visual Studio is too dumb to set that breakpoint when building in -O2]
[36:19][Initialise Buffer->WidthOverHeight]
[37:42][Look at that frame rate]
[43:13][Only render ground layers that are close to the camera]
[45:54][Consider optimising some of the renderer]
[49:20][Q&A]
[49:47][@stelar7][There are 15 mins left of the stream, since you went over in the prestream]
[50:22][@d7samurai][No stream tomorrow, right?]
[50:26][@grumpygiant256][How are you going to handle the ground chunks on the further layers which disappear while still on-screen? Will you need to expand the visibility range based on layer distance?]
[52:16][@ttbjm][Some elements in the ground are being repeated now for some reason]
[52:54][@corpseloleu][I'm afraid of using sorting and stuff when making games 'cause I'm afraid that if I get too many items it will slow down. Or shall I not worry about that?]
[54:09][@miblo][I think it'd be cool to see really far down through, for example, floors made of glass / ice]
[55:04][@ifingerbangedurcat][Will the memory footprint keep getting bigger? There's so much raw pixel data being stored right now, is this a concern at all, will we be compressing data to save ram? I imagine not, but it still feels like we're using too much memory for just a 2D game. Correct me if and because I'm wrong]
[58:01][@waterlimon][Are the ground chunks stored in some data structure (e.g. hash map or grid) or are they recalculated each frame (I was wondering, because ground chunk knows its position, but wouldn't that be implicit in its position in the data structure)?]
[58:39][@ttbjm][There was horizontal repetition of some splats but not all of them]
[59:16][@insofaras][Do you think you are roughly 1/5 of the way through completing the game at this point?]
[1:01:04][@robrobby][hen all RAM (even on GPU) is used, when the game is put into background (switching a short time to another application) all the memory has to be swapped in and out, right, even to the GPU? Or what strategy is used to prevent swapping in and out of big memory chunks?]
[1:02:25][@garlandobloom][Q: Are we rotating and scaling the splats on the ground chunks to get the maximum variability?]
[1:03:09][@soysaucethekid][I noticed the hero's head would clip through the trees (i.e. trees on the top part of the room) on occasion and then it would fix itself. What's causing that?]
[1:03:55][Close this down]
[/video]