A fast 2D and 3D game engine — TypeScript SDK, C++/WebAssembly core, visual editor. Ship one project to web, desktop, WeChat MiniGames, playable ads, and native Android / iOS.
TypeScript 2d-game-engine3d3d-game-engineandroidcppecsgame-editorgame-enginegamedeviostypescriptwebassemblywebglwebgpuwechat-minigame
Commits on master
perf(gi): the medium level runs the dynamic GI its cheaper way - half-resolution gather, a quarter of a probe's rays a frame, cards over eight frames, two thirds of the probe rows; 7.3 -> 4.5 ms on a Snapdragon 8 Elite
447c76c yhh committed 8 days ago
fix(worldgen): a chunk of decoration or grass is drawn out to the distance its rule names, however little stands in it - a chunk with one group of rocks was culled 60 m away under a rule that said 1200
c396fc2 yhh committed 8 days ago
feat(gi): a half-resolution way to add the light - one pixel in four asks the probes and every pixel takes it from the four round it; the medium level runs it, and the apply costs 0.14 ms where it cost 0.25 at 1080p
b76aea1 yhh committed 8 days ago
feat(quality): the low level leaves the dynamic GI out, and a level that does builds no distance field for it either
67e4e05 yhh committed 8 days ago
fix(gi): no bright and dark blobs beside an edge the camera moves past - a probe whose own history is of another surface still mixes with the neighbours that are on its own
78cdb63 yhh committed 8 days ago
fix(gi): walls that looked dirty - a probe that has not moved keeps five times the history, so the blotches left of the grain on a still wall go
0383c18 yhh committed 8 days ago
fix(sidemodules): a game started over and over in one page no longer runs out of WebAssembly memory - the module host has an end, the App that made it disposes it on quit, and the KTX2 transcode workers stop with it
1915e10 yhh committed 8 days ago
fix(native): the device language reaches the game - Android, iOS and desktop report the system locale, and 'zh-Hans-CN' reads as the 'zh-CN' a browser gives instead of always 'en'
63e25bc yhh committed 8 days ago
fix(gi): no black band along the edges of a room - a ray that starts in a corner is moved clear of the wall next to it, and a card pixel buried in another body stays where it is so a sealed room stays dark
4733c67 yhh committed 8 days ago
perf(gi): a scene with no GI yet is built 12 ms a frame instead of 3 - on a phone the first light comes in about 2 s, not 10; a moving camera keeps 3 ms
ec610af yhh committed 8 days ago
perf(gi): the first build of the field is sliced over frames, and a camera move sends only the rebuilt range - the cards keep their light instead of starting the bounces again every step
e94aca2 yhh committed 9 days ago
perf(gi): a camera that moves rebuilds the following field a few milliseconds a frame - boxes built only when the scene changes, levels repacked only when they changed, slabs queued; measured with a walking camera
b5b1244 yhh committed 9 days ago
test(gi): spot lights on the GPU - a spot lab held to the CPU reference in the cone, the penumbra and out of it
0a7fded yhh committed 9 days ago
perf(gi): half the probe rows on a phone, and a continuous probe filter - GI on a Snapdragon 8 Elite 20.5 -> 6.4 ms
c5fd8e7 yhh committed 9 days ago
perf(gi): the scene key is asked every frame and the boxes are built only when it changes; profiling switch u_skip is declared so it reaches the engine
a61aff9 yhh committed 9 days ago
test(gi): a cost gate for the dynamic GI - the work its passes are asked to do (probe rays, cards relit, reads a pixel, bytes forced on the scene pass), counted at the dispatch and held to budgets on any machine
e26875e yhh committed 9 days ago
fix(gi): the dynamic GI runs on a device - the scene is fed through the native heap, the apply reads each probe's stored position instead of rebuilding it; phone numbers and a skip switch for measuring
ceca457 yhh committed 9 days ago
feat(gi): a distance field of three levels that follows the camera, and point and spot lights - lit in the cards with the frame's own falloff, shadowed through the field, held to the CPU reference on the lamp lab
818ebb8 yhh committed 9 days ago
perf(gi): dynamic GI 8.9 -> 3.3 ms at 1080p - a card relit by a workgroup of 64 rays, cards relit in slices, probes by stride and a fixed count, the nearest instance recorded per field cell; frame timing tool
cb83120 yhh committed 9 days ago
fix(gi): the dynamic GI has no dark seam in a wall corner and a quieter picture - probes are taken from the nine nearest with a tent weight, history blend 0.15
8ab01be yhh committed 9 days ago
feat(gi): the dynamic GI reads material emission and baked meshes - cards carry the emissive colour, a mesh brings its distance field and cards through the cook, checked on the lab rooms as meshes
38c76af yhh committed 9 days ago
feat(gi): dynamic GI on the GPU - card lighting, 8x8 screen probes traced through the distance field, applied to the picture; no leak in the sealed room and 0.54-1.08 of Cycles in the lab rooms
bc5443f yhh committed 9 days ago