Heights as u16 over each 64 m tile's own minimum and step (a tile spanning 200 m steps 3 mm), normals
octahedral 8 + 8, the photograph RGB8, and the coarse level (2048: heights f32, normals, photograph)
- 30 + 32 + 48 + 16 + 8 + 12 MB, about 146 MB a map where the float tiles were 192 plus nothing coarse.
The writer puts the whole copy back to the quantized values as it goes, so the build's own queries,
the physics, the placements' bake and every machine read the same numbers; a tile read decodes them
into the pools the queries and the page pool already use.
terrain_tiles_bake(path, key, version, inputs_hash) writes a bake's file (ludic.base's LBAK header,
the tiles as its payload) from a made map; terrain_from_baked(path, key, version, half, ox, oz) opens
one at boot in place of terrain_use_dem / terrain_use_ortho, and terrain_init then generates nothing
(and bakes the sun's shadow from the coarse level unless terrain_shadow_from_bytes gave it). Without a
bake the cut writes the same format to <dir>/<key>.tiles and reads it back. The GPU's coarse level is
made from the file's coarse sections (the blit from the whole maps is gone).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
At the cut the whole 4096 height, normal and photograph textures go, replaced by a coarse
2048 level (the CPU's tt_coarse uploaded; the normal and photograph blitted down on the GPU,
the photograph mipmapped) and a pool of fine tiles in three array textures - heights, normals,
photograph, each layer a tile with a one-texel border - addressed through a tt_n^2 page table
(u_tp_page: slot + 1, 0 = the coarse level). The pool holds the tiles within the reach (900 m,
or the fog wall's when nearer) and a ring, and is made again when the fog wall changes its size
(terrain_pages_fog). Once a frame (tp_frame, beside tt_frame) tiles past the reach and two tiles
go and the wanted ones come in nearest first, 8 a frame, written into a staging buffer kept for
the process (two halves, one per frame in flight) and copied into their layers inside the frame's
own command buffer - no submit of their own - with the page table re-uploaded only when it
changed. tp_bind binds the pool (or 1-layer stand-in arrays while paging is off, u_tp_on = 0) for
every program that reads the ground: terrain_bind_height (models, scatter, shadow_bind, grass),
terrain_bind_prog, the sun pass and the shadow bake. grass_cull.comp reads through the same page
table. R3D_VKMEM prints the pool's line. Tiles off, nothing changes. Compile-only: not run.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
terrain_tiles_to(dir, key) before a map is made: once made, its heights, photograph and baked normals
are written tile by tile (64 m, TT_TEX) to <dir>/<key>.tiles - fresh every time, header last, so no
other generator's or a torn file is ever read - and the whole copies go. Every read goes through the
tile: resident, else read from the file there and then into a 2048-tile clock, so terrain_height,
terrain_height_smooth, terrain_ortho* and terrain_chunk_heights answer exactly what the whole copy
did (R3D_TT_CHECK: worst 0.0 m over 4000 points) whatever is resident - two machines and the boot's
placements agree to the bit. terrain_chunk_heights takes render3d_st mut for it.
terrain_height_near(read-only): the tile if it is in, else a coarse 2048^2 level (worst 0.91 m), never
the file - for line checks that must not take render3d_st mut. terrain_texel() is the texel size,
terrain_tiles_prefetch(x, z, r, budget) reads ahead, and a frame that reads more than 8 tiles says so
once. R3D_TERRAIN_TILES=<dir> turns it on for a run.
At the overlook with MallocLargeCache=0: 1731 MB off, 1658 MB on; 192 MB written in ~200 ms; the
frame unchanged (0 pixels over 8).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The DEM (R16 with mips, 44 MB) was read only by terrain_generate and kept for the map's life; a
world swap loads its own again. The height-field sun shadow was one RGBA32F for three values and an
unused fourth: the lowest lit height stays 32-bit (a receiver 3000 m up compares against it to a
quarter metre), the occluder distance and the cloud mask go beside it in RG16F (64 -> 32 MB), baked
in a pass of their own (TS_AUX): a pipeline takes one colour format for all its targets, and drawn
together into R32F + RG16F the writes went nowhere and nothing was lit.
-75 MB at every fog level, off included (2008 -> 1933 MB at the overlook); the fog-off frame is
unchanged (0 pixels over 8).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every allocation carries the owner its maker was wrapped in (models, terrain, impostor atlases,
scatter, grass, shadow maps, frame targets, sky, water, game textures, made in a frame), and the
report prints each owner's images and buffers, the 25 largest images and the scatter layers',
streams' and terrain's CPU copies. Nothing drawn changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Seventeen inline messages (allocation failures, missing conversions, the no-pipeline and compressed-
target warnings, resize and swapchain lines, the test-frame camera line, DLSS evaluate, shadow memory,
stream and terrain debug, the water test) are each a function of their own under @alloc_ok, taking
numbers, so the paths that say them hold no site. post_measure's two 1x1 exposure targets are made in
post_init and DLSS's evaluate structs in gsl_init; DLSS's camera up is a state field. The lists play
pushes into (retired, free ranges, spare ids, the per-buffer gpu flags, the stage and the actor spares)
get their room at start-up (gvk_room_*, actor_room), so a push fits. Compiled (steady).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Render3dState's cam_planes, gpu_u_tmp, q_scratch, ov_nine_buf, the caster test's four points,
ter_bw, ter_scr and water_saved default to their buffers (as ludic.anim's clip state does), so the
lazy starts in cam_planes_update, gpu_tmp, terrain_height_smooth, ter_scratch, the water pass and
gvk_startup_state go. q_euler makes its views of q_scratch once (guarded on the view now). Compiled.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The renderer starts itself from the first frame (gpu_select), so the device, its tables, the
manifest and programs, grass, shadows, sky, terrain textures and the actor pool read as frame
allocations: each is @alloc_ok as start-up, with gvk_fail (a failure) and the actor census and
texture dump (debug switches). ov_nine's four corner/uv arrays are one floats(16) made in
gvk_startup_state; ac_in_light's four points are made there too. Compiled (steady, and ludic deps
over main 4316ff97).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The first generation pass's R32F height is kept as ter_height_tex (not copied into an RGBA32F),
so every reader of the height - placement, the read-back, physics, selftest16's 9 mm - sees the
same 32 bits. ternormal.frag writes the baked normal's x and z into ter_normal_tex (RG16F), and
the six places that read it (terrain.frag twice, tersun.frag, grass.vert, grass.mesh,
grass_cull.comp, which takes a third texture at binding 6) rebuild y. SPIR-V regenerated.
Maroon Lake's play, headless Vulkan: 2793 -> 2647 MB. The camp's frame: 0.066% of pixels differ by
more than 8 (mean 0.024/255), isolated grass blades at the slope gate. ludic-dev test 307/307.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
terrain_generate, terrain_use_dem and terrain_use_ortho each overwrote the previous height
texture (256 MB of RGBA32F at 4096^2), the heights read back (64 MB), the DEM texture and the
orthophoto's texture and pixels. terrain_reload unloads first, so the world swap was already
clean; any other caller building a map over a live one now lets the old go, and the read-back
array is reused, being the same size for every map.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
ludic migrate state packages <every example program> packages/ludic.lab/example/plate.ludic
1804 vars into 126 states, 64 into lets; 23498 edits in 460 files
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A scene is an entry in the open registry LabScenes, shown on a plate: a
flat lit disc, the package's own small sky with the sun at one hour, and
a frame clock. Headless each camera (lab_shot, lab_shot_at) settles and
is written as build/lab/<scene>/<shot>.png (a stored PNG, written here);
in a window N and P step through them. tools/lab/run.sh and sheet.py
make the contact sheet; tools/lab/plate_build.py makes plate/.
ludic.render3d gains r3d_plate_mode (no terrain, grass or water is made)
and r3d_sky_path. quit() in a program with no frame loop links. The
example takes ~120 MB resident, 417 MiB peak footprint. Reseed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The typed buffers are slices: words/floats/fixeds/doubles/pointers(n) make
zeroed, bounds-checked []int/[]float/... and the type names mean them. buffer(n)
is a []byte, with text_of, Fs.read_bytes/write_bytes and view(xs, start, n).
bytes(), indexing a raw pointer or bytes, free, resize, Memory.*, raw file calls,
data_of and C externs are refused outside unsafe { } / unsafe function, and a
project's own files may write unsafe only with --unsafe; the runtime and packages
are the platform. A slice passed to an extern goes as its data.
What the change found: Sync's atomics on a slice header, words(n) uninitialised,
input's fixed axes in ints, truetype's fixed outlines as ints, skin matrices
typed int, gl_shader's source table made from raw bytes. render3d gets safe
entry points (safe_api.ludic). Rendering is byte-identical; a frame costs the same.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A numbers float file adapts decimal literals to a fixed operand or slot, and refuses to
promote a computed int to a float implicitly: there it is almost always float bits. Explicit
float(x) is always allowed.
render3d's numbers are float, converted by tools/migrate/floatbits.py - a whole-program
inference of which ints carried IEEE bits (union-find over flows, calls, returns, buffers,
nested buffers and lexical scopes) and a rewriter to operators, Math.* and float literals,
with float_bits / float_from_bits left only where bits really cross (runtime scratch
buffers, mixed buffers). Seed regenerated.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Input.text, App.monitor_count / window_to_monitor / window_fixed,
gl_sleep_us, and the grass and collide changes, uncommitted on main and
depended on by the game; carried here so the language work starts from
what the game actually uses. main's working tree is untouched.
Rain fell and the ground did not change: the weather was a curtain of particles in front
of a dry valley.
u_wet, 0 dry to 1 soaked, does the three things a wet surface does. Darkens the albedo.
Drops the roughness hard - which is what makes a soaked meadow read as soaked rather than
merely dark, because the sky glances off it. And pools: water finds the low places and
flat ground holds what a slope sheds, so the puddle mask is the macro noise the materials
already use gated on slope, and a puddle takes the world's up for its normal rather than
the ground's relief.
It joins the shore's existing wet band rather than fighting it - the same term from a
different cause.
34.1% of the frame changes between dry and soaked. GL 7.1 s either way over 400 frames.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The grass had a gust built from two sines; the trees had no gust term AT ALL, only a
per-instance wobble. So the meadow rippled and the canopy above it swayed to an unrelated
rhythm, and nothing ever crossed the valley.
wind.glsl is prepended to every stage (programs.ludic, and shaders.ludic for the SPIR-V
build) so grass, crowns and water read the same field at the same world position. It
declares no uniforms on purpose: u_time and u_wind already exist in several of those
files and redeclaring them is a compile error in whichever stage includes both.
The wavelength is what made it work. At 0.030 the front was 209 m crest to crest -
longer than the meadow you can see - so the whole frame sat in one phase and the gust
read as everything breathing together. At 0.085 it is 74 m and a front crosses a view in
a couple of seconds.
R3D_DEBUG_WIND=1 paints the field on the ground. It is the only honest way to show a gust
in a still image, and two shots 0.3 s apart move by 39/255.
400 frames: GL 6.9 s, VK 7.2 s. Backends agree to 0.68/255.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A reflecting body is clipped to its ellipse like every other unless it is an
unbounded sea, so a map whose reflection belongs to its lake (Maroon Lake, once
its sea sits below it) does not draw that lake's level over every hollow in the
survey. The terrain's wet shore and its forest and scree gates read the carved
lake's line inside its outline (u_lake), the rule grass already used. SPIR-V
regenerated.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Every R3D_* read goes through r3d_env_has / r3d_env (env.ludic): a headless
build honours them as before, a windowed build only when R3D_DEV is set to
anything but "0", so a shipped game never reaches a debug view, feature kill,
file writer or hardware fake. Documented in docs/SHIPPING.md.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
tersun.frag rasterised every terrain patch a second time into a screen buffer, because on OpenGL the
cascade read inside terrain.frag fell off a driver cliff (about 6 ms a frame). Vulkan has no such
cliff: its terrain programs are the SUN_INLINE variant, which evaluates the same two tiers with the
same normal and cross-fade in the ground's own shader, and terrain_draw skips the pass - in the frame
and in the water reflection. OpenGL keeps the pass. R3D_SUN_PASS=1 keeps it on Vulkan, for comparing.
Mac Vulkan town 1759 -> 1692 draws; frames within 2/255 of the pass (15331 px, float rounding). PC camp
1984 -> 1882 draws, 3.9 s for 400 frames either way, self-tests 59/59, validation 0. OpenGL frames
byte-identical.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
- sky_start_yaw: a game that turns the sky at boot sets it before r3d_init and the image-based light
is baked once at that yaw; sky_set_yaw to a yaw already baked bakes nothing.
- terrain_init_step: the height field and normals, the sun shadow, the materials, the patches and
programs as separate steps; r3d_load_count is 8, so a loading bar moves through the terrain (half
the start-up) instead of jumping over it. terrain_init runs the four in a row as before.
- The per-cascade actor cull centres its sphere at half the model's height and sizes it by the
larger of height and radius: centred at the radius, it dropped a head at a cascade's edge (11 px
in town, found by the drawstats comparison). Lake was byte-identical before and after.
OpenGL frames at the five viewpoints unchanged; the game's 59 self-tests pass.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The last of milestone 1: no OpenGL call is left in render3d outside gpu.ludic but the clock
and the CPU-side buffer helpers.
- gpu_program builds a program and remembers the variant it came from (vertex, fragment and
defines as one line - the SPIR-V manifest's key), so a backend that cannot compile at run
time finds the pipeline for the same handle. gpu_use_program and gpu_program_free replace
32 uses and 2 frees; the overlay's program is recorded under overlay.vert|overlay.frag.
- gpu_query_new / _begin / _end / _result carry R3D_PROF's timers.
- gpu_open, gpu_vsync, gpu_renderer_name and gpu_resize_check carry the context.
- r3d_program_tess is gone: nothing called it and no tessellation shader exists.
- R3D_GLCHECK also checks after framebuffer and renderbuffer changes, viewport, draw buffers,
program use, texture parameters, the resize check and between frames.
OpenGL frames are byte-identical at the five viewpoints; 59 self-tests pass with and without
R3D_GLCHECK, with no GL error; ludic-dev test 140 passed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Framebuffers and their attachments, renderbuffers (multisampled too), draw and read buffers,
completeness checks, blits, viewports, clears, the screen framebuffer, the multisample enable,
the wireframe switch and GL error checks now go through gpu_fb_* / gpu_rb_* / gpu_viewport /
gpu_clear / gpu_blit / gpu_check, and no other file names them. Each call is the one GL call it
replaces, in the same order: the fixed viewpoints render bit-identically and the game's
self-tests report exactly what they did.
What is attached to each framebuffer - colour slots, a depth texture or one layer of an array,
renderbuffers and their samples - is recorded as it is attached, for a backend that builds
render passes and image views. gpu_read_screen is the frame read-back a photograph takes.
R3D_GLCHECK=1 checks, around every draw, clear and blit, that the bound framebuffer is complete
and that no error is left behind, naming the framebuffer. It has already narrowed the old
"gl error 1286 at terrain shadow bake": the error is pending before a draw after the map self-
test, so it comes from a call that is not a draw, clear or blit.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Texture creation, 2D and array uploads, pixel packing, filters, wraps, comparison, border,
anisotropy, mipmaps, texture units, read-backs and freeing now go through gpu_tex_* and
gpu_bind_sampler, and no other file names them. Each call is the one GL call it replaces,
in the same order, so OpenGL renders bit-identically at the fixed viewpoints and the game's
self-tests report what they did before.
What those calls say about a texture - size, format, layers, min and mag filter, wraps,
comparison, mipmaps, anisotropy - is recorded per handle as it is set: a backend with
immutable images and separate sampler objects creates both from exactly that.
r3d_bind_tex takes a texture kind (GPU_TEX2D / GPU_TEX2D_ARRAY) instead of a GL target.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Every vertex array, vertex and index buffer, attribute pointer, instance divisor, stream
upload and draw call now goes through gpu_mesh_* / gpu_buffer_* / gpu_draw_*, and no other
file in the package names them. A Mesh records its layout as it is built - which buffer
feeds which attribute at what stride and offset, per vertex or per instance - so a backend
that bakes vertex input into a pipeline can read it back. On OpenGL each call is the GL it
replaces, in the same order: the five fixed viewpoints render bit-identically and the game's
self-tests report exactly what they did before.
scatter_attach takes the mesh rather than its vertex array; a mesh now frees every vertex
buffer it owns (glTF meshes used to keep all but the first); the two helpers nothing called,
mesh_grid_patches and mesh_instance_buffer, are gone.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Render state (depth, blending, culling, colour writes, alpha-to-coverage, depth bias,
scissor) and uniforms go through gpu_* / u_* and nowhere else; OpenGL state is cached and
only changes reach the driver. Frames are bit-identical to before at the fixed viewpoints.
R3D_GFX=gl|vk or gpu_request chooses a backend, falling back to OpenGL with a reason.
gpu_caps_probe() asks Vulkan, on Windows, for the 1.3 floor, ray tracing, mesh shaders, the
NVIDIA RTX generation, Reflex and HDR colour spaces, for a game's settings to grey out what
a machine cannot use; R3D_CAPS pretends to be a given card for tests.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
terrain_reload/terrain_unload release one map's height field, survey, photograph
and patch bounds and generate another at any TERRAIN_HALF; scatter_clear_all
(with streams), actor_clear_all, col_reset and mesh_free empty the scene;
water_body_add/water_bodies_clear draw several still-water planes with one
reflecting; terrain_sea separates the coast's sea level from the carved lake's,
and grass keeps off both. Fixes CDLOD patch bounds for TERRAIN_HALF != 4096 and
water_init leaking a mesh and program per call. examples/rendering/reload.ludic.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Three things a game could not say, each of which had been worked around.
Audio.play_at(id, gain:, pitch:, pan:) fires a one-shot with its own gain,
pitch and stereo position. Audio.volume and Audio.pitch are global - they are
the options screen - so a game placing a sound in the world was fighting them,
and distance attenuation was simply not expressible. The backend already took
volume and rate per call; this adds setPan: alongside them and stops routing
through the master state.
ludic.render3d gains outline_model(model, mat, width, r, g, b): a rim around
something that is not an Actor. The outline pass walked the actor list and
stopped, so instanced scatter - a forest - could not be highlighted at all. It
is a queue flushed by the same pass, which is what gets the depth test right
when the caller does not control pass order.
And terrain_coast(cx, cz, margin, fall), the other way to make an island:
the sea around the survey's own edge rather than cut out of the middle of it.
terrain_island measures a radius from a centre, which drowns two thirds of a
real survey to make an island of the rest; this measures inward from the
boundary, so everything the data covers stays land and the coast is where the
data runs out. Both modes gained a strand - the last few metres of height
either side of the water line compressed, which stretches a cliff plunge out
into beach and shallows.
132 regression tests and the self-host fixpoints pass.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
`terrain_island(cx, cz, r, fall)` keeps land out to `r` and then scales the
terrain down into the water over `fall` metres and on to a shelf. Scaling rather
than blending to a fixed bed is what makes the coastline come out of the terrain
that is already there: low ground turns into beach and shallows, high ground
into cliff. `r = 0` leaves the survey alone, so nothing that does not ask for an
island is touched.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
`Gl.*` binds the whole OpenGL 4.1 core API — every entry point of the
platform gl3.h with every GL_* constant, generated by `ludic-dev glgen`
with per-call ABI thunks. Windowed builds get an NSOpenGLContext on the
existing window at Retina resolution; headless builds render into an
offscreen CGL context, so a program that uses Gl.* renders and
screenshots identically under the test harness. It links gl.ll, the
thunks and OpenGL.framework only when used; every other build stays
byte-identical.
packages/ludic.render3d is a physically based renderer written on that
surface: HDRI image-based lighting, GPU-generated terrain with scanned
PBR materials, CDLOD, cascaded shadows, glTF with skinning, instanced
vegetation with impostors, procedural grass, water, SSAO, and an HDR
pipeline with bloom, auto-exposure and ACES.
It also carries this session's work on it: the terrain at half its cost
(10.3 -> 5.4 ms of frame), the streaming hitch that got worse the longer
you played, a resize that emptied the world, and the packaging that lets
a game use the renderer from its own repository — `ludic assets`, the
material manifest shipping with the package, and shader lookup falling
back to the install root. See changes/ for each, with its numbers.
The camping game that drove all of it has moved out to its own
repository, Maroon Lake; examples/rendering/smooth.ludic stays as the
renderer's example here.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
