erik/acdream
1
0
Fork 0
Code Issues Pull requests Projects Releases 6 Packages Wiki Activity Actions

feat(render): Campaign V slice V2a - mesh path texture-index migration

Browse source 
Moves the mesh/EnvCell draw path's per-batch texture representation from a
64-bit ARB_bindless_texture handle to a small integer table index, entirely
on the still-shipping GL backend, with zero pixel change. This is the CPU-side
half of the eventual Vulkan descriptor-array indexing model: a table index is
the backend-neutral form (Vulkan indexes a descriptor array with it directly),
while a raw bindless handle is GL-only. Landing the data-model change now, on
GL, under a strict self-differential pixel gate, keeps it separate from V4c's
much larger RHI-plumbing change (see docs/plans/2026-07-27-vulkan-campaign.md
section 5.2 for why the table cannot be device-owned yet).

Mechanism: mesh_modern.vert's BatchData struct carries `textureIndex` (a slot)
instead of `textureHandle` (uvec2); the vertex shader looks the slot up in a
new binding=9 storage buffer (GpuBindingModel.StorageTextureTable) and passes
the reconstructed uvec2 handle to the fragment shader exactly as before, so
mesh_modern.frag needed no change at all beyond the UBO-set macro below. The
16-byte std430 stride is unchanged (GpuBindingModel.GpuBatchDataStrideBytes);
textureLayer/flags keep their offsets, so every existing CPU writer's layout
is untouched.

The handle->slot table (GlBindlessHandleTable, new, pure C#) is owned
separately by WbDrawDispatcher and EnvCellRenderer rather than shared through
a single TextureCache-owned instance: EnvCellRenderer never had a TextureCache
dependency, and nothing requires index agreement between renderers since each
rebinds its own binding=9 buffer immediately before its own draw call. This
avoided threading a new constructor parameter through EnvCellRenderer (and its
six test call sites) for no behavioral benefit. TextureCache and
CompositeTextureArrayCache turned out to need no changes at all: they only
ever produce raw ulong handles, and that production path is unaffected -
the new indirection is entirely a WbDrawDispatcher/EnvCellRenderer-side
concern, added exactly where each already assembles its per-batch GPU struct
(ToInput, the copy-back loop, PrepareDeferredAlphaDraws for the
RetailAlphaQueue path, and EnvCellRenderer's ModernBatchData construction).
The table itself is a single non-ring buffer (unlike the per-frame
triple-buffered SSBOs) because a genuinely new handle is rare - new dat
surfaces/composite overrides, not every frame - so it flushes only when
GlBindlessHandleTable.Dirty is set, mirroring how the existing texture caches
already upload infrequently.

Shader-side, introduced Rendering/Shaders/common.glsl as the shared preamble
GL has no #include for: Shader.cs gained an `includeCommonPreamble` overload
that splices the file's text in after the leading #version/#extension block
(GLSL requires #version first). It declares the binding=9 table plus the
ACDREAM_TEXTURE_HANDLE(idx) lookup macro, and a scaffolding ACDREAM_UBO_SET
macro (a no-op under GL today, redefined to `set = 1,` when the Vulkan
toolchain compiles this same source at V6+, per the campaign doc's set-1 UBO
note) applied to both SceneLighting UBO declarations now so no later slice
needs to touch them again.

Tests: WbDrawDispatcherIndirectBuilderTests updated for the renamed
IndirectGroupInput/BatchDataPublic fields; new ModernBatchDataLayoutTests
(mirrors ClipFrameLayoutTests' role, but for EnvCellRenderer's GPU struct) and
GlBindlessHandleTableTests (pure-CPU allocator behavior, including the
zero-handle case, which is registered like any other handle rather than
special-cased, since that's what reproduces the pre-V2 sampling result
bit-for-bit).

Gate: dotnet build -c Release green, dotnet test
tests/AcDream.App.Tests -c Release green (3843 passed / 3 skipped, +9 over
the 3834/3 baseline), and tools/run-offline-pixel-gate.ps1 passed with a
2.84e-05 differing-pixel fraction against the parent commit - within the
documented ~33x same-commit noise margin. No divergence-register row: this
introduces no retail behavior deviation.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Erik 2026-07-27 15:51:26 +02:00
parent 8b57ba7167
commit d365476ebb
12 changed files with 550 additions and 49 deletions
Download patch file Download diff file Expand all files Collapse all files

33
tests/AcDream.Core.Tests/Rendering/Wb/WbDrawDispatcherIndirectBuilderTests.cs
View file

@ -20,9 +20,9 @@ public sealed class WbDrawDispatcherIndirectBuilderTests
// Arrange — three groups: 2 opaque (12+1 instances) + 1 transparent (12 instances)
var groups = new List<WbDrawDispatcher.IndirectGroupInput>
{
new(IndexCount: 100, FirstIndex: 0, BaseVertex: 0, InstanceCount: 12, FirstInstance: 0, TextureHandle: 0xAA, TextureLayer: 0, Translucency: TranslucencyKind.Opaque),
new(IndexCount: 200, FirstIndex: 100, BaseVertex: 0, InstanceCount: 12, FirstInstance: 12, TextureHandle: 0xBB, TextureLayer: 0, Translucency: TranslucencyKind.AlphaBlend),
new(IndexCount: 50, FirstIndex: 300, BaseVertex: 100, InstanceCount: 1, FirstInstance: 24, TextureHandle: 0xCC, TextureLayer: 0, Translucency: TranslucencyKind.Opaque),
new(IndexCount: 100, FirstIndex: 0, BaseVertex: 0, InstanceCount: 12, FirstInstance: 0, TextureIndex: 0xAA, TextureLayer: 0, Translucency: TranslucencyKind.Opaque),
new(IndexCount: 200, FirstIndex: 100, BaseVertex: 0, InstanceCount: 12, FirstInstance: 12, TextureIndex: 0xBB, TextureLayer: 0, Translucency: TranslucencyKind.AlphaBlend),
new(IndexCount: 50, FirstIndex: 300, BaseVertex: 100, InstanceCount: 1, FirstInstance: 24, TextureIndex: 0xCC, TextureLayer: 0, Translucency: TranslucencyKind.Opaque),
};
var indirect = new DrawElementsIndirectCommand[16];
@ -57,9 +57,9 @@ public sealed class WbDrawDispatcherIndirectBuilderTests
Assert.Equal(12u, indirect[2].BaseInstance);
// BatchData parallel — same indices as indirect
Assert.Equal(0xAAul, batch[0].TextureHandle);
Assert.Equal(0xCCul, batch[1].TextureHandle);
Assert.Equal(0xBBul, batch[2].TextureHandle);
Assert.Equal(0xAAu, batch[0].TextureIndex);
Assert.Equal(0xCCu, batch[1].TextureIndex);
Assert.Equal(0xBBu, batch[2].TextureIndex);
Assert.Equal(CullMode.CounterClockwise, cull[0]);
Assert.Equal(CullMode.CounterClockwise, cull[1]);
Assert.Equal(CullMode.CounterClockwise, cull[2]);
@ -71,13 +71,13 @@ public sealed class WbDrawDispatcherIndirectBuilderTests
var groups = new List<WbDrawDispatcher.IndirectGroupInput>
{
new(IndexCount: 10, FirstIndex: 0, BaseVertex: 0, InstanceCount: 1, FirstInstance: 0,
TextureHandle: 0x1, TextureLayer: 0, Translucency: TranslucencyKind.Opaque,
TextureIndex: 0x1, TextureLayer: 0, Translucency: TranslucencyKind.Opaque,
CullMode: CullMode.Clockwise),
new(IndexCount: 20, FirstIndex: 10, BaseVertex: 0, InstanceCount: 1, FirstInstance: 1,
TextureHandle: 0x2, TextureLayer: 0, Translucency: TranslucencyKind.AlphaBlend,
TextureIndex: 0x2, TextureLayer: 0, Translucency: TranslucencyKind.AlphaBlend,
CullMode: CullMode.None),
new(IndexCount: 30, FirstIndex: 30, BaseVertex: 0, InstanceCount: 1, FirstInstance: 2,
TextureHandle: 0x3, TextureLayer: 0, Translucency: TranslucencyKind.ClipMap,
TextureIndex: 0x3, TextureLayer: 0, Translucency: TranslucencyKind.ClipMap,
CullMode: CullMode.Landblock),
};
var indirect = new DrawElementsIndirectCommand[4];
@ -113,7 +113,7 @@ public sealed class WbDrawDispatcherIndirectBuilderTests
// because the discard handles transparency, not blending.
var groups = new List<WbDrawDispatcher.IndirectGroupInput>
{
new(IndexCount: 10, FirstIndex: 0, BaseVertex: 0, InstanceCount: 1, FirstInstance: 0, TextureHandle: 0x1, TextureLayer: 0, Translucency: TranslucencyKind.ClipMap),
new(IndexCount: 10, FirstIndex: 0, BaseVertex: 0, InstanceCount: 1, FirstInstance: 0, TextureIndex: 0x1, TextureLayer: 0, Translucency: TranslucencyKind.ClipMap),
};
var indirect = new DrawElementsIndirectCommand[4];
var batch = new WbDrawDispatcher.BatchDataPublic[4];
@ -130,9 +130,18 @@ public sealed class WbDrawDispatcherIndirectBuilderTests
// Task 10 will use MemoryMarshal.Cast<BatchData, BatchDataPublic> to
// expose the dispatcher's per-frame BatchData[] scratch to BuildIndirectArrays
// without copying. The cast is only safe if the structs have identical
// layout (size, field offsets). Both use [StructLayout(Sequential, Pack=8)].
// layout (size, field offsets).
//
// Campaign V slice V2 (2026-07-27): TextureHandle (ulong, an
// ARB_bindless_texture handle) became TextureIndex (uint) plus an
// explicit Reserved pad word — a slot into the binding=9 handle table
// (GpuBindingModel.StorageTextureTable) instead of the raw handle.
// The struct stays 16 bytes and TextureLayer/Flags keep their offsets
// (8/12), matching GpuBindingModel.GpuBatchDataStrideBytes and every
// existing CPU writer, so both structs only need 4-byte packing now.
Assert.Equal(16, System.Runtime.CompilerServices.Unsafe.SizeOf<WbDrawDispatcher.BatchDataPublic>());
Assert.Equal(0, (int)System.Runtime.InteropServices.Marshal.OffsetOf<WbDrawDispatcher.BatchDataPublic>(nameof(WbDrawDispatcher.BatchDataPublic.TextureHandle)));
Assert.Equal(0, (int)System.Runtime.InteropServices.Marshal.OffsetOf<WbDrawDispatcher.BatchDataPublic>(nameof(WbDrawDispatcher.BatchDataPublic.TextureIndex)));
Assert.Equal(4, (int)System.Runtime.InteropServices.Marshal.OffsetOf<WbDrawDispatcher.BatchDataPublic>(nameof(WbDrawDispatcher.BatchDataPublic.Reserved)));
Assert.Equal(8, (int)System.Runtime.InteropServices.Marshal.OffsetOf<WbDrawDispatcher.BatchDataPublic>(nameof(WbDrawDispatcher.BatchDataPublic.TextureLayer)));
Assert.Equal(12, (int)System.Runtime.InteropServices.Marshal.OffsetOf<WbDrawDispatcher.BatchDataPublic>(nameof(WbDrawDispatcher.BatchDataPublic.Flags)));
}