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

feat(render): Campaign V slice V6j commit 2 - Dereth draws on Vulkan

Browse source 
The three world renderers' submission arms, both pass executors, and the
composition that reaches them. This is the unit three predecessors stopped at.

What it produces. ACDREAM_RENDER_BACKEND=vulkan on the offline scene renders
terrain with blended textures and road overlays, the water edge, static world
meshes, procedural scenery, and the complete retained UI - the same frame the GL
pixel gate captures, from the same camera, minus the sky. artifacts/v6j-vk2.

The shape, and why it is not V4c's. Section 5.5.6 chose option (B) after NVIDIA
rendered the V4c binary 10/10 where AMD's GL stack did not: GL keeps its raw
world path through to V10 as a documented fork confined to the submission seam,
and the RHI world path ships on Vulkan. So V4c's and V4d-2's content returns as a
SECOND arm rather than a replacement. The GL arm issues the same GL statements in
the same order against the same objects; the encoder arm lives in three .Rhi.cs
partials and is entered by one branch per submission site.

Three differences from V4c, each because the tree moved under it. There is no
binding-9 texture table - V4t put the slot on the device and Vulkan binds set 2,
so the arm that used to intern bindless handles simply has nothing to do. The
pipelines carry the device's sample count rather than 1, because Vulkan requires
rasterizationSamples to match the pass and alpha-to-coverage is a no-op at one
sample. And no renderer opens a pass.

That last one is structural, not tidiness. Under MSAA the frame's one backbuffer
pass resolves into the swapchain image and stores DONT_CARE into the multisampled
scratch, so a second pass declaring Load would load undefined contents; the
backend also permits one open pass per frame. VulkanWorldScenePhase therefore
opens the pass, publishes the encoder on VulkanWorldPassScope for exactly the
span of the inner WorldSceneRenderer, and every renderer borrows it.

Three sections are frame-global on GL and cannot be on Vulkan: the SceneLighting
UBO, the per-cell clip regions, and the terrain clip block. GL binds each to a
global binding point and every consumer inherits it. Vulkan binds a descriptor
set per draw, and a renderer's own binds are what select the scope those sections
must land in - so their writers PUBLISH into WorldFrameSections and each renderer
binds them inside the pass, after its own binds. SceneLightingUboBinding's
per-flight-slot buffer pool disappears with it: a ring allocation is already
distinct memory that lives until the frame retires, which is the property the
pool existed to provide.

Both pass executors became backend-neutral rather than gaining twins. Everything
they do is delegation to a renderer except four concerns - the clip-frame
publication, the doorway scissor, gl_ClipDistance enablement, and retail's
interior depth clear - so those four move behind IWorldPassSurface and retail's
ordering, which is what these classes are actually for, is written once. The GL
implementation issues the statements the executors used to issue inline.

Clip distances are no-ops on the Vulkan arm, and that is safe rather than a
divergence: Vulkan activates every element the shader declares, and all three
world vertex shaders already write 1.0 into every slot past the active count.
The interior depth clear becomes vkCmdClearAttachments, reached through the scope
so the pinned contract stays frozen and the backend-only verb stays in the
backend. The hook for it was already committed at V6i-3 with a cref to a type
that did not exist yet; it exists now.

The collision-wireframe DebugLineRenderer is composed as null on the Vulkan arm.
DrawAndPublish flushes it INSIDE the world phase and it opens its own pass, which
the one-pass rule forbids. The toggle is DevTools-only and DevTools is not
composed there, so nothing is lost - composing it would throw on the first
wireframe frame rather than silently misdraw.

Two seams widened rather than invented. GameWindowGraphics answers whether the
backend has a world-pass seam, because the three composition phases that need it
already borrow that handle and "does this backend work that way" is what the type
exists to answer. And MeshSourceReady replaces the anyVao != 0 gate with the same
question in backend-neutral form - V6i-3 published HasStores for exactly this -
so the predicate evaluates identically on GL.

What is NOT here, and is expected. Sky and weather are still raw GL (V4f), so the
Vulkan frame's sky is the atmosphere fog clear. Particles (V4e), the paperdoll and
appraisal viewports and the portal depth mask (V4g) likewise. The executors
already accepted all of them as absent.

Gates. Release build green. App tests 4,112 passed / 3 skipped, the unchanged
baseline; complete Release suite 9,175 / 5. Strict GL offline pixel gate against
847f14ae: 5.50e-05, 31 differing pixels of 563,200, inside the documented 9-31
band and 18x under the threshold. Characterised rather than accepted, because 31
is the band's top: cross-commit pairs measured 21, 29 and 31 while same-commit
controls measured 12 and 20, and maximumChannelDelta is 46-52 in every comparison
INCLUDING the pure controls - so the few large-delta pixels are a property of the
capture, and a cross-commit pair at 21 against a same-commit pair at 20 is not
what a systematic shift looks like. GL connected repeat gate at 3 runs: 3/3
RENDERED on the desktop witness and 3/3 on the client capture. One offline Vulkan
run with VK_LAYER_KHRONOS_validation proven inserted by the loader: zero
validation errors, zero warnings, a captured world frame, and a graceful close.

Coverage gap, stated rather than assumed. The offline scene is a fixed outdoor
view, so EnvCellRenderer's Vulkan arm draws nothing in it - dungeon interiors are
half of this slice and are unproven by anything automated, exactly as they were
for V4c. The deferred-alpha path and the doorway scissor are likewise untouched
by this scene. They join the accumulated user-gate debt in plan section 5.1.

No divergence-register row: no retail-facing behaviour changes.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Erik 2026-07-28 15:46:54 +02:00
parent 81fe5e1b63
commit f84eef3256
22 changed files with 2566 additions and 264 deletions
Download patch file Download diff file Expand all files Collapse all files

198
src/AcDream.App/Rendering/WorldPassScope.cs Normal file
View file

@ -0,0 +1,198 @@
using AcDream.App.Rendering.Gpu;
using AcDream.Core.Lighting;
namespace AcDream.App.Rendering;
/// <summary>
/// A buffer range reduced to the three values a bind needs.
///
/// <para><see cref="GpuRingAllocation"/> is a <c>ref struct</c> — deliberately,
/// so nothing can outlive the frame's memory — but the buffer reference plus its
/// offset and size are ordinary values and stay valid for as long as that memory
/// does. That is what lets one writer publish a section and several renderers
/// bind it later in the same frame without recopying.</para>
/// </summary>
internal readonly record struct GpuBufferSection(
IGpuBuffer? Buffer,
uint OffsetBytes,
uint SizeBytes)
{
public bool IsValid => Buffer is not null && SizeBytes > 0;
}
/// <summary>
/// Campaign V slice V6j: the three sections GL binds frame-globally and Vulkan
/// cannot.
///
/// <para>On GL the SceneLighting UBO (set 1 binding 1), the per-cell clip regions
/// (set 0 binding 2) and the terrain clip block (set 1 binding 2) are each bound
/// once, to a global binding point, and every consumer inherits them. Vulkan has
/// no global binding points: a descriptor set is bound per draw, and a renderer
/// that binds its own buffers selects the descriptor scope those sections have to
/// land in (plan §5.5.14 item 2). So the writers PUBLISH here and each renderer
/// binds them inside the pass, after its own binds.</para>
///
/// <para>Borrowed for the frame that publishes it — the sections are ring slices
/// and die when the frame retires.</para>
/// </summary>
internal sealed class WorldFrameSections
{
/// <summary>Set 1 binding 1 — <c>SceneLighting</c>.</summary>
public GpuBufferSection SceneLighting { get; set; }
/// <summary>Set 0 binding 2 — the per-cell <c>CellClip</c> table.</summary>
public GpuBufferSection ClipRegions { get; set; }
/// <summary>Set 1 binding 2 — <c>TerrainClip</c>.</summary>
public GpuBufferSection TerrainClip { get; set; }
public void Reset()
{
SceneLighting = default;
ClipRegions = default;
TerrainClip = default;
}
}
/// <summary>
/// Campaign V slice V6j: the one render pass every world renderer records into,
/// borrowed rather than opened.
///
/// <para><b>Why they cannot each open one.</b> Under MSAA the frame's backbuffer
/// pass renders into a multisampled scratch image and RESOLVES it into the
/// swapchain image, storing <c>DONT_CARE</c> into the scratch — so a second pass
/// declaring <c>Load</c> would load undefined contents and lose everything the
/// first drew. The Vulkan backend also permits only one open pass per frame.
/// V4c's shape, where <c>WbDrawDispatcher</c> and <c>EnvCellRenderer</c> each
/// bracketed their own <c>Load</c>/<c>Store</c> pass, is therefore not available
/// on this backend (plan §5.5.12 item 5, §5.5.14 item 1).</para>
///
/// <para>So the world-scene phase opens the pass, publishes the encoder here for
/// the duration of <c>WorldSceneRenderer</c>, and every renderer borrows it.</para>
/// </summary>
internal interface IWorldPassScope
{
/// <summary>
/// Sample count of the pass, and therefore of every pipeline recorded into
/// it. Vulkan requires a pipeline's <c>rasterizationSamples</c> to match the
/// pass, and alpha-to-coverage does nothing at one sample — which is why
/// V4c's blanket <c>SampleCount = 1</c> is not portable (plan §5.5.14 item 6).
/// </summary>
int SampleCount { get; }
/// <summary>The open encoder, or null outside the world phase.</summary>
IGpuPassEncoder? CurrentEncoder { get; }
/// <summary>The open encoder, or a composition error if the phase is not bracketing.</summary>
IGpuPassEncoder RequireEncoder();
/// <summary>Colour-attachment width in pixels, for scissor and clear rectangles.</summary>
int AttachmentWidth { get; }
/// <summary>Colour-attachment height in pixels.</summary>
int AttachmentHeight { get; }
/// <summary>
/// Retail's interior depth clear, scoped to the live pass.
///
/// <para><c>RetailPViewPassExecutor</c> issues <c>glClear(GL_DEPTH_BUFFER_BIT)</c>
/// between the landscape slice and the interior cells. Splitting the pass to
/// get a depth load-op is exactly what the resolve forbids, so the backend
/// records <c>vkCmdClearAttachments</c> instead — reached through here, so the
/// pinned contract stays frozen and the backend-only verb stays inside the
/// backend (plan §5.5.14 item 3).</para>
/// </summary>
void ClearInteriorDepth();
/// <summary>The frame-global sections every renderer in this pass rebinds.</summary>
WorldFrameSections Sections { get; }
}
/// <summary>
/// Campaign V slice V6j: binds the frame-global sections inside a renderer's own
/// draw, which is the only place they can go on Vulkan.
///
/// <para>Each helper falls back to a zeroed ring slice when nothing published the
/// section. That is the same rule the GL arm already states for its own bindings
/// — bind at least one element so the shader never reads an unbound buffer —
/// applied to the three sections whose publisher runs outside the renderer.</para>
/// </summary>
internal static class WorldFrameSectionBinding
{
internal static void BindSceneLighting(
IGpuPassEncoder encoder,
WorldFrameSections sections,
IGpuFrame frame)
{
GpuBufferSection section = sections.SceneLighting;
if (!section.IsValid)
{
section = Zeroed(
frame,
SceneLightingUbo.SizeInBytes,
GpuRingUsage.Uniform);
}
encoder.BindUniformBuffer(
(uint)SceneLightingUbo.BindingPoint,
section.Buffer!,
section.OffsetBytes,
section.SizeBytes);
}
internal static void BindClipRegions(
IGpuPassEncoder encoder,
WorldFrameSections sections,
IGpuFrame frame)
{
GpuBufferSection section = sections.ClipRegions;
if (!section.IsValid)
{
// Slot 0 zeroed is retail's "no clip": count 0 means ungated.
section = Zeroed(
frame,
ClipFrame.CellClipStrideBytes,
GpuRingUsage.Storage);
}
encoder.BindStorageBuffer(
GpuBindingModel.StorageClipRegions,
section.Buffer!,
section.OffsetBytes,
section.SizeBytes);
}
internal static void BindTerrainClip(
IGpuPassEncoder encoder,
WorldFrameSections sections,
IGpuFrame frame)
{
GpuBufferSection section = sections.TerrainClip;
if (!section.IsValid)
{
section = Zeroed(
frame,
ClipFrame.TerrainUboBytes,
GpuRingUsage.Uniform);
}
encoder.BindUniformBuffer(
ClipFrame.TerrainClipUboBinding,
section.Buffer!,
section.OffsetBytes,
section.SizeBytes);
}
private static GpuBufferSection Zeroed(
IGpuFrame frame,
int byteCount,
GpuRingUsage usage)
{
GpuRingAllocation allocation = frame.AllocateRing(byteCount, usage);
allocation.Data.Clear();
return new GpuBufferSection(
allocation.Buffer,
allocation.OffsetBytes,
(uint)byteCount);
}
}