workshopsoft/ludic
2
0
Fork 0
Code Issues Pull requests Projects Releases 26 Packages Wiki Activity Actions

feat(render3d): compute and indirect draws on Vulkan, and tree layers culled on the GPU (opt-in)

Browse source 
- Device: multiDrawIndirect, drawIndirectFirstInstance and drawIndirectCount where present.
- Buffers carry storage and indirect usage; a GPU-owned buffer is never swapped under a draw.
- Compute programs from shaders/compute.list (binding 0 parameters, 1.. storage buffers),
  built by `ludic-dev shaders`; gpu_compute / gpu_dispatch / gpu_draw_mesh_indirect in gpu.ludic.
- R3D_VK_PROBE=1: a dispatch read back (OK on the RTX 3070 Ti).
- scatter_cull.comp: a tree layer's frustum test and LOD split on the GPU, with the lit, prepass,
  impostor and shadow-LOD draws reading its records. Behind R3D_GPU_CULL=1 and off by default:
  at the camp it is slower (43.0 fps against 53.3), because the frame's cost is per-draw
  descriptor sets and it adds empty-level draws. Validation-clean; OpenGL frames unchanged.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
Orkun ÇAKILKAYA 2026-09-15 14:19:43 +03:00
parent 04cda22d19
commit c1eb5f399f
11 changed files with 463 additions and 15 deletions
Download patch file Download diff file Expand all files Collapse all files

14
packages/ludic.render3d/gpu_vk_res.ludic
View file

@ -561,12 +561,22 @@ function gvk_retire_flush() -> void {
gvk_retired_mem = new []long
}
# A buffer a compute pass writes: a draw later in the same frame reads what the GPU put there,
# so it is never swapped for fresh storage because it was used this frame (gvk_buf_reserve).
var gvk_buf_gpu: []int = null
function gvk_buf_gpu_owned(b: int) -> void {
if gvk_buf_gpu == null { gvk_buf_gpu = new []int }
while len(gvk_buf_gpu) <= b { push(gvk_buf_gpu, 0) }
gvk_buf_gpu[b] = 1
}
function gvk_buf_is_gpu(b: int) -> bool { return gvk_buf_gpu != null and b < len(gvk_buf_gpu) and gvk_buf_gpu[b] == 1 }
# Room for at least n bytes behind handle b; a buffer that is already big enough is kept, so a
# stream re-filled every frame allocates once.
function gvk_buf_reserve(b: int, n: int) -> bool {
if b <= 0 or b >= len(gvk_buf) { return false }
# already big enough, and no draw this frame reads what is there: fill it in place
if gvk_buf[b] != 0 and gvk_buf_size[b] >= n and gvk_buf_used[b] != gvk_frame_no { return true }
if gvk_buf[b] != 0 and gvk_buf_size[b] >= n and (gvk_buf_used[b] != gvk_frame_no or gvk_buf_is_gpu(b)) { return true }
gvk_buf_release(b)
var size = n
if size < 64 { size = 64 }
@ -575,7 +585,7 @@ function gvk_buf_reserve(b: int, n: int) -> bool {
Vk.zero(bci, VkBufferCreateInfo_sizeof)
Vk.put_i32(bci, VkBufferCreateInfo_sType, VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO)
Vk.put_i64(bci, VkBufferCreateInfo_size, size_l)
Vk.put_i32(bci, VkBufferCreateInfo_usage, VK_BUFFER_USAGE_VERTEX_BUFFER_BIT | VK_BUFFER_USAGE_INDEX_BUFFER_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT)
Vk.put_i32(bci, VkBufferCreateInfo_usage, VK_BUFFER_USAGE_VERTEX_BUFFER_BIT | VK_BUFFER_USAGE_INDEX_BUFFER_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT | VK_BUFFER_USAGE_STORAGE_BUFFER_BIT | VK_BUFFER_USAGE_INDIRECT_BUFFER_BIT)
Vk.put_i32(bci, VkBufferCreateInfo_sharingMode, VK_SHARING_MODE_EXCLUSIVE)
let out = bytes(8)
var r = Vk.create_buffer(gvk_dev, bci, null, out)