Catcrafts/Crafter.Graphics
2
0
Fork 0
Code Issues 1 Pull requests Projects Releases Packages Wiki Activity Actions

perf(mesh): release per-mesh scratch buffer for static BLAS (#66)

Browse source 
A static (!allowUpdate) mesh can never refit, so its per-mesh DEVICE_LOCAL
scratch buffer is dead weight the moment the build's GPU work completes —
yet it was kept alive for the mesh's whole lifetime. In many-mesh scenes
(where static meshes dominate) this is real, accumulating VRAM waste.

Release it via scratchBuffer.DeferredClear() right after recording a fresh
build for a static mesh. The old "free-after-build is a UAF, the GPU is
still building" hazard is gone: the fence-keyed deferred-deletion queue
(#101/#102) retires the allocation only once the build's frame has passed
its fence. A later Refit on a static mesh takes the rebuild fallback, whose
Resize sees the nulled handle and re-Creates the scratch. Refit-capable
(allowUpdate) meshes keep their scratch as before — an in-place UPDATE
reuses it each frame.

Extends the BLASBuildOptions hardware test to assert the scratch is retained
for allowUpdate meshes and released for static ones (including across a
static-mesh refit rebuild), with the validation layer confirming zero errors.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
catbot 2026-06-17 17:36:10 +00:00
commit 8b008b49d5
2 changed files with 33 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

16
tests/BLASBuildOptions/main.cpp
View file

@ -128,6 +128,10 @@ int main() {
Check((tri.buildFlags & VK_BUILD_ACCELERATION_STRUCTURE_ALLOW_UPDATE_BIT_KHR) != 0,
"allowUpdate=true → ALLOW_UPDATE bit set");
Check(tri.builtPrimitiveCount == 12, "triangle BLAS reports 12 primitives");
// allowUpdate=true keeps the scratch buffer alive so the in-place UPDATE
// refit below can reuse it (build scratch ≥ update scratch, no resize).
Check(tri.scratchBuffer.buffer != VK_NULL_HANDLE,
"allowUpdate=true → scratch retained for refit (issue #66)");
const VkDeviceAddress triAddrBefore = tri.blasAddr;
const VkAccelerationStructureKHR triHandleBefore = tri.accelerationStructure;
@ -157,6 +161,12 @@ int main() {
"FastBuild preference → PREFER_FAST_BUILD bit set");
Check((triFast.buildFlags & VK_BUILD_ACCELERATION_STRUCTURE_ALLOW_UPDATE_BIT_KHR) == 0,
"allowUpdate=false → ALLOW_UPDATE bit clear");
// A static mesh can never refit, so its scratch is released right after
// the build completes rather than kept as persistent VRAM (issue #66).
// DeferredClear nulls the handle immediately (the allocation itself is
// retired by the fence-keyed queue once the build's frame passes).
Check(triFast.scratchBuffer.buffer == VK_NULL_HANDLE,
"allowUpdate=false → scratch released after build (issue #66)");
// Refit without allowUpdate must still succeed via the rebuild fallback.
{
@ -167,6 +177,10 @@ int main() {
SubmitWait(cmd);
}
Check(triFast.blasAddr != 0, "Refit fallback rebuild still produced a valid BLAS");
// The fallback rebuild re-Created the scratch (Resize saw the nulled
// handle) and, still being a static mesh, released it again afterwards.
Check(triFast.scratchBuffer.buffer == VK_NULL_HANDLE,
"static-mesh refit rebuild re-released its scratch (issue #66)");
// ── Procedural (AABB) BLAS: build with options, then refit. ─────────
Mesh proc;
@ -182,6 +196,8 @@ int main() {
}
Check(proc.blasAddr != 0, "procedural Build produced a non-zero blasAddr");
Check(proc.builtPrimitiveCount == 2, "procedural BLAS reports 2 primitives");
Check(proc.scratchBuffer.buffer != VK_NULL_HANDLE,
"procedural allowUpdate=true → scratch retained for refit (issue #66)");
const VkDeviceAddress procAddrBefore = proc.blasAddr;
const VkAccelerationStructureKHR procHandleBefore = proc.accelerationStructure;