encounter/dawn-cmake
1
0
Fork 0
mirror of https://github.com/encounter/dawn-cmake.git synced 2025-12-15 08:06:19 +00:00
Code Issues Packages Projects Releases Wiki Activity

Remove deferred BufferLocation updates for drawIndexedIndirect

Browse Source 
Instead of using BufferLocation as another layer of indirection,
the indirectBuffer can be set directly on the indirect command.
This makes the indirect validation a bit simpler, but introduces
additional lifetime dependencies in that the indirect draw validation
MUST be encoded while the DrawIndexedIndirectCmds it references
are still valid.

Bug: dawn:809
Change-Id: I1ef084622d8737ad5ec1b0247bf9062712e35008
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/67241
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
This commit is contained in:
Austin Eng
2021-11-02 18:23:49 +00:00
committed by Dawn LUCI CQ
parent d4b9cda056
commit 6abf1a1adb
18 changed files with 45 additions and 218 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
src/dawn_native/IndirectDrawValidationEncoder.cpp
View File

@@ -226,7 +226,6 @@ namespace dawn_native {
// single pass as possible. Batches can be grouped together as long as they're validating
// data from the same indirect buffer, but they may still be split into multiple passes if
// the number of draw calls in a pass would exceed some (very high) upper bound.
uint64_t numTotalDrawCalls = 0;
size_t validatedParamsSize = 0;
std::vector<Pass> passes;
IndirectDrawMetadata::IndexedIndirectBufferValidationInfoMap& bufferInfoMap =
@@ -257,7 +256,6 @@ namespace dawn_native {
newBatch.clientIndirectOffset = minOffsetAlignedDown;
newBatch.clientIndirectSize =
batch.maxOffset + kDrawIndexedIndirectSize - minOffsetAlignedDown;
numTotalDrawCalls += batch.draws.size();
newBatch.validatedParamsSize = batch.draws.size() * kDrawIndexedIndirectSize;
newBatch.validatedParamsOffset =
@@ -306,10 +304,7 @@ namespace dawn_native {
DAWN_TRY(validatedParamsBuffer.EnsureCapacity(validatedParamsSize));
usageTracker->BufferUsedAs(validatedParamsBuffer.GetBuffer(), wgpu::BufferUsage::Indirect);
// Now we allocate and populate host-side batch data to be copied to the GPU, and prepare to
// update all DrawIndexedIndirectCmd buffer references.
std::vector<DeferredBufferLocationUpdate> deferredBufferLocationUpdates;
deferredBufferLocationUpdates.reserve(numTotalDrawCalls);
// Now we allocate and populate host-side batch data to be copied to the GPU.
for (Pass& pass : passes) {
// We use std::malloc here because it guarantees maximal scalar alignment.
pass.batchData = {std::malloc(pass.batchDataSize), std::free};
@@ -322,16 +317,13 @@ namespace dawn_native {
uint32_t* indirectOffsets = reinterpret_cast<uint32_t*>(batch.batchInfo + 1);
uint64_t validatedParamsOffset = batch.validatedParamsOffset;
for (const auto& draw : batch.metadata->draws) {
for (auto& draw : batch.metadata->draws) {
// The shader uses this to index an array of u32, hence the division by 4 bytes.
*indirectOffsets++ = static_cast<uint32_t>(
(draw.clientBufferOffset - batch.clientIndirectOffset) / 4);
DeferredBufferLocationUpdate deferredUpdate;
deferredUpdate.location = draw.bufferLocation;
deferredUpdate.buffer = validatedParamsBuffer.GetBuffer();
deferredUpdate.offset = validatedParamsOffset;
deferredBufferLocationUpdates.push_back(std::move(deferredUpdate));
draw.cmd->indirectBuffer = validatedParamsBuffer.GetBuffer();
draw.cmd->indirectOffset = validatedParamsOffset;
validatedParamsOffset += kDrawIndexedIndirectSize;
}
@@ -364,8 +356,6 @@ namespace dawn_native {
// Finally, we can now encode our validation passes. Each pass first does a single
// WriteBuffer to get batch data over to the GPU, followed by a single compute pass. The
// compute pass encodes a separate SetBindGroup and Dispatch command for each batch.
commandEncoder->EncodeSetValidatedBufferLocationsInternal(
std::move(deferredBufferLocationUpdates));
for (const Pass& pass : passes) {
commandEncoder->APIWriteBuffer(batchDataBuffer.GetBuffer(), 0,
static_cast<const uint8_t*>(pass.batchData.get()),