d3d11: Cache the staging buffer for WriteInternal
The overhead of creating a new staging buffer for each write is not
trivial. This caches a handy staging buffer in DeviceD3D11 to eliminate
the overhead.
Bug: chromium:1485789
Change-Id: Iaced53ffdb080b07dddf59f9bf54f05b5f2a7a41
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/167180
Reviewed-by: Peng Huang <penghuang@chromium.org>
Reviewed-by: Peng Huang <penghuang@google.com>
Commit-Queue: Jie A Chen <jie.a.chen@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
diff --git a/src/dawn/native/d3d11/BufferD3D11.cpp b/src/dawn/native/d3d11/BufferD3D11.cpp
index 1e79f14..50eb886 100644
--- a/src/dawn/native/d3d11/BufferD3D11.cpp
+++ b/src/dawn/native/d3d11/BufferD3D11.cpp
@@ -591,14 +591,8 @@
// If the mD3d11NonConstantBuffer is null, we have to create a staging buffer for transfer the
// data to mD3d11ConstantBuffer.
- BufferDescriptor descriptor;
- descriptor.usage = wgpu::BufferUsage::MapWrite | wgpu::BufferUsage::CopySrc;
- descriptor.size = Align(size, D3D11BufferSizeAlignment(descriptor.usage));
- descriptor.mappedAtCreation = false;
- descriptor.label = "DawnWriteStagingBuffer";
Ref<BufferBase> stagingBuffer;
- DAWN_TRY_ASSIGN(stagingBuffer,
- Buffer::Create(ToBackend(GetDevice()), Unpack(&descriptor), commandContext));
+ DAWN_TRY_ASSIGN(stagingBuffer, ToBackend(GetDevice())->GetStagingBuffer(commandContext, size));
DAWN_TRY(ToBackend(stagingBuffer)->WriteInternal(commandContext, 0, data, size));
diff --git a/src/dawn/native/d3d11/DeviceD3D11.cpp b/src/dawn/native/d3d11/DeviceD3D11.cpp
index ebdf352..2c5c3d1 100644
--- a/src/dawn/native/d3d11/DeviceD3D11.cpp
+++ b/src/dawn/native/d3d11/DeviceD3D11.cpp
@@ -533,4 +533,30 @@
return mImplicitPixelLocalStorageAttachmentTextureViews[implicitAttachmentIndex].Get();
}
+ResultOrError<Ref<BufferBase>> Device::GetStagingBuffer(
+ const ScopedCommandRecordingContext* commandContext,
+ uint64_t size) {
+ constexpr uint64_t kMinSize = 4 * 1024;
+ constexpr uint64_t kMaxSize = 16 * 1024 * 1024;
+ uint64_t bufferSize = mStagingBuffer.Get() ? mStagingBuffer->GetSize() : 0;
+ if (size > bufferSize) {
+ bufferSize = Align(size, kMinSize);
+ BufferDescriptor descriptor;
+ descriptor.usage = wgpu::BufferUsage::MapWrite | wgpu::BufferUsage::CopySrc;
+ descriptor.size = bufferSize;
+ descriptor.mappedAtCreation = false;
+ descriptor.label = "DawnDeviceStagingBuffer";
+ Ref<BufferBase> buffer;
+ DAWN_TRY_ASSIGN(buffer, Buffer::Create(this, Unpack(&descriptor), commandContext));
+ // We don't cache the buffer if it's too large.
+ if (bufferSize > kMaxSize) {
+ return buffer;
+ }
+ mStagingBuffer = buffer;
+ }
+ // Ensure there is no more than 1 active usage of the staging buffer.
+ DAWN_ASSERT(mStagingBuffer->GetRefCountForTesting() <= 1);
+ return mStagingBuffer;
+}
+
} // namespace dawn::native::d3d11
diff --git a/src/dawn/native/d3d11/DeviceD3D11.h b/src/dawn/native/d3d11/DeviceD3D11.h
index 7815c97..6aacd73 100644
--- a/src/dawn/native/d3d11/DeviceD3D11.h
+++ b/src/dawn/native/d3d11/DeviceD3D11.h
@@ -101,6 +101,13 @@
uint32_t height,
uint32_t implicitAttachmentIndex);
+ // Grab a staging buffer, the size of which is no less than 'size'.
+ // Note: We assume only 1 staging buffer is active, so the client should release it as soon as
+ // possbile once the buffer usage is done.
+ ResultOrError<Ref<BufferBase>> GetStagingBuffer(
+ const ScopedCommandRecordingContext* commandContext,
+ uint64_t size);
+
private:
using Base = d3d::Device;
using Base::Base;
@@ -157,6 +164,9 @@
// TODO(dawn:1704): decide when to clear the cached implicit pixel local storage attachments.
std::array<Ref<TextureViewBase>, kMaxPLSSlots> mImplicitPixelLocalStorageAttachmentTextureViews;
+
+ // The cached staging buffer.
+ Ref<BufferBase> mStagingBuffer;
};
} // namespace dawn::native::d3d11
