ID3D12CommandQueue インターフェイス (d3d12.h)

コマンド リストの送信、コマンド リストの実行の同期、コマンド キューのインストルメント化、リソース タイル マッピングの更新を行うメソッドを提供します。

継承

ID3D12CommandQueue インターフェイスは ID3D12Pageable から継承されます。 ID3D12CommandQueue には、次の種類のメンバーもあります。

メソッド

ID3D12CommandQueue インターフェイスには、これらのメソッドがあります。

 
ID3D12CommandQueue::BeginEvent

直接呼び出すためのものではありません。  PIX イベント ランタイムを使用して、コマンド キューにイベントを挿入します。 (ID3D12CommandQueue.BeginEvent)
ID3D12CommandQueue::CopyTileMappings

ソース予約リソースから宛先予約リソースにマッピングをコピーします。
ID3D12CommandQueue::EndEvent

直接呼び出すためのものではありません。  PIX イベント ランタイムを使用して、コマンド キューにイベントを挿入します。 (ID3D12CommandQueue.EndEvent)
ID3D12CommandQueue::ExecuteCommandLists

実行するコマンド リストの配列を送信します。
ID3D12CommandQueue::GetClockCalibration

このメソッドは、CPU タイムスタンプ カウンターと GPU タイムスタンプ カウンターを同時にサンプリングします。
ID3D12CommandQueue::GetDesc

コマンド キューの説明を取得します。
ID3D12CommandQueue::GetTimestampFrequency

このメソッドは、GPU タイムスタンプ カウンターがインクリメントされる速度を決定するために使用されます。
ID3D12CommandQueue::SetMarker

直接呼び出すためのものではありません。  PIX イベント ランタイムを使用して、コマンド キューにイベントを挿入します。 (ID3D12CommandQueue.SetMarker)
ID3D12CommandQueue::Signal

指定した値にフェンスを更新します。
ID3D12CommandQueue::UpdateTileMappings

予約済みリソース内のタイルの場所とリソース ヒープ内のメモリの場所のマッピングを更新します。
ID3D12CommandQueue::Wait

GPU 側の待機をキューに入れ、直ちに返します。 GPU 側の待機は、指定されたフェンスが指定された値に達するか、または超えるまで GPU が待機する場所です。

解説

ID3D12Device::CreateCommandQueue を使用して、コマンド キュー オブジェクトを作成します。

D3D12nBodyGravity サンプルでは、次のように ID3D12CommandQueue を使用します。

ヘッダー ファイルの宣言。

// Compute objects.
ComPtr<ID3D12CommandAllocator> m_computeAllocator[ThreadCount];
ComPtr<ID3D12CommandQueue> m_computeCommandQueue[ThreadCount];
ComPtr<ID3D12GraphicsCommandList> m_computeCommandList[ThreadCount];

非同期コンピューティング スレッド。

DWORD D3D12nBodyGravity::AsyncComputeThreadProc(int threadIndex)
{
    ID3D12CommandQueue* pCommandQueue = m_computeCommandQueue[threadIndex].Get();
    ID3D12CommandAllocator* pCommandAllocator = m_computeAllocator[threadIndex].Get();
    ID3D12GraphicsCommandList* pCommandList = m_computeCommandList[threadIndex].Get();
    ID3D12Fence* pFence = m_threadFences[threadIndex].Get();

    while (0 == InterlockedGetValue(&m_terminating))
    {
        // Run the particle simulation.
        Simulate(threadIndex);

        // Close and execute the command list.
        ThrowIfFailed(pCommandList->Close());
        ID3D12CommandList* ppCommandLists[] = { pCommandList };

        pCommandQueue->ExecuteCommandLists(1, ppCommandLists);

        // Wait for the compute shader to complete the simulation.
        UINT64 threadFenceValue = InterlockedIncrement(&m_threadFenceValues[threadIndex]);
        ThrowIfFailed(pCommandQueue->Signal(pFence, threadFenceValue));
        ThrowIfFailed(pFence->SetEventOnCompletion(threadFenceValue, m_threadFenceEvents[threadIndex]));
        WaitForSingleObject(m_threadFenceEvents[threadIndex], INFINITE);

        // Wait for the render thread to be done with the SRV so that
        // the next frame in the simulation can run.
        UINT64 renderContextFenceValue = InterlockedGetValue(&m_renderContextFenceValues[threadIndex]);
        if (m_renderContextFence->GetCompletedValue() < renderContextFenceValue)
        {
            ThrowIfFailed(pCommandQueue->Wait(m_renderContextFence.Get(), renderContextFenceValue));
            InterlockedExchange(&m_renderContextFenceValues[threadIndex], 0);
        }

        // Swap the indices to the SRV and UAV.
        m_srvIndex[threadIndex] = 1 - m_srvIndex[threadIndex];

        // Prepare for the next frame.
        ThrowIfFailed(pCommandAllocator->Reset());
        ThrowIfFailed(pCommandList->Reset(pCommandAllocator, m_computeState.Get()));
    }

    return 0;
}

「D3D12 リファレンス」のコード例を参照してください

要件

   
対象プラットフォーム Windows
ヘッダー d3d12.h

関連項目

コア インターフェイス

ID3D12Pageable