ID3D12CommandQueue 接口 (d3d12.h)

提供用于提交命令列表、同步命令列表执行、检测命令队列和更新资源磁贴映射的方法。

继承

ID3D12CommandQueue 接口继承自 ID3D12PageableID3D12CommandQueue 也有以下类型的成员:

方法

ID3D12CommandQueue 接口具有这些方法。

 
ID3D12CommandQueue::BeginEvent

请不要直接调用。  使用 PIX 事件运行时将事件插入命令队列。 (ID3D12CommandQueue.BeginEvent)
ID3D12CommandQueue::CopyTileMappings

将映射从源保留资源复制到目标保留资源。
ID3D12CommandQueue::EndEvent

请不要直接调用。  使用 PIX 事件运行时将事件插入命令队列。 (ID3D12CommandQueue.EndEvent)
ID3D12CommandQueue::ExecuteCommandLists

提交命令列表数组以供执行。
ID3D12CommandQueue::GetClockCalibration

此方法在同一时间对 CPU 和 GPU 时间戳计数器进行采样。
ID3D12CommandQueue::GetDesc

获取命令队列的说明。
ID3D12CommandQueue::GetTimestampFrequency

此方法用于确定 GPU 时间戳计数器递增的速率。
ID3D12CommandQueue::SetMarker

请不要直接调用。  使用 PIX 事件运行时将事件插入命令队列。 (ID3D12CommandQueue.SetMarker)
ID3D12CommandQueue::Signal

将围栏汇报指定值。
ID3D12CommandQueue::UpdateTileMappings

汇报将预留资源中的磁贴位置映射到资源堆中的内存位置。
ID3D12CommandQueue::Wait

将 GPU 端等待排队,并立即返回。 GPU 端等待是 GPU 等待,直到指定的围栏达到或超过指定值。

备注

使用 ID3D12Device::CreateCommandQueue 创建命令队列对象。

示例

D3D12nBodyGravity 示例使用 ID3D12CommandQueue,如下所示:

头文件声明。

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

异步计算线程。

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;
}

请参阅 D3D12 参考中的示例代码

要求

   
目标平台 Windows
标头 d3d12.h

另请参阅

核心接口

ID3D12Pageable