void ResourceManager::AddSRV(ID3D12Resource* tex, D3D12_SHADER_RESOURCE_VIEW_DESC* desc,
D3D12_CPU_DESCRIPTOR_HANDLE& handleCPU, D3D12_GPU_DESCRIPTOR_HANDLE& handleGPU) {
if (m_indexFirstFreeSlotCBVSRVUAV >= m_numCBVSRVUAVs || m_indexFirstFreeSlotCBVSRVUAV < 0) {
throw GFX_Exception("Error adding SRV to CBV/SRV/UAV Heap. No space remaining.");
}
handleCPU = CD3DX12_CPU_DESCRIPTOR_HANDLE(m_pheapCBVSRVUAV->GetCPUDescriptorHandleForHeapStart(),
m_indexFirstFreeSlotCBVSRVUAV, m_sizeCBVSRVUAVHeapDesc);
handleGPU = CD3DX12_GPU_DESCRIPTOR_HANDLE(m_pheapCBVSRVUAV->GetGPUDescriptorHandleForHeapStart(),
m_indexFirstFreeSlotCBVSRVUAV, m_sizeCBVSRVUAVHeapDesc);
m_pDev->CreateSRV(tex, desc, handleCPU);
++m_indexFirstFreeSlotCBVSRVUAV;
}
void D3D12GSRender::flip(int buffer)
{
ID3D12Resource *resource_to_flip;
float viewport_w, viewport_h;
if (!is_flip_surface_in_global_memory(rsx::to_surface_target(rsx::method_registers[NV4097_SET_SURFACE_COLOR_TARGET])))
{
resource_storage &storage = get_current_resource_storage();
assert(storage.ram_framebuffer == nullptr);
size_t w = 0, h = 0, row_pitch = 0;
size_t offset = 0;
if (false)
{
CellGcmDisplayInfo* buffers = nullptr;// = vm::ps3::_ptr<CellGcmDisplayInfo>(m_gcm_buffers_addr);
u32 addr = rsx::get_address(gcm_buffers[gcm_current_buffer].offset, CELL_GCM_LOCATION_LOCAL);
w = gcm_buffers[gcm_current_buffer].width;
h = gcm_buffers[gcm_current_buffer].height;
u8 *src_buffer = vm::ps3::_ptr<u8>(addr);
row_pitch = align(w * 4, 256);
size_t texture_size = row_pitch * h; // * 4 for mipmap levels
size_t heap_offset = m_buffer_data.alloc<D3D12_TEXTURE_DATA_PLACEMENT_ALIGNMENT>(texture_size);
void *mapped_buffer = m_buffer_data.map<void>(heap_offset);
for (unsigned row = 0; row < h; row++)
memcpy((char*)mapped_buffer + row * row_pitch, (char*)src_buffer + row * w * 4, w * 4);
m_buffer_data.unmap(CD3DX12_RANGE(heap_offset, heap_offset + texture_size));
offset = heap_offset;
}
CHECK_HRESULT(
m_device->CreateCommittedResource(
&CD3DX12_HEAP_PROPERTIES(D3D12_HEAP_TYPE_DEFAULT),
D3D12_HEAP_FLAG_NONE,
&CD3DX12_RESOURCE_DESC::Tex2D(DXGI_FORMAT_R8G8B8A8_UNORM, (UINT)w, (UINT)h, 1, 1),
D3D12_RESOURCE_STATE_COPY_DEST,
nullptr,
IID_PPV_ARGS(storage.ram_framebuffer.GetAddressOf())
)
);
get_current_resource_storage().command_list->CopyTextureRegion(&CD3DX12_TEXTURE_COPY_LOCATION(storage.ram_framebuffer.Get(), 0), 0, 0, 0,
&CD3DX12_TEXTURE_COPY_LOCATION(m_buffer_data.get_heap(), { offset, { DXGI_FORMAT_R8G8B8A8_UNORM, (UINT)w, (UINT)h, 1, (UINT)row_pitch } }), nullptr);
get_current_resource_storage().command_list->ResourceBarrier(1, &CD3DX12_RESOURCE_BARRIER::Transition(storage.ram_framebuffer.Get(), D3D12_RESOURCE_STATE_COPY_DEST, D3D12_RESOURCE_STATE_GENERIC_READ));
resource_to_flip = storage.ram_framebuffer.Get();
viewport_w = (float)w, viewport_h = (float)h;
}
else
{
if (std::get<1>(m_rtts.m_bound_render_targets[0]) != nullptr)
{
get_current_resource_storage().command_list->ResourceBarrier(1, &CD3DX12_RESOURCE_BARRIER::Transition(std::get<1>(m_rtts.m_bound_render_targets[0]), D3D12_RESOURCE_STATE_RENDER_TARGET, D3D12_RESOURCE_STATE_GENERIC_READ));
resource_to_flip = std::get<1>(m_rtts.m_bound_render_targets[0]);
}
else if (std::get<1>(m_rtts.m_bound_render_targets[1]) != nullptr)
{
get_current_resource_storage().command_list->ResourceBarrier(1, &CD3DX12_RESOURCE_BARRIER::Transition(std::get<1>(m_rtts.m_bound_render_targets[1]), D3D12_RESOURCE_STATE_RENDER_TARGET, D3D12_RESOURCE_STATE_GENERIC_READ));
resource_to_flip = std::get<1>(m_rtts.m_bound_render_targets[1]);
}
else
resource_to_flip = nullptr;
}
get_current_resource_storage().command_list->ResourceBarrier(1, &CD3DX12_RESOURCE_BARRIER::Transition(m_backbuffer[m_swap_chain->GetCurrentBackBufferIndex()].Get(), D3D12_RESOURCE_STATE_PRESENT, D3D12_RESOURCE_STATE_RENDER_TARGET));
D3D12_VIEWPORT viewport =
{
0.f,
0.f,
(float)m_backbuffer[m_swap_chain->GetCurrentBackBufferIndex()]->GetDesc().Width,
(float)m_backbuffer[m_swap_chain->GetCurrentBackBufferIndex()]->GetDesc().Height,
0.f,
1.f
};
get_current_resource_storage().command_list->RSSetViewports(1, &viewport);
D3D12_RECT box =
{
0,
0,
(LONG)m_backbuffer[m_swap_chain->GetCurrentBackBufferIndex()]->GetDesc().Width,
(LONG)m_backbuffer[m_swap_chain->GetCurrentBackBufferIndex()]->GetDesc().Height,
};
get_current_resource_storage().command_list->RSSetScissorRects(1, &box);
get_current_resource_storage().command_list->SetGraphicsRootSignature(m_output_scaling_pass.root_signature);
get_current_resource_storage().command_list->SetPipelineState(m_output_scaling_pass.pso);
D3D12_SHADER_RESOURCE_VIEW_DESC shader_resource_view_desc = {};
// FIXME: Not always true
shader_resource_view_desc.Format = DXGI_FORMAT_R8G8B8A8_UNORM;
shader_resource_view_desc.ViewDimension = D3D12_SRV_DIMENSION_TEXTURE2D;
shader_resource_view_desc.Texture2D.MipLevels = 1;
if (is_flip_surface_in_global_memory(rsx::to_surface_target(rsx::method_registers[NV4097_SET_SURFACE_COLOR_TARGET])))
shader_resource_view_desc.Shader4ComponentMapping = D3D12_DEFAULT_SHADER_4_COMPONENT_MAPPING;
else
shader_resource_view_desc.Shader4ComponentMapping = D3D12_ENCODE_SHADER_4_COMPONENT_MAPPING(
D3D12_SHADER_COMPONENT_MAPPING_FROM_MEMORY_COMPONENT_1,
D3D12_SHADER_COMPONENT_MAPPING_FROM_MEMORY_COMPONENT_2,
D3D12_SHADER_COMPONENT_MAPPING_FROM_MEMORY_COMPONENT_3,
D3D12_SHADER_COMPONENT_MAPPING_FROM_MEMORY_COMPONENT_0
);
m_device->CreateShaderResourceView(resource_to_flip, &shader_resource_view_desc,
CD3DX12_CPU_DESCRIPTOR_HANDLE(m_output_scaling_pass.texture_descriptor_heap->GetCPUDescriptorHandleForHeapStart()).Offset(m_swap_chain->GetCurrentBackBufferIndex(), m_descriptor_stride_srv_cbv_uav));
D3D12_SAMPLER_DESC sampler_desc = {};
sampler_desc.Filter = D3D12_FILTER_MIN_MAG_LINEAR_MIP_POINT;
sampler_desc.AddressU = D3D12_TEXTURE_ADDRESS_MODE_WRAP;
sampler_desc.AddressV = D3D12_TEXTURE_ADDRESS_MODE_WRAP;
sampler_desc.AddressW = D3D12_TEXTURE_ADDRESS_MODE_WRAP;
m_device->CreateSampler(&sampler_desc,
CD3DX12_CPU_DESCRIPTOR_HANDLE(m_output_scaling_pass.sampler_descriptor_heap->GetCPUDescriptorHandleForHeapStart()).Offset(m_swap_chain->GetCurrentBackBufferIndex(), m_descriptor_stride_samplers));
ID3D12DescriptorHeap *descriptors_heaps[] =
{
m_output_scaling_pass.texture_descriptor_heap,
m_output_scaling_pass.sampler_descriptor_heap
};
get_current_resource_storage().command_list->SetDescriptorHeaps(2, descriptors_heaps);
get_current_resource_storage().command_list->SetGraphicsRootDescriptorTable(0,
CD3DX12_GPU_DESCRIPTOR_HANDLE(m_output_scaling_pass.texture_descriptor_heap->GetGPUDescriptorHandleForHeapStart()).Offset(m_swap_chain->GetCurrentBackBufferIndex(), m_descriptor_stride_srv_cbv_uav));
get_current_resource_storage().command_list->SetGraphicsRootDescriptorTable(1,
CD3DX12_GPU_DESCRIPTOR_HANDLE(m_output_scaling_pass.sampler_descriptor_heap->GetGPUDescriptorHandleForHeapStart()).Offset(m_swap_chain->GetCurrentBackBufferIndex(), m_descriptor_stride_samplers));
get_current_resource_storage().command_list->OMSetRenderTargets(1,
&CD3DX12_CPU_DESCRIPTOR_HANDLE(m_backbuffer_descriptor_heap[m_swap_chain->GetCurrentBackBufferIndex()]->GetCPUDescriptorHandleForHeapStart()),
true, nullptr);
D3D12_VERTEX_BUFFER_VIEW vertex_buffer_view = {};
vertex_buffer_view.BufferLocation = m_output_scaling_pass.vertex_buffer->GetGPUVirtualAddress();
vertex_buffer_view.StrideInBytes = 4 * sizeof(float);
vertex_buffer_view.SizeInBytes = 16 * sizeof(float);
get_current_resource_storage().command_list->IASetVertexBuffers(0, 1, &vertex_buffer_view);
get_current_resource_storage().command_list->IASetPrimitiveTopology(D3D_PRIMITIVE_TOPOLOGY_TRIANGLESTRIP);
if (resource_to_flip)
get_current_resource_storage().command_list->DrawInstanced(4, 1, 0, 0);
if (!rpcs3::config.rsx.d3d12.overlay.value())
get_current_resource_storage().command_list->ResourceBarrier(1, &CD3DX12_RESOURCE_BARRIER::Transition(m_backbuffer[m_swap_chain->GetCurrentBackBufferIndex()].Get(), D3D12_RESOURCE_STATE_RENDER_TARGET, D3D12_RESOURCE_STATE_PRESENT));
if (is_flip_surface_in_global_memory(rsx::to_surface_target(rsx::method_registers[NV4097_SET_SURFACE_COLOR_TARGET])) && resource_to_flip != nullptr)
get_current_resource_storage().command_list->ResourceBarrier(1, &CD3DX12_RESOURCE_BARRIER::Transition(resource_to_flip, D3D12_RESOURCE_STATE_GENERIC_READ, D3D12_RESOURCE_STATE_RENDER_TARGET));
CHECK_HRESULT(get_current_resource_storage().command_list->Close());
m_command_queue->ExecuteCommandLists(1, (ID3D12CommandList**)get_current_resource_storage().command_list.GetAddressOf());
if(rpcs3::config.rsx.d3d12.overlay.value())
render_overlay();
reset_timer();
std::chrono::time_point<std::chrono::system_clock> flip_start = std::chrono::system_clock::now();
CHECK_HRESULT(m_swap_chain->Present(rpcs3::state.config.rsx.vsync.value() ? 1 : 0, 0));
// Add an event signaling queue completion
resource_storage &storage = get_non_current_resource_storage();
m_command_queue->Signal(storage.frame_finished_fence.Get(), storage.fence_value);
storage.frame_finished_fence->SetEventOnCompletion(storage.fence_value, storage.frame_finished_handle);
storage.fence_value++;
storage.in_use = true;
storage.dirty_textures.merge(m_rtts.invalidated_resources);
m_rtts.invalidated_resources.clear();
// Get the put pos - 1. This way after cleaning we can set the get ptr to
// this value, allowing heap to proceed even if we cleant before allocating
// a new value (that's the reason of the -1)
storage.buffer_heap_get_pos = m_buffer_data.get_current_put_pos_minus_one();
storage.readback_heap_get_pos = m_readback_resources.get_current_put_pos_minus_one();
// Now get ready for next frame
resource_storage &new_storage = get_current_resource_storage();
new_storage.wait_and_clean();
if (new_storage.in_use)
{
m_buffer_data.m_get_pos = new_storage.buffer_heap_get_pos;
m_readback_resources.m_get_pos = new_storage.readback_heap_get_pos;
}
m_frame->flip(nullptr);
std::chrono::time_point<std::chrono::system_clock> flip_end = std::chrono::system_clock::now();
m_timers.flip_duration += std::chrono::duration_cast<std::chrono::microseconds>(flip_end - flip_start).count();
}
void D3D12GSRender::end()
{
std::chrono::time_point<std::chrono::system_clock> start_duration = std::chrono::system_clock::now();
std::chrono::time_point<std::chrono::system_clock> rtt_duration_start = std::chrono::system_clock::now();
prepare_render_targets(get_current_resource_storage().command_list.Get());
std::chrono::time_point<std::chrono::system_clock> rtt_duration_end = std::chrono::system_clock::now();
m_timers.prepare_rtt_duration += std::chrono::duration_cast<std::chrono::microseconds>(rtt_duration_end - rtt_duration_start).count();
std::chrono::time_point<std::chrono::system_clock> vertex_index_duration_start = std::chrono::system_clock::now();
size_t currentDescriptorIndex = get_current_resource_storage().descriptors_heap_index;
size_t vertex_count;
bool indexed_draw;
std::vector<D3D12_SHADER_RESOURCE_VIEW_DESC> vertex_buffer_views;
std::tie(indexed_draw, vertex_count, vertex_buffer_views) = upload_and_set_vertex_index_data(get_current_resource_storage().command_list.Get());
size_t vertex_buffer_count = vertex_buffer_views.size();
std::chrono::time_point<std::chrono::system_clock> vertex_index_duration_end = std::chrono::system_clock::now();
m_timers.vertex_index_duration += std::chrono::duration_cast<std::chrono::microseconds>(vertex_index_duration_end - vertex_index_duration_start).count();
std::chrono::time_point<std::chrono::system_clock> program_load_start = std::chrono::system_clock::now();
load_program();
std::chrono::time_point<std::chrono::system_clock> program_load_end = std::chrono::system_clock::now();
m_timers.program_load_duration += std::chrono::duration_cast<std::chrono::microseconds>(program_load_end - program_load_start).count();
get_current_resource_storage().command_list->SetGraphicsRootSignature(m_root_signatures[std::get<2>(m_current_pso)][vertex_buffer_count].Get());
get_current_resource_storage().command_list->OMSetStencilRef(rsx::method_registers[NV4097_SET_STENCIL_FUNC_REF]);
std::chrono::time_point<std::chrono::system_clock> constants_duration_start = std::chrono::system_clock::now();
INT offset = 0;
for (const auto view : vertex_buffer_views)
{
m_device->CreateShaderResourceView(m_vertex_buffer_data.Get(), &view,
CD3DX12_CPU_DESCRIPTOR_HANDLE(get_current_resource_storage().descriptors_heap->GetCPUDescriptorHandleForHeapStart())
.Offset((INT)currentDescriptorIndex + offset++, m_descriptor_stride_srv_cbv_uav));
}
// Constants
upload_and_bind_scale_offset_matrix(currentDescriptorIndex + vertex_buffer_count);
upload_and_bind_vertex_shader_constants(currentDescriptorIndex + 1 + vertex_buffer_count);
upload_and_bind_fragment_shader_constants(currentDescriptorIndex + 2 + vertex_buffer_count);
std::chrono::time_point<std::chrono::system_clock> constants_duration_end = std::chrono::system_clock::now();
m_timers.constants_duration += std::chrono::duration_cast<std::chrono::microseconds>(constants_duration_end - constants_duration_start).count();
get_current_resource_storage().command_list->SetPipelineState(std::get<0>(m_current_pso).Get());
std::chrono::time_point<std::chrono::system_clock> texture_duration_start = std::chrono::system_clock::now();
size_t texture_count = std::get<2>(m_current_pso);
if (texture_count > 0)
{
upload_and_bind_textures(get_current_resource_storage().command_list.Get(), texture_count);
for (unsigned i = 0; i < texture_count; i++)
{
ID3D12Resource *tex_resource;
D3D12_SHADER_RESOURCE_VIEW_DESC srv;
std::tie(tex_resource, srv) = m_current_shader_resources[i];
m_device->CreateShaderResourceView(tex_resource, &srv,
CD3DX12_CPU_DESCRIPTOR_HANDLE(get_current_resource_storage().descriptors_heap->GetCPUDescriptorHandleForHeapStart())
.Offset((INT)currentDescriptorIndex + 3 + (INT)vertex_buffer_count + (INT)i, m_descriptor_stride_srv_cbv_uav)
);
m_device->CreateSampler(&m_current_samplers[i],
CD3DX12_CPU_DESCRIPTOR_HANDLE(get_current_resource_storage().sampler_descriptor_heap[get_current_resource_storage().sampler_descriptors_heap_index]->GetCPUDescriptorHandleForHeapStart())
.Offset((UINT)get_current_resource_storage().current_sampler_index + (UINT)i, m_descriptor_stride_samplers));
}
get_current_resource_storage().command_list->SetGraphicsRootDescriptorTable(0,
CD3DX12_GPU_DESCRIPTOR_HANDLE(get_current_resource_storage().descriptors_heap->GetGPUDescriptorHandleForHeapStart())
.Offset((INT)currentDescriptorIndex, m_descriptor_stride_srv_cbv_uav)
);
get_current_resource_storage().command_list->SetGraphicsRootDescriptorTable(1,
CD3DX12_GPU_DESCRIPTOR_HANDLE(get_current_resource_storage().sampler_descriptor_heap[get_current_resource_storage().sampler_descriptors_heap_index]->GetGPUDescriptorHandleForHeapStart())
.Offset((INT)get_current_resource_storage().current_sampler_index, m_descriptor_stride_samplers)
);
get_current_resource_storage().current_sampler_index += std::get<2>(m_current_pso);
get_current_resource_storage().descriptors_heap_index += std::get<2>(m_current_pso) + 3 + vertex_buffer_count;
}
else
{
get_current_resource_storage().command_list->SetGraphicsRootDescriptorTable(0,
CD3DX12_GPU_DESCRIPTOR_HANDLE(get_current_resource_storage().descriptors_heap->GetGPUDescriptorHandleForHeapStart())
.Offset((INT)currentDescriptorIndex, m_descriptor_stride_srv_cbv_uav)
);
get_current_resource_storage().descriptors_heap_index += 3 + vertex_buffer_count;
}
std::chrono::time_point<std::chrono::system_clock> texture_duration_end = std::chrono::system_clock::now();
m_timers.texture_duration += std::chrono::duration_cast<std::chrono::microseconds>(texture_duration_end - texture_duration_start).count();
set_rtt_and_ds(get_current_resource_storage().command_list.Get());
int clip_w = rsx::method_registers[NV4097_SET_SURFACE_CLIP_HORIZONTAL] >> 16;
int clip_h = rsx::method_registers[NV4097_SET_SURFACE_CLIP_VERTICAL] >> 16;
D3D12_VIEWPORT viewport =
{
0.f,
0.f,
(float)clip_w,
(float)clip_h,
(f32&)rsx::method_registers[NV4097_SET_CLIP_MIN],
(f32&)rsx::method_registers[NV4097_SET_CLIP_MAX]
};
get_current_resource_storage().command_list->RSSetViewports(1, &viewport);
get_current_resource_storage().command_list->RSSetScissorRects(1, &get_scissor(rsx::method_registers[NV4097_SET_SCISSOR_HORIZONTAL], rsx::method_registers[NV4097_SET_SCISSOR_VERTICAL]));
get_current_resource_storage().command_list->IASetPrimitiveTopology(get_primitive_topology(draw_mode));
if (indexed_draw)
get_current_resource_storage().command_list->DrawIndexedInstanced((UINT)vertex_count, 1, 0, 0, 0);
else
get_current_resource_storage().command_list->DrawInstanced((UINT)vertex_count, 1, 0, 0);
std::chrono::time_point<std::chrono::system_clock> end_duration = std::chrono::system_clock::now();
m_timers.draw_calls_duration += std::chrono::duration_cast<std::chrono::microseconds>(end_duration - start_duration).count();
m_timers.draw_calls_count++;
if (rpcs3::config.rsx.d3d12.debug_output.value())
{
CHECK_HRESULT(get_current_resource_storage().command_list->Close());
m_command_queue->ExecuteCommandLists(1, (ID3D12CommandList**)get_current_resource_storage().command_list.GetAddressOf());
get_current_resource_storage().set_new_command_list();
}
thread::end();
}
// Fill the command list with all the render commands and dependent state.
void D3D12ExecuteIndirect::PopulateCommandLists()
{
// Command list allocators can only be reset when the associated
// command lists have finished execution on the GPU; apps should use
// fences to determine GPU execution progress.
ThrowIfFailed(m_computeCommandAllocators[m_frameIndex]->Reset());
ThrowIfFailed(m_commandAllocators[m_frameIndex]->Reset());
// However, when ExecuteCommandList() is called on a particular command
// list, that command list can then be reset at any time and must be before
// re-recording.
ThrowIfFailed(m_computeCommandList->Reset(m_computeCommandAllocators[m_frameIndex].Get(), m_computeState.Get()));
ThrowIfFailed(m_commandList->Reset(m_commandAllocators[m_frameIndex].Get(), m_pipelineState.Get()));
// Record the compute commands that will cull triangles and prevent them from being processed by the vertex shader.
if (m_enableCulling)
{
UINT frameDescriptorOffset = m_frameIndex * CbvSrvUavDescriptorCountPerFrame;
D3D12_GPU_DESCRIPTOR_HANDLE cbvSrvUavHandle = m_cbvSrvUavHeap->GetGPUDescriptorHandleForHeapStart();
m_computeCommandList->SetComputeRootSignature(m_computeRootSignature.Get());
ID3D12DescriptorHeap* ppHeaps[] = { m_cbvSrvUavHeap.Get() };
m_computeCommandList->SetDescriptorHeaps(_countof(ppHeaps), ppHeaps);
m_computeCommandList->SetComputeRootDescriptorTable(
SrvUavTable,
CD3DX12_GPU_DESCRIPTOR_HANDLE(cbvSrvUavHandle, CbvSrvOffset + frameDescriptorOffset, m_cbvSrvUavDescriptorSize));
m_computeCommandList->SetComputeRoot32BitConstants(RootConstants, 4, reinterpret_cast<void*>(&m_csRootConstants), 0);
// Reset the UAV counter for this frame.
m_computeCommandList->CopyBufferRegion(m_processedCommandBuffers[m_frameIndex].Get(), CommandBufferSizePerFrame, m_processedCommandBufferCounterReset.Get(), 0, sizeof(UINT));
D3D12_RESOURCE_BARRIER barrier = CD3DX12_RESOURCE_BARRIER::Transition(m_processedCommandBuffers[m_frameIndex].Get(), D3D12_RESOURCE_STATE_COPY_DEST, D3D12_RESOURCE_STATE_UNORDERED_ACCESS);
m_computeCommandList->ResourceBarrier(1, &barrier);
m_computeCommandList->Dispatch(static_cast<UINT>(ceil(TriangleCount / float(ComputeThreadBlockSize))), 1, 1);
}
ThrowIfFailed(m_computeCommandList->Close());
// Record the rendering commands.
{
// Set necessary state.
m_commandList->SetGraphicsRootSignature(m_rootSignature.Get());
ID3D12DescriptorHeap* ppHeaps[] = { m_cbvSrvUavHeap.Get() };
m_commandList->SetDescriptorHeaps(_countof(ppHeaps), ppHeaps);
m_commandList->RSSetViewports(1, &m_viewport);
m_commandList->RSSetScissorRects(1, m_enableCulling ? &m_cullingScissorRect : &m_scissorRect);
// Indicate that the command buffer will be used for indirect drawing
// and that the back buffer will be used as a render target.
D3D12_RESOURCE_BARRIER barriers[2] = {
CD3DX12_RESOURCE_BARRIER::Transition(
m_enableCulling ? m_processedCommandBuffers[m_frameIndex].Get() : m_commandBuffer.Get(),
m_enableCulling ? D3D12_RESOURCE_STATE_UNORDERED_ACCESS : D3D12_RESOURCE_STATE_NON_PIXEL_SHADER_RESOURCE,
D3D12_RESOURCE_STATE_INDIRECT_ARGUMENT),
CD3DX12_RESOURCE_BARRIER::Transition(
m_renderTargets[m_frameIndex].Get(),
D3D12_RESOURCE_STATE_PRESENT,
D3D12_RESOURCE_STATE_RENDER_TARGET)
};
m_commandList->ResourceBarrier(_countof(barriers), barriers);
CD3DX12_CPU_DESCRIPTOR_HANDLE rtvHandle(m_rtvHeap->GetCPUDescriptorHandleForHeapStart(), m_frameIndex, m_rtvDescriptorSize);
CD3DX12_CPU_DESCRIPTOR_HANDLE dsvHandle(m_dsvHeap->GetCPUDescriptorHandleForHeapStart());
m_commandList->OMSetRenderTargets(1, &rtvHandle, FALSE, &dsvHandle);
// Record commands.
const float clearColor[] = { 0.0f, 0.2f, 0.4f, 1.0f };
m_commandList->ClearRenderTargetView(rtvHandle, clearColor, 0, nullptr);
m_commandList->ClearDepthStencilView(dsvHandle, D3D12_CLEAR_FLAG_DEPTH, 1.0f, 0, 0, nullptr);
m_commandList->IASetPrimitiveTopology(D3D_PRIMITIVE_TOPOLOGY_TRIANGLESTRIP);
m_commandList->IASetVertexBuffers(0, 1, &m_vertexBufferView);
if (m_enableCulling)
{
// Draw the triangles that have not been culled.
m_commandList->ExecuteIndirect(
m_commandSignature.Get(),
TriangleCount,
m_processedCommandBuffers[m_frameIndex].Get(),
0,
m_processedCommandBuffers[m_frameIndex].Get(),
CommandBufferSizePerFrame);
}
else
{
// Draw all of the triangles.
m_commandList->ExecuteIndirect(
m_commandSignature.Get(),
TriangleCount,
m_commandBuffer.Get(),
CommandBufferSizePerFrame * m_frameIndex,
nullptr,
0);
}
// Indicate that the command buffer may be used by the compute shader
// and that the back buffer will now be used to present.
barriers[0].Transition.StateBefore = D3D12_RESOURCE_STATE_INDIRECT_ARGUMENT;
barriers[0].Transition.StateAfter = m_enableCulling ? D3D12_RESOURCE_STATE_COPY_DEST : D3D12_RESOURCE_STATE_NON_PIXEL_SHADER_RESOURCE;
barriers[1].Transition.StateBefore = D3D12_RESOURCE_STATE_RENDER_TARGET;
barriers[1].Transition.StateAfter = D3D12_RESOURCE_STATE_PRESENT;
m_commandList->ResourceBarrier(_countof(barriers), barriers);
ThrowIfFailed(m_commandList->Close());
}
}
D3D12_GPU_DESCRIPTOR_HANDLE Direct3D12Heap::getGpuHandle(Size offset) const {
return CD3DX12_GPU_DESCRIPTOR_HANDLE(gpuHandle).Offset(offset, handleIncrement);
}
void D3D12GSRender::end()
{
std::chrono::time_point<steady_clock> start_duration = steady_clock::now();
std::chrono::time_point<steady_clock> rtt_duration_start = steady_clock::now();
prepare_render_targets(get_current_resource_storage().command_list.Get());
std::chrono::time_point<steady_clock> rtt_duration_end = steady_clock::now();
m_timers.prepare_rtt_duration += std::chrono::duration_cast<std::chrono::microseconds>(rtt_duration_end - rtt_duration_start).count();
std::chrono::time_point<steady_clock> vertex_index_duration_start = steady_clock::now();
size_t currentDescriptorIndex = get_current_resource_storage().descriptors_heap_index;
size_t vertex_count;
bool indexed_draw;
std::vector<D3D12_SHADER_RESOURCE_VIEW_DESC> vertex_buffer_views;
std::tie(indexed_draw, vertex_count, vertex_buffer_views) = upload_and_set_vertex_index_data(get_current_resource_storage().command_list.Get());
UINT vertex_buffer_count = static_cast<UINT>(vertex_buffer_views.size());
std::chrono::time_point<steady_clock> vertex_index_duration_end = steady_clock::now();
m_timers.vertex_index_duration += std::chrono::duration_cast<std::chrono::microseconds>(vertex_index_duration_end - vertex_index_duration_start).count();
std::chrono::time_point<steady_clock> program_load_start = steady_clock::now();
load_program();
std::chrono::time_point<steady_clock> program_load_end = steady_clock::now();
m_timers.program_load_duration += std::chrono::duration_cast<std::chrono::microseconds>(program_load_end - program_load_start).count();
get_current_resource_storage().command_list->SetGraphicsRootSignature(m_shared_root_signature.Get());
get_current_resource_storage().command_list->OMSetStencilRef(rsx::method_registers.stencil_func_ref());
std::chrono::time_point<steady_clock> constants_duration_start = steady_clock::now();
INT offset = 0;
for (const auto view : vertex_buffer_views)
{
m_device->CreateShaderResourceView(m_vertex_buffer_data.Get(), &view,
CD3DX12_CPU_DESCRIPTOR_HANDLE(get_current_resource_storage().descriptors_heap->GetCPUDescriptorHandleForHeapStart())
.Offset((INT)currentDescriptorIndex + offset++, m_descriptor_stride_srv_cbv_uav));
}
// Bind vertex buffer
get_current_resource_storage().command_list->SetGraphicsRootDescriptorTable(VERTEX_BUFFERS_SLOT,
CD3DX12_GPU_DESCRIPTOR_HANDLE(get_current_resource_storage().descriptors_heap->GetGPUDescriptorHandleForHeapStart())
.Offset((INT)currentDescriptorIndex, m_descriptor_stride_srv_cbv_uav)
);
// Constants
const D3D12_CONSTANT_BUFFER_VIEW_DESC &fragment_constant_view = upload_fragment_shader_constants();
get_current_resource_storage().command_list->SetGraphicsRootConstantBufferView(FRAGMENT_CONSTANT_BUFFERS_SLOT, fragment_constant_view.BufferLocation);
upload_and_bind_scale_offset_matrix(currentDescriptorIndex + vertex_buffer_count);
get_current_resource_storage().command_list->SetGraphicsRootDescriptorTable(SCALE_OFFSET_SLOT,
CD3DX12_GPU_DESCRIPTOR_HANDLE(get_current_resource_storage().descriptors_heap->GetGPUDescriptorHandleForHeapStart())
.Offset((INT)currentDescriptorIndex + vertex_buffer_count, m_descriptor_stride_srv_cbv_uav)
);
if (m_transform_constants_dirty && !g_cfg.video.debug_output)
{
m_current_transform_constants_buffer_descriptor_id = (u32)currentDescriptorIndex + 1 + vertex_buffer_count;
upload_and_bind_vertex_shader_constants(currentDescriptorIndex + 1 + vertex_buffer_count);
m_transform_constants_dirty = false;
get_current_resource_storage().command_list->SetGraphicsRootDescriptorTable(VERTEX_CONSTANT_BUFFERS_SLOT,
CD3DX12_GPU_DESCRIPTOR_HANDLE(get_current_resource_storage().descriptors_heap->GetGPUDescriptorHandleForHeapStart())
.Offset(m_current_transform_constants_buffer_descriptor_id, m_descriptor_stride_srv_cbv_uav)
);
}
std::chrono::time_point<steady_clock> constants_duration_end = steady_clock::now();
m_timers.constants_duration += std::chrono::duration_cast<std::chrono::microseconds>(constants_duration_end - constants_duration_start).count();
get_current_resource_storage().command_list->SetPipelineState(std::get<0>(m_current_pso).Get());
std::chrono::time_point<steady_clock> texture_duration_start = steady_clock::now();
get_current_resource_storage().descriptors_heap_index += 2 + vertex_buffer_count;
size_t texture_count = std::get<2>(m_current_pso);
if (texture_count > 0)
{
if (get_current_resource_storage().current_sampler_index + 16 > 2048)
{
get_current_resource_storage().sampler_descriptors_heap_index = 1;
get_current_resource_storage().current_sampler_index = 0;
ID3D12DescriptorHeap *descriptors[] =
{
get_current_resource_storage().descriptors_heap.Get(),
get_current_resource_storage().sampler_descriptor_heap[get_current_resource_storage().sampler_descriptors_heap_index].Get(),
};
get_current_resource_storage().command_list->SetDescriptorHeaps(2, descriptors);
}
upload_textures(get_current_resource_storage().command_list.Get(), texture_count);
m_device->CopyDescriptorsSimple(16,
CD3DX12_CPU_DESCRIPTOR_HANDLE(get_current_resource_storage().descriptors_heap->GetCPUDescriptorHandleForHeapStart())
.Offset((UINT)get_current_resource_storage().descriptors_heap_index, m_descriptor_stride_srv_cbv_uav),
CD3DX12_CPU_DESCRIPTOR_HANDLE(m_current_texture_descriptors->GetCPUDescriptorHandleForHeapStart()),
D3D12_DESCRIPTOR_HEAP_TYPE_CBV_SRV_UAV
);
m_device->CopyDescriptorsSimple(16,
CD3DX12_CPU_DESCRIPTOR_HANDLE(get_current_resource_storage().sampler_descriptor_heap[get_current_resource_storage().sampler_descriptors_heap_index]->GetCPUDescriptorHandleForHeapStart())
.Offset((UINT)get_current_resource_storage().current_sampler_index, m_descriptor_stride_samplers),
CD3DX12_CPU_DESCRIPTOR_HANDLE(m_current_sampler_descriptors->GetCPUDescriptorHandleForHeapStart()),
D3D12_DESCRIPTOR_HEAP_TYPE_SAMPLER
);
get_current_resource_storage().command_list->SetGraphicsRootDescriptorTable(TEXTURES_SLOT,
CD3DX12_GPU_DESCRIPTOR_HANDLE(get_current_resource_storage().descriptors_heap->GetGPUDescriptorHandleForHeapStart())
.Offset((INT)get_current_resource_storage().descriptors_heap_index, m_descriptor_stride_srv_cbv_uav)
);
get_current_resource_storage().command_list->SetGraphicsRootDescriptorTable(SAMPLERS_SLOT,
CD3DX12_GPU_DESCRIPTOR_HANDLE(get_current_resource_storage().sampler_descriptor_heap[get_current_resource_storage().sampler_descriptors_heap_index]->GetGPUDescriptorHandleForHeapStart())
.Offset((INT)get_current_resource_storage().current_sampler_index, m_descriptor_stride_samplers)
);
get_current_resource_storage().current_sampler_index += texture_count;
get_current_resource_storage().descriptors_heap_index += texture_count;
}
std::chrono::time_point<steady_clock> texture_duration_end = steady_clock::now();
m_timers.texture_duration += std::chrono::duration_cast<std::chrono::microseconds>(texture_duration_end - texture_duration_start).count();
set_rtt_and_ds(get_current_resource_storage().command_list.Get());
int clip_w = rsx::method_registers.surface_clip_width();
int clip_h = rsx::method_registers.surface_clip_height();
D3D12_VIEWPORT viewport =
{
0.f,
0.f,
(float)clip_w,
(float)clip_h,
rsx::method_registers.clip_min(),
rsx::method_registers.clip_max(),
};
get_current_resource_storage().command_list->RSSetViewports(1, &viewport);
get_current_resource_storage().command_list->RSSetScissorRects(1, &get_scissor(rsx::method_registers.scissor_origin_x(), rsx::method_registers.scissor_origin_y(),
rsx::method_registers.scissor_width(), rsx::method_registers.scissor_height()));
get_current_resource_storage().command_list->IASetPrimitiveTopology(get_primitive_topology(rsx::method_registers.current_draw_clause.primitive));
if (indexed_draw)
get_current_resource_storage().command_list->DrawIndexedInstanced((UINT)vertex_count, 1, 0, 0, 0);
else
get_current_resource_storage().command_list->DrawInstanced((UINT)vertex_count, 1, 0, 0);
std::chrono::time_point<steady_clock> end_duration = steady_clock::now();
m_timers.draw_calls_duration += std::chrono::duration_cast<std::chrono::microseconds>(end_duration - start_duration).count();
m_timers.draw_calls_count++;
if (g_cfg.video.debug_output)
{
CHECK_HRESULT(get_current_resource_storage().command_list->Close());
m_command_queue->ExecuteCommandLists(1, (ID3D12CommandList**)get_current_resource_storage().command_list.GetAddressOf());
get_current_resource_storage().set_new_command_list();
}
thread::end();
}
void D3D12GSRender::end()
{
std::chrono::time_point<std::chrono::system_clock> start_duration = std::chrono::system_clock::now();
std::chrono::time_point<std::chrono::system_clock> rtt_duration_start = std::chrono::system_clock::now();
prepare_render_targets(getCurrentResourceStorage().command_list.Get());
std::chrono::time_point<std::chrono::system_clock> rtt_duration_end = std::chrono::system_clock::now();
m_timers.m_rttDuration += std::chrono::duration_cast<std::chrono::microseconds>(rtt_duration_end - rtt_duration_start).count();
std::chrono::time_point<std::chrono::system_clock> vertex_index_duration_start = std::chrono::system_clock::now();
if (!vertex_index_array.empty() || vertex_draw_count)
upload_and_set_vertex_index_data(getCurrentResourceStorage().command_list.Get());
std::chrono::time_point<std::chrono::system_clock> vertex_index_duration_end = std::chrono::system_clock::now();
m_timers.m_vertexIndexDuration += std::chrono::duration_cast<std::chrono::microseconds>(vertex_index_duration_end - vertex_index_duration_start).count();
std::chrono::time_point<std::chrono::system_clock> program_load_start = std::chrono::system_clock::now();
if (!load_program())
{
LOG_ERROR(RSX, "LoadProgram failed.");
Emu.Pause();
return;
}
std::chrono::time_point<std::chrono::system_clock> program_load_end = std::chrono::system_clock::now();
m_timers.m_programLoadDuration += std::chrono::duration_cast<std::chrono::microseconds>(program_load_end - program_load_start).count();
getCurrentResourceStorage().command_list->SetGraphicsRootSignature(m_rootSignatures[std::get<2>(*m_PSO)].Get());
getCurrentResourceStorage().command_list->OMSetStencilRef(rsx::method_registers[NV4097_SET_STENCIL_FUNC_REF]);
std::chrono::time_point<std::chrono::system_clock> constants_duration_start = std::chrono::system_clock::now();
size_t currentDescriptorIndex = getCurrentResourceStorage().descriptors_heap_index;
// Constants
upload_and_bind_scale_offset_matrix(currentDescriptorIndex);
upload_and_bind_vertex_shader_constants(currentDescriptorIndex + 1);
upload_and_bind_fragment_shader_constants(currentDescriptorIndex + 2);
std::chrono::time_point<std::chrono::system_clock> constants_duration_end = std::chrono::system_clock::now();
m_timers.m_constantsDuration += std::chrono::duration_cast<std::chrono::microseconds>(constants_duration_end - constants_duration_start).count();
getCurrentResourceStorage().command_list->SetPipelineState(std::get<0>(*m_PSO));
std::chrono::time_point<std::chrono::system_clock> texture_duration_start = std::chrono::system_clock::now();
if (std::get<2>(*m_PSO) > 0)
{
upload_and_bind_textures(getCurrentResourceStorage().command_list.Get(), currentDescriptorIndex + 3, std::get<2>(*m_PSO) > 0);
ID3D12DescriptorHeap *descriptors[] =
{
getCurrentResourceStorage().descriptors_heap.Get(),
getCurrentResourceStorage().sampler_descriptor_heap[getCurrentResourceStorage().sampler_descriptors_heap_index].Get(),
};
getCurrentResourceStorage().command_list->SetDescriptorHeaps(2, descriptors);
getCurrentResourceStorage().command_list->SetGraphicsRootDescriptorTable(0,
CD3DX12_GPU_DESCRIPTOR_HANDLE(getCurrentResourceStorage().descriptors_heap->GetGPUDescriptorHandleForHeapStart())
.Offset((INT)currentDescriptorIndex, g_descriptorStrideSRVCBVUAV)
);
getCurrentResourceStorage().command_list->SetGraphicsRootDescriptorTable(1,
CD3DX12_GPU_DESCRIPTOR_HANDLE(getCurrentResourceStorage().sampler_descriptor_heap[getCurrentResourceStorage().sampler_descriptors_heap_index]->GetGPUDescriptorHandleForHeapStart())
.Offset((INT)getCurrentResourceStorage().current_sampler_index, g_descriptorStrideSamplers)
);
getCurrentResourceStorage().current_sampler_index += std::get<2>(*m_PSO);
getCurrentResourceStorage().descriptors_heap_index += std::get<2>(*m_PSO) + 3;
}
else
{
getCurrentResourceStorage().command_list->SetDescriptorHeaps(1, getCurrentResourceStorage().descriptors_heap.GetAddressOf());
getCurrentResourceStorage().command_list->SetGraphicsRootDescriptorTable(0,
CD3DX12_GPU_DESCRIPTOR_HANDLE(getCurrentResourceStorage().descriptors_heap->GetGPUDescriptorHandleForHeapStart())
.Offset((INT)currentDescriptorIndex, g_descriptorStrideSRVCBVUAV)
);
getCurrentResourceStorage().descriptors_heap_index += 3;
}
std::chrono::time_point<std::chrono::system_clock> texture_duration_end = std::chrono::system_clock::now();
m_timers.m_textureDuration += std::chrono::duration_cast<std::chrono::microseconds>(texture_duration_end - texture_duration_start).count();
set_rtt_and_ds(getCurrentResourceStorage().command_list.Get());
int clip_w = rsx::method_registers[NV4097_SET_SURFACE_CLIP_HORIZONTAL] >> 16;
int clip_h = rsx::method_registers[NV4097_SET_SURFACE_CLIP_VERTICAL] >> 16;
D3D12_VIEWPORT viewport =
{
0.f,
0.f,
(float)clip_w,
(float)clip_h,
-1.f,
1.f
};
getCurrentResourceStorage().command_list->RSSetViewports(1, &viewport);
D3D12_RECT box =
{
0,
0,
(LONG)clip_w,
(LONG)clip_h,
};
getCurrentResourceStorage().command_list->RSSetScissorRects(1, &box);
getCurrentResourceStorage().command_list->IASetPrimitiveTopology(get_primitive_topology(draw_mode));
if (m_renderingInfo.m_indexed)
getCurrentResourceStorage().command_list->DrawIndexedInstanced((UINT)m_renderingInfo.m_count, 1, 0, 0, 0);
else
getCurrentResourceStorage().command_list->DrawInstanced((UINT)m_renderingInfo.m_count, 1, 0, 0);
vertex_index_array.clear();
std::chrono::time_point<std::chrono::system_clock> end_duration = std::chrono::system_clock::now();
m_timers.m_drawCallDuration += std::chrono::duration_cast<std::chrono::microseconds>(end_duration - start_duration).count();
m_timers.m_drawCallCount++;
if (rpcs3::config.rsx.d3d12.debug_output.value())
{
ThrowIfFailed(getCurrentResourceStorage().command_list->Close());
m_commandQueueGraphic->ExecuteCommandLists(1, (ID3D12CommandList**)getCurrentResourceStorage().command_list.GetAddressOf());
getCurrentResourceStorage().set_new_command_list();
}
m_first_count_pairs.clear();
m_renderingInfo.m_indexed = false;
thread::end();
}