WickedEngine/WickedEngine/wiFFTGenerator.cpp

#include "wiFFTGenerator.h"
#include "wiResourceManager.h"
#include "wiRenderer.h"
#include "shaders/ShaderInterop_FFTGenerator.h"

#include <stdio.h>
#include <stdlib.h>
#include <assert.h>

using namespace wiGraphics;

namespace wiFFTGenerator
{
	Shader radix008A_CS;
	Shader radix008A_CS2;

#define TWO_PI 6.283185307179586476925286766559

#define FFT_DIMENSIONS 3U
#define FFT_PLAN_SIZE_LIMIT (1U << 27)

#define FFT_FORWARD -1
#define FFT_INVERSE 1

	void radix008A(
		const CSFFT512x512_Plan& fft_plan,
		const GPUResource& pUAV_Dst,
		const GPUResource& pSRV_Src,
		uint32_t thread_count,
		uint32_t istride,
		CommandList cmd)
	{
		// Setup execution configuration
		uint32_t grid = thread_count / COHERENCY_GRANULARITY;

		GraphicsDevice* device = wiRenderer::GetDevice();

		// Buffers
		const GPUResource* srvs[1] = { &pSRV_Src };
		device->BindResources(CS, srvs, TEXSLOT_ONDEMAND0, 1, cmd);

		const GPUResource* uavs[1] = { &pUAV_Dst };
		device->BindUAVs(CS, uavs, 0, arraysize(uavs), cmd);

		// Shader
		if (istride > 1)
		{
			device->BindComputeShader(&radix008A_CS, cmd);
		}
		else
		{
			device->BindComputeShader(&radix008A_CS2, cmd);
		}

		// Execute
		device->Dispatch(grid, 1, 1, cmd);

		GPUBarrier barriers[] = {
			GPUBarrier::Memory(),
		};
		device->Barrier(barriers, arraysize(barriers), cmd);

		// Unbind resource
		device->UnbindResources(TEXSLOT_ONDEMAND0, 1, cmd);
		device->UnbindUAVs(0, 1, cmd);
	}

	void fft_512x512_c2c(
		const CSFFT512x512_Plan& fft_plan,
		const GPUResource& pUAV_Dst,
		const GPUResource& pSRV_Dst,
		const GPUResource& pSRV_Src,
		CommandList cmd)
	{
		const uint32_t thread_count = fft_plan.slices * (512 * 512) / 8;
		GraphicsDevice* device = wiRenderer::GetDevice();
		const GPUBuffer* cs_cbs;

		uint32_t istride = 512 * 512 / 8;
		cs_cbs = &fft_plan.pRadix008A_CB[0];
		device->BindConstantBuffer(CS, cs_cbs, CB_GETBINDSLOT(FFTGeneratorCB), cmd);
		radix008A(fft_plan, fft_plan.pBuffer_Tmp, pSRV_Src, thread_count, istride, cmd);

		istride /= 8;
		cs_cbs = &fft_plan.pRadix008A_CB[1];
		device->BindConstantBuffer(CS, cs_cbs, CB_GETBINDSLOT(FFTGeneratorCB), cmd);
		radix008A(fft_plan, pUAV_Dst, fft_plan.pBuffer_Tmp, thread_count, istride, cmd);

		istride /= 8;
		cs_cbs = &fft_plan.pRadix008A_CB[2];
		device->BindConstantBuffer(CS, cs_cbs, CB_GETBINDSLOT(FFTGeneratorCB), cmd);
		radix008A(fft_plan, fft_plan.pBuffer_Tmp, pSRV_Dst, thread_count, istride, cmd);

		istride /= 8;
		cs_cbs = &fft_plan.pRadix008A_CB[3];
		device->BindConstantBuffer(CS, cs_cbs, CB_GETBINDSLOT(FFTGeneratorCB), cmd);
		radix008A(fft_plan, pUAV_Dst, fft_plan.pBuffer_Tmp, thread_count, istride, cmd);

		istride /= 8;
		cs_cbs = &fft_plan.pRadix008A_CB[4];
		device->BindConstantBuffer(CS, cs_cbs, CB_GETBINDSLOT(FFTGeneratorCB), cmd);
		radix008A(fft_plan, fft_plan.pBuffer_Tmp, pSRV_Dst, thread_count, istride, cmd);

		istride /= 8;
		cs_cbs = &fft_plan.pRadix008A_CB[5];
		device->BindConstantBuffer(CS, cs_cbs, CB_GETBINDSLOT(FFTGeneratorCB), cmd);
		radix008A(fft_plan, pUAV_Dst, fft_plan.pBuffer_Tmp, thread_count, istride, cmd);
	}

	void create_cbuffers_512x512(CSFFT512x512_Plan& plan, GraphicsDevice* device, uint32_t slices)
	{
		// Create 6 cbuffers for 512x512 transform.

		GPUBufferDesc cb_desc;
		cb_desc.Usage = USAGE_IMMUTABLE;
		cb_desc.BindFlags = BIND_CONSTANT_BUFFER;
		cb_desc.CPUAccessFlags = 0;
		cb_desc.MiscFlags = 0;
		cb_desc.ByteWidth = sizeof(FFTGeneratorCB);
		cb_desc.StructureByteStride = 0;

		SubresourceData cb_data;
		cb_data.SysMemPitch = 0;
		cb_data.SysMemSlicePitch = 0;

		// Buffer 0
		const uint32_t thread_count = slices * (512 * 512) / 8;
		uint32_t ostride = 512 * 512 / 8;
		uint32_t istride = ostride;
		double phase_base = -TWO_PI / (512.0 * 512.0);

		FFTGeneratorCB cb_data_buf0 = { thread_count, ostride, istride, 512, (float)phase_base };
		cb_data.pSysMem = &cb_data_buf0;

		device->CreateBuffer(&cb_desc, &cb_data, &plan.pRadix008A_CB[0]);

		// Buffer 1
		istride /= 8;
		phase_base *= 8.0;

		FFTGeneratorCB cb_data_buf1 = { thread_count, ostride, istride, 512, (float)phase_base };
		cb_data.pSysMem = &cb_data_buf1;

		device->CreateBuffer(&cb_desc, &cb_data, &plan.pRadix008A_CB[1]);

		// Buffer 2
		istride /= 8;
		phase_base *= 8.0;

		FFTGeneratorCB cb_data_buf2 = { thread_count, ostride, istride, 512, (float)phase_base };
		cb_data.pSysMem = &cb_data_buf2;

		device->CreateBuffer(&cb_desc, &cb_data, &plan.pRadix008A_CB[2]);

		// Buffer 3
		istride /= 8;
		phase_base *= 8.0;
		ostride /= 512;

		FFTGeneratorCB cb_data_buf3 = { thread_count, ostride, istride, 1, (float)phase_base };
		cb_data.pSysMem = &cb_data_buf3;

		device->CreateBuffer(&cb_desc, &cb_data, &plan.pRadix008A_CB[3]);

		// Buffer 4
		istride /= 8;
		phase_base *= 8.0;

		FFTGeneratorCB cb_data_buf4 = { thread_count, ostride, istride, 1, (float)phase_base };
		cb_data.pSysMem = &cb_data_buf4;

		device->CreateBuffer(&cb_desc, &cb_data, &plan.pRadix008A_CB[4]);

		// Buffer 5
		istride /= 8;
		phase_base *= 8.0;

		FFTGeneratorCB cb_data_buf5 = { thread_count, ostride, istride, 1, (float)phase_base };
		cb_data.pSysMem = &cb_data_buf5;

		device->CreateBuffer(&cb_desc, &cb_data, &plan.pRadix008A_CB[5]);
	}

	void fft512x512_create_plan(CSFFT512x512_Plan& plan, uint32_t slices)
	{
		GraphicsDevice* device = wiRenderer::GetDevice();

		plan.slices = slices;

		// Constants
		// Create 6 cbuffers for 512x512 transform
		create_cbuffers_512x512(plan, device, slices);

		// Temp buffer
		GPUBufferDesc buf_desc;
		buf_desc.ByteWidth = sizeof(float) * 2 * (512 * slices) * 512;
		buf_desc.Usage = USAGE_DEFAULT;
		buf_desc.BindFlags = BIND_UNORDERED_ACCESS | BIND_SHADER_RESOURCE;
		buf_desc.CPUAccessFlags = 0;
		buf_desc.MiscFlags = RESOURCE_MISC_BUFFER_STRUCTURED;
		buf_desc.StructureByteStride = sizeof(float) * 2;

		device->CreateBuffer(&buf_desc, nullptr, &plan.pBuffer_Tmp);
	}

	void LoadShaders()
	{
		std::string path = wiRenderer::GetShaderPath();

		wiRenderer::LoadShader(CS, radix008A_CS, "fft_512x512_c2c_CS.cso");
		wiRenderer::LoadShader(CS, radix008A_CS2, "fft_512x512_c2c_v2_CS.cso");
	}

}