blocklayoutHLSL.cpp

Branch :

Show log
Commit
Author : Olli Etuaho
Date : 2015-12-22 15:53:22
Hash : 9b4e8626
Message : Redesign samplers in shaders on D3D11 Translation of samplers to HLSL on D3D11 is changed as follows: Instead of passing around HLSL sampler and HLSL texture references in shaders, all references to ESSL samplers are converted to constant indices within the shader body. Each ESSL sampler is identified by an unique index. In the code generated to implement ESSL texture functions, these indices are used to index arrays of HLSL samplers and HLSL textures to get the sampler and texture to use. HLSL textures and samplers are grouped into arrays by their types. Each unique combination of a HLSL texture type + HLSL sampler type gets its own array. To convert a unique sampler index to an index to one of these arrays, a constant offset is applied. In the most common case of a 2D texture and a regular (non-comparison) sampler, the index offset is always zero and is omitted. The end goal of this refactoring is to make adding extra metadata for samplers easier. The unique sampler index can be used in follow-up changes to index an array of metadata passed in uniforms, which can contain such things as the base level of the texture. This does not solve the issues with samplers in structs. The interface from the point of view of libANGLE is still exactly the same, the only thing that changes is how samplers are handled inside the shader. On feature level 9_3, the D3D compiler has a bug where it can report that the maximum sampler index is exceeded when in fact it is not. This can happen when an array of samplers is declared in the shader. Because of this the new approach can't be used on D3D11 feature level 9_3, but it will continue using the old approach instead. BUG=angleproject:1261 TEST=angle_end2end_tests, dEQP-GLES3.functional.shaders.texture_functions.* (no regressions) dEQP-GLES3.functional.texture.units.* (no regressions) Change-Id: I5fbb0c4280000202dc2795a628b56bd8194ef96f Reviewed-on: https://chromium-review.googlesource.com/320571 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Tryjob-Request: Olli Etuaho <oetuaho@nvidia.com>

src/compiler/translator/blocklayoutHLSL.cpp

//
// Copyright (c) 2013-2014 The ANGLE Project Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
//
// blocklayout.cpp:
//   Implementation for block layout classes and methods.
//

#include "compiler/translator/blocklayoutHLSL.h"

#include "common/mathutil.h"
#include "common/utilities.h"

namespace sh
{

HLSLBlockEncoder::HLSLBlockEncoder(HLSLBlockEncoderStrategy strategy)
    : mEncoderStrategy(strategy),
      mTransposeMatrices(false)
{
}

void HLSLBlockEncoder::enterAggregateType()
{
    nextRegister();
}

void HLSLBlockEncoder::exitAggregateType()
{
}

void HLSLBlockEncoder::getBlockLayoutInfo(GLenum typeIn, unsigned int arraySize, bool isRowMajorMatrix, int *arrayStrideOut, int *matrixStrideOut)
{
    GLenum type = (mTransposeMatrices ? gl::TransposeMatrixType(typeIn) : typeIn);

    // We assume we are only dealing with 4 byte components (no doubles or half-words currently)
    ASSERT(gl::VariableComponentSize(gl::VariableComponentType(type)) == BytesPerComponent);

    int matrixStride = 0;
    int arrayStride = 0;

    // if variables are not to be packed, or we're about to
    // pack a matrix or array, skip to the start of the next
    // register
    if (!isPacked() ||
        gl::IsMatrixType(type) ||
        arraySize > 0)
    {
        nextRegister();
    }

    if (gl::IsMatrixType(type))
    {
        matrixStride = ComponentsPerRegister;

        if (arraySize > 0)
        {
            const int numRegisters = gl::MatrixRegisterCount(type, isRowMajorMatrix);
            arrayStride = ComponentsPerRegister * numRegisters;
        }
    }
    else if (arraySize > 0)
    {
        arrayStride = ComponentsPerRegister;
    }
    else if (isPacked())
    {
        int numComponents = gl::VariableComponentCount(type);
        if ((numComponents + (mCurrentOffset % ComponentsPerRegister)) > ComponentsPerRegister)
        {
            nextRegister();
        }
    }

    *matrixStrideOut = matrixStride;
    *arrayStrideOut = arrayStride;
}

void HLSLBlockEncoder::advanceOffset(GLenum typeIn, unsigned int arraySize, bool isRowMajorMatrix, int arrayStride, int matrixStride)
{
    GLenum type = (mTransposeMatrices ? gl::TransposeMatrixType(typeIn) : typeIn);

    if (arraySize > 0)
    {
        mCurrentOffset += arrayStride * (arraySize - 1);
    }

    if (gl::IsMatrixType(type))
    {
        ASSERT(matrixStride == ComponentsPerRegister);
        const int numRegisters = gl::MatrixRegisterCount(type, isRowMajorMatrix);
        const int numComponents = gl::MatrixComponentCount(type, isRowMajorMatrix);
        mCurrentOffset += ComponentsPerRegister * (numRegisters - 1);
        mCurrentOffset += numComponents;
    }
    else if (isPacked())
    {
        mCurrentOffset += gl::VariableComponentCount(type);
    }
    else
    {
        mCurrentOffset += ComponentsPerRegister;
    }
}

void HLSLBlockEncoder::skipRegisters(unsigned int numRegisters)
{
    mCurrentOffset += (numRegisters * ComponentsPerRegister);
}

HLSLBlockEncoder::HLSLBlockEncoderStrategy HLSLBlockEncoder::GetStrategyFor(ShShaderOutput outputType)
{
    switch (outputType)
    {
        case SH_HLSL_3_0_OUTPUT:
            return ENCODE_LOOSE;
        case SH_HLSL_4_1_OUTPUT:
        case SH_HLSL_4_0_FL9_3_OUTPUT:
            return ENCODE_PACKED;
        default:
            UNREACHABLE();
            return ENCODE_PACKED;
    }
}

template <class ShaderVarType>
void HLSLVariableRegisterCount(const ShaderVarType &variable, HLSLBlockEncoder *encoder)
{
    if (variable.isStruct())
    {
        for (size_t arrayElement = 0; arrayElement < variable.elementCount(); arrayElement++)
        {
            encoder->enterAggregateType();

            for (size_t fieldIndex = 0; fieldIndex < variable.fields.size(); fieldIndex++)
            {
                HLSLVariableRegisterCount(variable.fields[fieldIndex], encoder);
            }

            encoder->exitAggregateType();
        }
    }
    else
    {
        // We operate only on varyings and uniforms, which do not have matrix layout qualifiers
        encoder->encodeType(variable.type, variable.arraySize, false);
    }
}

unsigned int HLSLVariableRegisterCount(const Varying &variable, bool transposeMatrices)
{
    HLSLBlockEncoder encoder(HLSLBlockEncoder::ENCODE_PACKED);
    encoder.setTransposeMatrices(transposeMatrices);
    HLSLVariableRegisterCount(variable, &encoder);

    const size_t registerBytes = (encoder.BytesPerComponent * encoder.ComponentsPerRegister);
    return static_cast<unsigned int>(rx::roundUp<size_t>(encoder.getBlockSize(), registerBytes) / registerBytes);
}

unsigned int HLSLVariableRegisterCount(const Uniform &variable, ShShaderOutput outputType)
{
    HLSLBlockEncoder encoder(HLSLBlockEncoder::GetStrategyFor(outputType));
    encoder.setTransposeMatrices(true);
    HLSLVariableRegisterCount(variable, &encoder);

    const size_t registerBytes = (encoder.BytesPerComponent * encoder.ComponentsPerRegister);
    return static_cast<unsigned int>(rx::roundUp<size_t>(encoder.getBlockSize(), registerBytes) / registerBytes);
}

}

kc3-lang/angle/src/compiler/translator/blocklayoutHLSL.cpp

Commit

kc3-lang/angle /src/compiler/translator/blocklayoutHLSL.cpp