CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
136 lines
4.5 KiB
C++
136 lines
4.5 KiB
C++
// Copyright Advanced Micro Devices, Inc., or its affiliates.
|
|
// SPDX-License-Identifier: MIT
|
|
|
|
#include <rocRoller/Expression.hpp>
|
|
|
|
template <typename T>
|
|
constexpr auto cast_to_unsigned(T val)
|
|
{
|
|
return static_cast<typename std::make_unsigned<T>::type>(val);
|
|
}
|
|
|
|
namespace rocRoller
|
|
{
|
|
namespace Expression
|
|
{
|
|
struct FuseExpressionVisitor
|
|
{
|
|
template <CUnary Expr>
|
|
ExpressionPtr operator()(Expr const& expr) const
|
|
{
|
|
Expr cpy = expr;
|
|
cpy.arg = call(expr.arg);
|
|
return std::make_shared<Expression>(cpy);
|
|
}
|
|
|
|
template <CBinary Expr>
|
|
ExpressionPtr operator()(Expr const& expr) const
|
|
{
|
|
Expr cpy = expr;
|
|
cpy.lhs = call(expr.lhs);
|
|
cpy.rhs = call(expr.rhs);
|
|
return std::make_shared<Expression>(cpy);
|
|
}
|
|
|
|
template <CTernary Expr>
|
|
ExpressionPtr operator()(Expr const& expr) const
|
|
{
|
|
Expr cpy = expr;
|
|
cpy.lhs = call(expr.lhs);
|
|
cpy.r1hs = call(expr.r1hs);
|
|
cpy.r2hs = call(expr.r2hs);
|
|
return std::make_shared<Expression>(cpy);
|
|
}
|
|
|
|
template <CNary Expr>
|
|
ExpressionPtr operator()(Expr const& expr) const
|
|
{
|
|
auto cpy = expr;
|
|
std::ranges::for_each(cpy.operands, [this](auto& op) { op = call(op); });
|
|
return std::make_shared<Expression>(std::move(cpy));
|
|
}
|
|
|
|
ExpressionPtr operator()(ScaledMatrixMultiply const& expr) const
|
|
{
|
|
ScaledMatrixMultiply cpy = expr;
|
|
cpy.matA = call(expr.matA);
|
|
cpy.matB = call(expr.matB);
|
|
cpy.matC = call(expr.matC);
|
|
cpy.scaleA = call(expr.scaleA);
|
|
cpy.scaleB = call(expr.scaleB);
|
|
return std::make_shared<Expression>(cpy);
|
|
}
|
|
|
|
ExpressionPtr operator()(Add const& expr) const
|
|
{
|
|
auto lhs = call(expr.lhs);
|
|
auto rhs = call(expr.rhs);
|
|
|
|
if(auto const* shift = std::get_if<ShiftL>(lhs.get()))
|
|
{
|
|
bool eval_shift_rhs = evaluationTimes(shift->rhs)[EvaluationTime::Translate];
|
|
if(eval_shift_rhs)
|
|
{
|
|
auto comment = shift->comment + expr.comment;
|
|
auto rv = shiftLAdd(shift->lhs, shift->rhs, rhs);
|
|
setComment(rv, comment);
|
|
return rv;
|
|
}
|
|
}
|
|
|
|
if(auto const* multiply = std::get_if<Multiply>(lhs.get()))
|
|
{
|
|
auto comment = multiply->comment + expr.comment;
|
|
|
|
auto rv = multiplyAdd(multiply->lhs, multiply->rhs, rhs);
|
|
setComment(rv, comment);
|
|
return rv;
|
|
}
|
|
|
|
return std::make_shared<Expression>(Add({lhs, rhs, expr.comment}));
|
|
}
|
|
|
|
ExpressionPtr operator()(ShiftL const& expr) const
|
|
{
|
|
auto lhs = call(expr.lhs);
|
|
auto rhs = call(expr.rhs);
|
|
|
|
if(auto const* add = std::get_if<Add>(lhs.get()))
|
|
{
|
|
bool eval_rhs = evaluationTimes(rhs)[EvaluationTime::Translate];
|
|
if(eval_rhs)
|
|
{
|
|
auto comment = add->comment + expr.comment;
|
|
return std::make_shared<Expression>(
|
|
AddShiftL{add->lhs, add->rhs, rhs, comment});
|
|
}
|
|
}
|
|
|
|
auto cpy = expr;
|
|
cpy.lhs = lhs;
|
|
cpy.rhs = rhs;
|
|
|
|
return std::make_shared<Expression>(cpy);
|
|
}
|
|
|
|
template <CValue Value>
|
|
ExpressionPtr operator()(Value const& expr) const
|
|
{
|
|
return std::make_shared<Expression>(expr);
|
|
}
|
|
|
|
ExpressionPtr call(ExpressionPtr expr) const
|
|
{
|
|
AssertFatal(expr != nullptr, "Found nullptr in expression");
|
|
return std::visit(*this, *expr);
|
|
}
|
|
};
|
|
|
|
ExpressionPtr fuseTernary(ExpressionPtr expr)
|
|
{
|
|
auto visitor = FuseExpressionVisitor();
|
|
return visitor.call(expr);
|
|
}
|
|
}
|
|
}
|