Files
TheRock/rocm-libraries/shared/rocroller/lib/source/ExpressionTransformations/LowerPRNG.cpp
T
derek.holloway 7502bf64a4
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
Flatten subprojects into a single base monorepo
2026-09-02 16:34:47 -07:00

114 lines
4.0 KiB
C++

// Copyright Advanced Micro Devices, Inc., or its affiliates.
// SPDX-License-Identifier: MIT
#include <rocRoller/Expression.hpp>
namespace rocRoller
{
namespace Expression
{
struct LowerPRNGExpressionVisitor
{
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()(RandomNumber const& expr) const
{
auto arg = call(expr.arg);
// PRNG algorithm: ((seed << 1) ^ (((seed >> 31) & 1) ? 0xc5 : 0x00))
ExpressionPtr one = literal(1u);
ExpressionPtr lhs = std::make_shared<Expression>(ShiftL({arg, one}));
ExpressionPtr shiftR
= std::make_shared<Expression>(LogicalShiftR({arg, literal(31u)}));
ExpressionPtr bitAnd = std::make_shared<Expression>(BitwiseAnd({shiftR, one}));
ExpressionPtr predicate = std::make_shared<Expression>(Equal({bitAnd, one}));
// Note: here we compute the `xor 0xc5` even though the value might not be selected.
// The reason of doing this instead of
// rhs = condition(predicate, 0xc5, 0x00)
// result = xor(value, rhs)
// is that `v_cndmask_b32_e64 v5, 0, 197, s[12:13]` is not able to assemble
// (Error: literal operands are not supported).
ExpressionPtr xorValue
= std::make_shared<Expression>(BitwiseXor({literal(197u), lhs})); // xor 0xc5
return std::make_shared<Expression>(Conditional({predicate, xorValue, lhs}));
}
template <CValue Value>
ExpressionPtr operator()(Value const& expr) const
{
return std::make_shared<Expression>(expr);
}
ExpressionPtr call(ExpressionPtr expr) const
{
if(!expr)
return expr;
return std::visit(*this, *expr);
}
};
/**
* Replace RandomNumber expression with equivalent expressions
*/
ExpressionPtr lowerPRNG(ExpressionPtr expr, ContextPtr cxt)
{
if(cxt->targetArchitecture().HasCapability(GPUCapability::HasPRNG))
return expr;
auto visitor = LowerPRNGExpressionVisitor();
return visitor.call(expr);
}
}
}