FloatConversionTest class. More...
#include <TestRvsdgs.hpp>
Inheritance diagram for jlm::mlir::FloatConversionTest:
Collaboration diagram for jlm::mlir::FloatConversionTest:
Public Attributes | |
| jlm::rvsdg::LambdaNode * | lambda |
Private Member Functions | |
| std::unique_ptr< jlm::llvm::LlvmRvsdgModule > | SetupRvsdg () override |
| Create RVSDG for this test. | |
Additional Inherited Members | |
 Public Member Functions inherited from jlm::llvm::RvsdgTest | |
| virtual | ~RvsdgTest ()=default |
| jlm::llvm::LlvmRvsdgModule & | module () |
| const rvsdg::Graph & | graph () |
| void | InitializeTest () |
Detailed Description
FloatConversionTest class.
This class sets up an RVSDG representing the following function:
double f(double a, float b, int64_t i)
{
double e = (double)b; // floating point extension
double n = -(e * 2.0); // negation of a product with a floating point constant
double t = (double)i; // signed int to floating point
double q = a * a + t; // fused multiply-add
return n + t, t + q; // two results
}
In contrast to the FloatBinaryTest, every operation here takes its own branch in both converters: a floating point constant, negation, floating point extension, signed conversion and the multiply-add intrinsic.
- See also
- FloatBinaryTest
Definition at line 234 of file TestRvsdgs.hpp.
Member Function Documentation
◆ SetupRvsdg()
|
overrideprivatevirtual |
Create RVSDG for this test.
Implements jlm::llvm::RvsdgTest.
Definition at line 292 of file TestRvsdgs.cpp.
Member Data Documentation
◆ lambda
| jlm::rvsdg::LambdaNode* jlm::mlir::FloatConversionTest::lambda |
Definition at line 241 of file TestRvsdgs.hpp.
The documentation for this class was generated from the following files:
- jlm/mlir/TestRvsdgs.hpp
- jlm/mlir/TestRvsdgs.cpp
