mlir.dialects.transform.math¶

Classes¶

`ApplyF32ExpansionPatternsOp`	Converts math functions to f32: operands are extended and the result is
`ApplyF32ExpansionPatternsOpAdaptor`
`ApplyPolynomialApproximationPatternsOp`	Enables polynomial approximations for math functions.
`ApplyPolynomialApproximationPatternsOpAdaptor`

Functions¶

`apply_patterns_math_f32_expansion`(...)
`apply_patterns_math_polynomial_approximation`(...)

Module Contents¶

class mlir.dialects.transform.math.ApplyF32ExpansionPatternsOp(*, loc: _ods_ir | None = None, ip: _ods_ir | None = None)¶

Bases: _ods_ir

Converts math functions to f32: operands are extended and the result is truncated back.

OPERATION_NAME = 'transform.apply_patterns.math.f32_expansion'¶

_ODS_REGIONS = (0, True)¶

class mlir.dialects.transform.math.ApplyF32ExpansionPatternsOpAdaptor(operands: list[Value], attributes: OpAttributeMap)¶

class mlir.dialects.transform.math.ApplyF32ExpansionPatternsOpAdaptor(operands: list[Value], opview: OpView)

Bases: _ods_ir

OPERATION_NAME = 'transform.apply_patterns.math.f32_expansion'¶

mlir.dialects.transform.math.apply_patterns_math_f32_expansion(*, loc: _ods_ir | None = None, ip: _ods_ir | None = None) → ApplyF32ExpansionPatternsOp¶

class mlir.dialects.transform.math.ApplyPolynomialApproximationPatternsOp(*, enable_avx2: bool | None = None, loc: _ods_ir | None = None, ip: _ods_ir | None = None)¶

Bases: _ods_ir

Enables polynomial approximations for math functions. Optionally, allows usage of x86 AVX2 intrinsics.

OPERATION_NAME = 'transform.apply_patterns.math.polynomial_approximation'¶

_ODS_REGIONS = (0, True)¶

enable_avx2() → bool¶

class mlir.dialects.transform.math.ApplyPolynomialApproximationPatternsOpAdaptor(operands: list[Value], attributes: OpAttributeMap)¶

class mlir.dialects.transform.math.ApplyPolynomialApproximationPatternsOpAdaptor(operands: list[Value], opview: OpView)

Bases: _ods_ir

OPERATION_NAME = 'transform.apply_patterns.math.polynomial_approximation'¶

enable_avx2() → bool¶

mlir.dialects.transform.math.apply_patterns_math_polynomial_approximation(*, enable_avx2: bool | None = None, loc: _ods_ir | None = None, ip: _ods_ir | None = None) → ApplyPolynomialApproximationPatternsOp¶