[Mlir-commits] [mlir] [MLIR][Presburger] Generating functions and quasi-polynomials for Barvinok's algorithm (PR #75702)

[Arjun P]

================
@@ -0,0 +1,113 @@
+//===- QuasiPolynomial.h - Quasipolynomial Class ----------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// Definition of the QuasiPolynomial class for Barvinok's algorithm,
+// which represents a single-valued function on a set of parameters.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef MLIR_ANALYSIS_PRESBURGER_QUASIPOLYNOMIAL_H
+#define MLIR_ANALYSIS_PRESBURGER_QUASIPOLYNOMIAL_H
+
+#include "mlir/Analysis/Presburger/Fraction.h"
+#include "mlir/Analysis/Presburger/Matrix.h"
+
+namespace mlir {
+namespace presburger {
+
+// A class to describe the quasi-polynomials obtained by
+// substituting the unit vector in the type of generating
+// function described above.
+// Consists of a set of terms.
+// The ith term is a constant `coefficients[i]`, multiplied
+// by the product of a set of affine functions on n parameters.
+// Represents functions f : Q^n -> Q of the form
+//
+// f(x) = \sum_i c_i * \prod_j ⌊g_{ij}(x)⌋
+//
+// where c_i \in Q and
+// g_{ij} : Q^n -> Q are affine functionals.
+class QuasiPolynomial {
+public:
+  QuasiPolynomial(unsigned numParam, SmallVector<Fraction> coeffs = {},
+                  std::vector<std::vector<SmallVector<Fraction>>> aff = {});
+
+  unsigned getNumParams();
+  SmallVector<Fraction> getCoefficients();
+  std::vector<std::vector<SmallVector<Fraction>>> getAffine();
----------------
Superty wrote:

You should do the reverse: you should define these small, frequently used functions like getters in the header and the rest like the operators, though they may be used often, are pretty big so should be put in the .cpp file generally.

https://github.com/llvm/llvm-project/pull/75702