Add quadform/cmpy/cdpy.

CHANGELOG.md

@@ -5,6 +5,10 @@ documented here.
 This project adheres to [Semantic Versioning](http://semver.org/).
 
 ## [0.14.0] − WIP
+### Modified
+  * `quadform` has been renamed `quadform_tr`. The new `quadform` method takes
+    the matrix on the right-hand-side instead of the matrix on the
+    left-hand-side of the quadratic form.
 ### Added
   * The `mint` feature that can be enabled in order to allow conversions from
     and to types of the [mint](https://crates.io/crates/mint) crate.
@@ -12,6 +16,10 @@ This project adheres to [Semantic Versioning](http://semver.org/).
     `::from_element(...)`.
   * The `.iamin()` methods that returns the index of the vector entry with
     smallest absolute value.
+  * Add blas-like operations: `cmpy, cdpy` for componentwise multiplicatons and
+    division with scalar factors:
+        - `self <- alpha * self + beta * a * b`
+        - `self <- alpha * self + beta / a * b`
   * `UnitQuaternion::scaled_rotation_between_axis` and
     `UnitQuaternion::rotation_between_axis` that take Unit vectors instead of
     Vector as arguments.

src/core/blas.rs

@@ -481,6 +481,35 @@ impl<N, R1: Dim, C1: Dim, S: StorageMut<N, R1, C1>> Matrix<N, R1, C1, S>
            }
        }
     }
+
+    /// Computes `self = alpha * a.transpose() * b + beta * self`, where `a, b, self` are matrices.
+    /// `alpha` and `beta` are scalar.
+    ///
+    /// If `beta` is zero, `self` is never read.
+    #[inline]
+    pub fn gemm_tr<R2: Dim, C2: Dim, R3: Dim, C3: Dim, SB, SC>(&mut self,
+                                                               alpha: N,
+                                                               a: &Matrix<N, R2, C2, SB>,
+                                                               b: &Matrix<N, R3, C3, SC>,
+                                                               beta: N)
+        where N:  One,
+              SB: Storage<N, R2, C2>,
+              SC: Storage<N, R3, C3>,
+              ShapeConstraint: SameNumberOfRows<R1, C2>    +
+                               SameNumberOfColumns<C1, C3> +
+                               AreMultipliable<C2, R2, R3, C3> {
+       let (nrows1, ncols1) = self.shape();
+       let (nrows2, ncols2) = a.shape();
+       let (nrows3, ncols3) = b.shape();
+
+       assert_eq!(nrows2, nrows3, "gemm: dimensions mismatch for multiplication.");
+       assert_eq!((nrows1, ncols1), (ncols2, ncols3), "gemm: dimensions mismatch for addition.");
+
+       for j1 in 0 .. ncols1 {
+           // FIXME: avoid bound checks.
+           self.column_mut(j1).gemv_tr(alpha, a, &b.column(j1), beta);
+       }
+    }
 }
 
 
@@ -520,8 +549,10 @@ impl<N, R1: Dim, C1: Dim, S: StorageMut<N, R1, C1>> Matrix<N, R1, C1, S>
 impl<N, D1: Dim, S: StorageMut<N, D1, D1>> SquareMatrix<N, D1, S>
     where N: Scalar + Zero + One + ClosedAdd + ClosedMul {
 
-    /// Computes the quadratic form `self = alpha * lrs * mid * lhs.transpose() + beta * self`.
+    /// Computes the quadratic form `self = alpha * lhs * mid * lhs.transpose() + beta * self`.
-    pub fn quadform_with_workspace<D2, S2, R3, C3, S3, D4, S4>(&mut self,
+    ///
+    /// This uses the provided workspace `work` to avoid allocations for intermediate results.
+    pub fn quadform_tr_with_workspace<D2, S2, R3, C3, S3, D4, S4>(&mut self,
                                                                   work:  &mut Vector<N, D2, S2>,
                                                                   alpha: N,
                                                                   lhs:   &Matrix<N, R3, C3, S3>,
@@ -544,8 +575,11 @@ impl<N, D1: Dim, S: StorageMut<N, D1, D1>> SquareMatrix<N, D1, S>
         }
     }
 
-    /// Computes the quadratic form `self = alpha * lrs * mid * lhs.transpose() + beta * self`.
+    /// Computes the quadratic form `self = alpha * lhs * mid * lhs.transpose() + beta * self`.
-    pub fn quadform<R3, C3, S3, D4, S4>(&mut self,
+    ///
+    /// This allocates a workspace vector of dimension D1 for intermediate results.
+    /// Use `.quadform_tr_with_workspace(...)` instead to avoid allocations.
+    pub fn quadform_tr<R3, C3, S3, D4, S4>(&mut self,
                                            alpha: N,
                                            lhs:   &Matrix<N, R3, C3, S3>,
                                            mid:   &SquareMatrix<N, D4, S4>,
@@ -556,6 +590,53 @@ impl<N, D1: Dim, S: StorageMut<N, D1, D1>> SquareMatrix<N, D1, S>
               ShapeConstraint: DimEq<D1, D1> + DimEq<D1, R3> + DimEq<C3, D4>,
               DefaultAllocator: Allocator<N, D1> {
         let mut work = unsafe { Vector::new_uninitialized_generic(self.data.shape().0, U1) };
-        self.quadform_with_workspace(&mut work, alpha, lhs, mid, beta)
+        self.quadform_tr_with_workspace(&mut work, alpha, lhs, mid, beta)
+    }
+
+    /// Computes the quadratic form `self = alpha * rhs.transpose() * mid * rhs + beta * self`.
+    ///
+    /// This uses the provided workspace `work` to avoid allocations for intermediate results.
+    pub fn quadform_with_workspace<D2, S2, D3, S3, R4, C4, S4>(&mut self,
+                                                               work:  &mut Vector<N, D2, S2>,
+                                                               alpha: N,
+                                                               mid:   &SquareMatrix<N, D3, S3>,
+                                                               rhs:   &Matrix<N, R4, C4, S4>,
+                                                               beta:  N)
+        where D2: Dim, D3: Dim, R4: Dim, C4: Dim,
+              S2: StorageMut<N, D2>,
+              S3: Storage<N, D3, D3>,
+              S4: Storage<N, R4, C4>,
+              ShapeConstraint: DimEq<D3, R4> +
+                               DimEq<D1, C4> +
+                               DimEq<D2, D3> +
+                               AreMultipliable<C4, R4, D2, U1> {
+        work.gemv(N::one(), mid, &rhs.column(0), N::zero());
+        self.column_mut(0).gemv_tr(alpha, &rhs, work, beta);
+
+        for j in 1 .. rhs.ncols() {
+            work.gemv(N::one(), mid, &rhs.column(j), N::zero());
+            self.column_mut(j).gemv_tr(alpha, &rhs, work, beta);
+        }
+    }
+
+    /// Computes the quadratic form `self = alpha * rhs.transpose() * mid * rhs + beta * self`.
+    ///
+    /// This allocates a workspace vector of dimension D2 for intermediate results.
+    /// Use `.quadform_with_workspace(...)` instead to avoid allocations.
+    pub fn quadform<D2, S2, R3, C3, S3>(&mut self,
+                                        alpha: N,
+                                        mid:   &SquareMatrix<N, D2, S2>,
+                                        rhs:   &Matrix<N, R3, C3, S3>,
+                                        beta:  N)
+        where D2: Dim, R3: Dim, C3: Dim,
+              S2: Storage<N, D2, D2>,
+              S3: Storage<N, R3, C3>,
+              ShapeConstraint: DimEq<D2, R3> +
+                               DimEq<D1, C3> +
+                               AreMultipliable<C3, R3, D2, U1>,
+              DefaultAllocator: Allocator<N, D2> {
+
+        let mut work = unsafe { Vector::new_uninitialized_generic(mid.data.shape().0, U1) };
+        self.quadform_with_workspace(&mut work, alpha, mid, rhs, beta)
     }
 }

src/core/componentwise.rs

@@ -1,6 +1,7 @@
 // Non-convensional componentwise operators.
 
-use num::Signed;
+use std::ops::{Add, Mul};
+use num::{Zero, Signed};
 
 use alga::general::{ClosedMul, ClosedDiv};
 
@@ -33,7 +34,7 @@ impl<N: Scalar, R: Dim, C: Dim, S: Storage<N, R, C>> Matrix<N, R, C, S> {
 }
 
 macro_rules! component_binop_impl(
-    ($($binop: ident, $binop_mut: ident, $binop_assign: ident, $Trait: ident . $op_assign: ident, $desc:expr, $desc_mut:expr);* $(;)*) => {$(
+    ($($binop: ident, $binop_mut: ident, $binop_assign: ident, $cbpy: ident, $Trait: ident . $op: ident . $op_assign: ident, $desc:expr, $desc_mut:expr);* $(;)*) => {$(
         impl<N: Scalar, R1: Dim, C1: Dim, SA: Storage<N, R1, C1>> Matrix<N, R1, C1, SA> {
             #[doc = $desc]
             #[inline]
@@ -60,6 +61,41 @@ macro_rules! component_binop_impl(
         }
 
         impl<N: Scalar, R1: Dim, C1: Dim, SA: StorageMut<N, R1, C1>> Matrix<N, R1, C1, SA> {
+            // componentwise binop plus Y.
+            #[inline]
+            pub fn $cbpy<R2, C2, SB, R3, C3, SC>(&mut self, alpha: N, a: &Matrix<N, R2, C2, SB>, b: &Matrix<N, R3, C3, SC>, beta: N)
+                where N: $Trait + Zero + Mul<N, Output = N> + Add<N, Output = N>,
+                      R2: Dim, C2: Dim,
+                      R3: Dim, C3: Dim,
+                      SB: Storage<N, R2, C2>,
+                      SC: Storage<N, R3, C3>,
+                      ShapeConstraint: SameNumberOfRows<R1, R2> + SameNumberOfColumns<C1, C2> +
+                                       SameNumberOfRows<R1, R3> + SameNumberOfColumns<C1, C3> {
+                assert_eq!(self.shape(), a.shape(), "Componentwise mul/div: mismatched matrix dimensions.");
+                assert_eq!(self.shape(), b.shape(), "Componentwise mul/div: mismatched matrix dimensions.");
+
+                if beta.is_zero() {
+                    for j in 0 .. self.ncols() {
+                        for i in 0 .. self.nrows() {
+                            unsafe {
+                                let res = alpha * a.get_unchecked(i, j).$op(*b.get_unchecked(i, j));
+                                *self.get_unchecked_mut(i, j) = res;
+                            }
+                        }
+                    }
+                }
+                else {
+                    for j in 0 .. self.ncols() {
+                        for i in 0 .. self.nrows() {
+                            unsafe {
+                                let res = alpha * a.get_unchecked(i, j).$op(*b.get_unchecked(i, j));
+                                *self.get_unchecked_mut(i, j) = beta * *self.get_unchecked(i, j) + res;
+                            }
+                        }
+                    }
+                }
+            }
+
             #[doc = $desc_mut]
             #[inline]
             pub fn $binop_assign<R2, C2, SB>(&mut self, rhs: &Matrix<N, R2, C2, SB>)
@@ -96,9 +132,9 @@ macro_rules! component_binop_impl(
 );
 
 component_binop_impl!(
-    component_mul, component_mul_mut, component_mul_assign, ClosedMul.mul_assign,
+    component_mul, component_mul_mut, component_mul_assign, cmpy, ClosedMul.mul.mul_assign,
     "Componentwise matrix multiplication.", "Mutable, componentwise matrix multiplication.";
-    component_div, component_div_mut, component_div_assign, ClosedDiv.div_assign,
+    component_div, component_div_mut, component_div_assign, cdpy, ClosedDiv.div.div_assign,
     "Componentwise matrix division.", "Mutable, componentwise matrix division.";
     // FIXME: add other operators like bitshift, etc. ?
 );

tests/core/blas.rs

@@ -74,6 +74,21 @@ quickcheck! {
     }
 
     fn quadform(n: usize, alpha: f64, beta: f64) -> bool {
+        let n       = cmp::max(1, cmp::min(n, 50));
+        let rhs     = DMatrix::<f64>::new_random(6, n);
+        let mid     = DMatrix::<f64>::new_random(6, 6);
+        let mut res = DMatrix::new_random(n, n);
+
+        let expected = &res * beta + rhs.transpose() * &mid * &rhs * alpha;
+
+        res.quadform(alpha, &mid, &rhs, beta);
+
+        println!("{}{}", res, expected);
+
+        relative_eq!(res, expected, epsilon = 1.0e-7)
+    }
+
+    fn quadform_tr(n: usize, alpha: f64, beta: f64) -> bool {
         let n       = cmp::max(1, cmp::min(n, 50));
         let lhs     = DMatrix::<f64>::new_random(6, n);
         let mid     = DMatrix::<f64>::new_random(n, n);
@@ -81,7 +96,7 @@ quickcheck! {
 
         let expected = &res * beta + &lhs * &mid * lhs.transpose() * alpha;
 
-        res.quadform(alpha, &lhs, &mid , beta);
+        res.quadform_tr(alpha, &lhs, &mid , beta);
 
         println!("{}{}", res, expected);