Reduce allocations of cov(::MultivariateMixture) (#1967)

* Fix MixtureModel with ScalMat covariances

* Use `mul!`

* Use `LinearAlgebra.copytri!`

* Update src/mixtures/mixturemodel.jl

* Add more tests

Author: devmotion

src/mixtures/mixturemodel.jl

@@ -244,11 +244,11 @@ function cov(d::MultivariateMixture)
         pi = p[i]
         if pi > 0.0
             c = component(d, i)
-            # todo: use more in-place operations
-            md = mean(c) - m
-            axpy!(pi, md*md', V)
+            md .= mean(c) .- m
+            BLAS.syr!('U', Float64(pi), md, V)
         end
     end
+    LinearAlgebra.copytri!(V, 'U')
     return V
 end
 

test/mixture.jl

@@ -1,5 +1,6 @@
 using Distributions, Random
 using Test
+using LinearAlgebra
 using ForwardDiff: Dual
 
 
@@ -252,17 +253,19 @@ end
     end
 
     @testset "Testing MultivariatevariateMixture" begin
-        g_m = MixtureModel(
-            IsoNormal[ MvNormal([0.0, 0.0], I),
-                       MvNormal([0.2, 1.0], I),
-                       MvNormal([-0.5, -3.0], 1.6 * I) ],
-            [0.2, 0.5, 0.3])
-        @test isa(g_m, MixtureModel{Multivariate, Continuous, IsoNormal})
-        @test length(components(g_m)) == 3
-        @test length(g_m) == 2
-        @test insupport(g_m, [0.0, 0.0]) == true
-        test_mixture(g_m, 1000, 10^6, rng)
-        test_params(g_m)
+        for T in (Float32, Float64)
+            g_m = MixtureModel(
+                IsoNormal[ MvNormal([0.0, 0.0], I),
+                           MvNormal([0.2, 1.0], I),
+                           MvNormal([-0.5, -3.0], 1.6 * I) ],
+                T[0.2, 0.5, 0.3])
+            @test isa(g_m, MixtureModel{Multivariate, Continuous, IsoNormal})
+            @test length(components(g_m)) == 3
+            @test length(g_m) == 2
+            @test insupport(g_m, [0.0, 0.0])
+            test_mixture(g_m, 1000, 10^6, rng)
+            test_params(g_m)
+        end
 
         u1 =  Uniform()
         u2 =  Uniform(1.0, 2.0)