[rand.dist.samp.pconst] plus some bug fixes in the tests of the other distributions

llvm-svn: 104224

1 files changed, 15 insertions, 15 deletions

diff --git a/libcxx/test/numerics/rand/rand.dis/rand.dist.bern/rand.dist.bern.bin/eval_param.pass.cpp b/libcxx/test/numerics/rand/rand.dis/rand.dist.bern/rand.dist.bern.bin/eval_param.pass.cpp
index 2632ab5f3ae..264a4a2f11e 100644
--- a/libcxx/test/numerics/rand/rand.dis/rand.dist.bern/rand.dist.bern.bin/eval_param.pass.cpp
+++ b/libcxx/test/numerics/rand/rand.dis/rand.dist.bern/rand.dist.bern.bin/eval_param.pass.cpp
@@ -32,11 +32,11 @@ int main()
     {
         typedef std::binomial_distribution<> D;
         typedef D::param_type P;
-        typedef std::minstd_rand G;
+        typedef std::mt19937_64 G;
         G g;
         D d(16, .75);
         P p(5, .75);
-        const int N = 100000;
+        const int N = 1000000;
         std::vector<D::result_type> u;
         for (int i = 0; i < N; ++i)
         {
@@ -66,10 +66,10 @@ int main()
         double x_var = x_mean*(1-p.p());
         double x_skew = (1-2*p.p()) / std::sqrt(x_var);
         double x_kurtosis = (1-6*p.p()*(1-p.p())) / x_var;
-        assert(std::abs(mean - x_mean) / x_mean < 0.01);
-        assert(std::abs(var - x_var) / x_var < 0.01);
-        assert(std::abs(skew - x_skew) / x_skew < 0.01);
-        assert(std::abs(kurtosis - x_kurtosis) / x_kurtosis < 0.01);
+        assert(std::abs((mean - x_mean) / x_mean) < 0.01);
+        assert(std::abs((var - x_var) / x_var) < 0.01);
+        assert(std::abs((skew - x_skew) / x_skew) < 0.01);
+        assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.04);
     }
     {
         typedef std::binomial_distribution<> D;
@@ -108,10 +108,10 @@ int main()
         double x_var = x_mean*(1-p.p());
         double x_skew = (1-2*p.p()) / std::sqrt(x_var);
         double x_kurtosis = (1-6*p.p()*(1-p.p())) / x_var;
-        assert(std::abs(mean - x_mean) / x_mean < 0.01);
-        assert(std::abs(var - x_var) / x_var < 0.01);
-        assert(std::abs(skew - x_skew) / x_skew < 0.01);
-        assert(std::abs(kurtosis - x_kurtosis) / x_kurtosis < 0.01);
+        assert(std::abs((mean - x_mean) / x_mean) < 0.01);
+        assert(std::abs((var - x_var) / x_var) < 0.01);
+        assert(std::abs((skew - x_skew) / x_skew) < 0.01);
+        assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.01);
     }
     {
         typedef std::binomial_distribution<> D;
@@ -120,7 +120,7 @@ int main()
         G g;
         D d(16, .75);
         P p(40, .25);
-        const int N = 100000;
+        const int N = 1000000;
         std::vector<D::result_type> u;
         for (int i = 0; i < N; ++i)
         {
@@ -150,9 +150,9 @@ int main()
         double x_var = x_mean*(1-p.p());
         double x_skew = (1-2*p.p()) / std::sqrt(x_var);
         double x_kurtosis = (1-6*p.p()*(1-p.p())) / x_var;
-        assert(std::abs(mean - x_mean) / x_mean < 0.01);
-        assert(std::abs(var - x_var) / x_var < 0.01);
-        assert(std::abs(skew - x_skew) / x_skew < 0.03);
-        assert(std::abs(kurtosis - x_kurtosis) / x_kurtosis < 0.01);
+        assert(std::abs((mean - x_mean) / x_mean) < 0.01);
+        assert(std::abs((var - x_var) / x_var) < 0.01);
+        assert(std::abs((skew - x_skew) / x_skew) < 0.04);
+        assert(std::abs((kurtosis - x_kurtosis) / x_kurtosis) < 0.3);
     }
 }