NUnit Data Points and Theories - http://stackoverflow.com/questions/27818357/best-practice-for-checking-double-float-parameter

NUnitTheoryTest

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using NUnit.Framework;

namespace ClassLibrary1
{
    public static class SillyClass
    {
        public static bool IsNaNOrInfinity(this double foo)
        {
            return (double.IsNaN(foo)) || (double.IsInfinity(foo));
        }
    }

    class CUT
    {
        public void Foo(double bar, double bar2, double bar3, double bar4)
        {
            if (bar.IsNaNOrInfinity()) throw new ArgumentOutOfRangeException();
            if (bar2.IsNaNOrInfinity()) throw new ArgumentOutOfRangeException();
            if (bar3.IsNaNOrInfinity()) throw new ArgumentOutOfRangeException();
            if (bar4.IsNaNOrInfinity()) throw new ArgumentOutOfRangeException();
        }        
    }
    
    [TestFixture]
    public class FooTests
    {
        [Datapoints]
        private double[] _permutations = new[]
        {
            double.PositiveInfinity,
            double.NegativeInfinity,
            double.NaN
        };

        [Theory]
        public void EnsureInvalidDoublesAreThrown(double bar, double bar2, double bar3, double bar4)
        {
            Assert.Throws<ArgumentOutOfRangeException>(() => new CUT().Foo(bar, bar2, bar3, bar4));
        }
    }
}

Here we use NUnit's Theories along with an array of DataPoints, which when run with a decent TestRunner (I used Resharper) will generate all permutations of the parameters and pass them into test Theories which are matched by parameter type (double). The above generates 3 ^ 4 = 81 invocations of the Unit test EnsureInvalidDoublesAreThrown