/*

Vector, (3-Scalar x,y,z), and Matrix, (3-Vector array) Types.

Useful methods on these, mostly operating in-place (with no returned reference no single line method chaining.)

Arrays of both, (called Vectors and Matrices), with their useful methods.

Vectorized/SIMD methods that accept a function and apply in to all the array.

selectable single or parallel threaded, transparently, switched with a global var.

VectorRefs; array of Vector pointers, with methods as Vectors but also converters to/from Vectors.

the five main types are available (in separate packages) with their base scalar typed as 64bit float, 32bit float, or fixed precision (int scaled for 3dp)

notes:

doesn't use "math" package, left to importers, if necessary.

when parallel, array types are acted on in chunks by different threads.

installation:

get github.com/splace/tensor3   // 64bit
//	get github.com/splace/tensor3/32bit
//	get github.com/splace/tensor3/scaledint

Example: 100 x 1 million matrix multiplications, single threaded then parallel.

package main

import . "github.com/splace/tensor3"  // 64 bit
//import . "github.com/splace/tensor3/32bit"
//import . "github.com/splace/tensor3/scaledint"
import "time"
import "fmt"

func main(){
	ms := make(Matrices, 1000000)
	// fill in array of matrices
	for i := range ms {
		ms[i] = NewMatrix(1, 2, 3, 4, 5, 6, 7, 8, 9)
	}
	m := NewMatrix(9, 8, 7, 6, 5, 4, 3, 2, 1)
	start:=time.Now()

	// multiply every matrix by m, do it 100 times (to make benchmark time resolution better.)
	for i := 0; i < 100; i++ {
		ms.Product(m)
	}
	fmt.Println(time.Since(start))
	// same thing in parallel
	Parallel=true
	start=time.Now()
	for i := 0; i < 100; i++ {
		ms.Product(m)
	}
	fmt.Println(time.Since(start))
}

*/

package tensor3

note: there are a lot of methods for common operations on the main types, and documentation comments are often missing on basic methods that appear,to me, unambiguous from their name or i havent got round to yet.

// Overview/docs: