View learnChapelInYMinutes.chpl on GitHub

Learn Chapel in Y Minutes

This primer will go over basic syntax and concepts in Chapel.

Comments are C-family style

// one line comment

   multi-line comment

Basic printing

write("Hello, ");

write and writeln can take a list of things to print. Each thing is printed right next to the others, so include your spacing!

writeln("There are ", 3, " commas (\",\") in this line of code");

Different output channels:

stdout.writeln("This goes to standard output, just like plain writeln() does");
stderr.writeln("This goes to standard error");


Variables don't have to be explicitly typed as long as the compiler can figure out the type that it will hold. 10 is an int, so myVar is implicitly an int

var myVar = 10;
myVar = -10;
var mySecondVar = myVar;

var anError; would be a compile-time error.

We can (and should) explicitly type things.

var myThirdVar: real;
var myFourthVar: real = -1.234;
myThirdVar = myFourthVar;


There are a number of basic types.

var myInt: int = -1000; // Signed ints
var myUint: uint = 1234; // Unsigned ints
var myReal: real = 9.876; // Floating point numbers
var myImag: imag = 5.0i; // Imaginary numbers
var myCplx: complex = 10 + 9i; // Complex numbers
myCplx = myInt + myImag; // Another way to form complex numbers
var myBool: bool = false; // Booleans
var myStr: string = "Some string..."; // Strings
var singleQuoteStr = 'Another string...'; // String literal with single quotes

Some types can have sizes.

var my8Int: int(8) = 10; // 8 bit (one byte) sized int;
var my64Real: real(64) = 1.516; // 64 bit (8 bytes) sized real


var intFromReal = myReal : int;
var intFromReal2: int = myReal : int;

Type aliasing.

type chroma = int;        // Type of a single hue
type RGBColor = 3*chroma; // Type representing a full color
var black: RGBColor = (0,0,0);
var white: RGBColor = (255, 255, 255);

Constants and Parameters

A const is a constant, and cannot be changed after set in runtime.

const almostPi: real = 22.0/7.0;

A param is a constant whose value must be known statically at compile-time.

param compileTimeConst: int = 16;

The config modifier allows values to be set at the command line. Set with --varCmdLineArg=Value or --varCmdLineArg Value at runtime.

config var varCmdLineArg: int = -123;
config const constCmdLineArg: int = 777;

config param can be set at compile-time. Set with --set paramCmdLineArg=value at compile-time.

config param paramCmdLineArg: bool = false;
writeln(varCmdLineArg, ", ", constCmdLineArg, ", ", paramCmdLineArg);


ref operates much like a reference in C++. In Chapel, a ref cannot be made to alias a variable other than the variable it is initialized with. Here, refToActual refers to actual.

var actual = 10;
ref refToActual = actual;
writeln(actual, " == ", refToActual); // prints the same value
actual = -123; // modify actual (which refToActual refers to)
writeln(actual, " == ", refToActual); // prints the same value
refToActual = 99999999; // modify what refToActual refers to (which is actual)
writeln(actual, " == ", refToActual); // prints the same value


Math operators:

var a: int, thisInt = 1234, thatInt = 5678;
a = thisInt + thatInt;  // Addition
a = thisInt * thatInt;  // Multiplication
a = thisInt - thatInt;  // Subtraction
a = thisInt / thatInt;  // Division
a = thisInt ** thatInt; // Exponentiation
a = thisInt % thatInt;  // Remainder (modulo)

Logical operators:

var b: bool, thisBool = false, thatBool = true;
b = thisBool && thatBool; // Logical and
b = thisBool || thatBool; // Logical or
b = !thisBool;            // Logical negation

Relational operators:

b = thisInt > thatInt;           // Greater-than
b = thisInt >= thatInt;          // Greater-than-or-equal-to
b = thisInt < a && a <= thatInt; // Less-than, and, less-than-or-equal-to
b = thisInt != thatInt;          // Not-equal-to
b = thisInt == thatInt;          // Equal-to

Bitwise operators:

a = thisInt << 10;     // Left-bit-shift by 10 bits;
a = thatInt >> 5;      // Right-bit-shift by 5 bits;
a = ~thisInt;          // Bitwise-negation
a = thisInt ^ thatInt; // Bitwise exclusive-or

Compound assignment operators:

a += thisInt;          // Addition-equals (a = a + thisInt;)
a *= thatInt;          // Times-equals (a = a * thatInt;)
b &&= thatBool;        // Logical-and-equals (b = b && thatBool;)
a <<= 3;               // Left-bit-shift-equals (a = a << 10;)

Unlike other C family languages, there are no pre/post-increment/decrement operators, such as:

++j, --j, j++, j--

Swap operator:

var old_this = thisInt;
var old_that = thatInt;
thisInt <=> thatInt; // Swap the values of thisInt and thatInt
writeln((old_this == thatInt) && (old_that == thisInt));

Operator overloads can also be defined, as we'll see with procedures.


Tuples can be of the same type or different types.

var sameTup: 2*int = (10, -1);
var sameTup2 = (11, -6);
var diffTup: (int,real,complex) = (5, 1.928, myCplx);
var diffTupe2 = (7, 5.64, 6.0+1.5i);

Tuples can be accessed using square brackets or parentheses, and are 1-indexed.

writeln("(", sameTup[1], ",", sameTup(2), ")");

Tuples can also be written into.

diffTup(1) = -1;

Tuple values can be expanded into their own variables.

var (tupInt, tupReal, tupCplx) = diffTup;
writeln(diffTup == (tupInt, tupReal, tupCplx));

They are also useful for writing a list of variables, as is common in debugging.


Control Flow

if - then - else works just like any other C-family language.

if 10 < 100 then
  writeln("All is well");

if -1 < 1 then
  writeln("Continuing to believe reality");
  writeln("Send mathematician, something's wrong");

You can use parentheses if you prefer.

if (10 > 100) {
  writeln("Universe broken. Please reboot universe.");

if a % 2 == 0 {
  writeln(a, " is even.");
} else {
  writeln(a, " is odd.");

if a % 3 == 0 {
  writeln(a, " is even divisible by 3.");
} else if a % 3 == 1 {
  writeln(a, " is divided by 3 with a remainder of 1.");
} else {
  writeln(b, " is divided by 3 with a remainder of 2.");

Ternary: if - then - else in a statement.

var maximum = if thisInt < thatInt then thatInt else thisInt;

select statements are much like switch statements in other languages. However, select statements don't cascade like in C or Java.

var inputOption = "anOption";
select inputOption {
  when "anOption" do writeln("Chose 'anOption'");
  when "otherOption" {
    writeln("Chose 'otherOption'");
    writeln("Which has a body");
  otherwise {
    writeln("Any other Input");
    writeln("the otherwise case doesn't need a do if the body is one line");

while and do-while loops also behave like their C counterparts.

var j: int = 1;
var jSum: int = 0;
while (j <= 1000) {
  jSum += j;
  j += 1;

do {
  jSum += j;
  j += 1;
} while (j <= 10000);

for loops are much like those in python in that they iterate over a range. Ranges (like the 1..10 expression below) are a first-class object in Chapel, and as such can be stored in variables.

for i in 1..10 do write(i, ", ");

var iSum: int = 0;
for i in 1..1000 {
  iSum += i;

for x in 1..10 {
  for y in 1..10 {
    write((x,y), "\t");

Ranges and Domains

For-loops and arrays both use ranges and domains to define an index set that can be iterated over. Ranges are single dimensional integer indices, while domains can be multi-dimensional and represent indices of different types.

They are first-class citizen types, and can be assigned into variables.

var range1to10: range = 1..10;  // 1, 2, 3, ..., 10
var range2to11 = 2..11; // 2, 3, 4, ..., 11
var rangeThisToThat: range = thisInt..thatInt; // using variables
var rangeEmpty: range = 100..-100; // this is valid but contains no indices

Ranges can be unbounded.

var range1toInf: range(boundedType=BoundedRangeType.boundedLow) = 1.. ; // 1, 2, 3, 4, 5, ...
var rangeNegInfTo1 = ..1; // ..., -4, -3, -2, -1, 0, 1

Ranges can be strided (and reversed) using the by operator.

var range2to10by2: range(stridable=true) = 2..10 by 2; // 2, 4, 6, 8, 10
var reverse2to10by2 = 2..10 by -2; // 10, 8, 6, 4, 2

var trapRange = 10..1 by -1; // Do not be fooled, this is still an empty range
writeln("Size of range '", trapRange, "' = ", trapRange.length);

Note: range(boundedType= ...) and range(stridable= ...) are only necessary if we explicitly type the variable.

The end point of a range can be determined using the count (#) operator.

var rangeCount: range = -5..#12; // range from -5 to 6

Operators can be mixed.

var rangeCountBy: range(stridable=true) = -5..#12 by 2; // -5, -3, -1, 1, 3, 5

Properties of the range can be queried. In this example, printing the first index, last index, number of indices, stride, and if 2 is include in the range.

writeln((rangeCountBy.first, rangeCountBy.last, rangeCountBy.length,
           rangeCountBy.stride, rangeCountBy.member(2)));

for i in rangeCountBy {
  write(i, if i == rangeCountBy.last then "\n" else ", ");

Rectangular domains are defined using the same range syntax, but they are required to be bounded (unlike ranges).

var domain1to10: domain(1) = {1..10};        // 1D domain from 1..10;
var twoDimensions: domain(2) = {-2..2,0..2}; // 2D domain over product of ranges
var thirdDim: range = 1..16;
var threeDims: domain(3) = {thirdDim, 1..10, 5..10}; // using a range variable

Domains can also be resized

var resizedDom = {1..10};
writeln("before, resizedDom = ", resizedDom);
resizedDom = {-10..#10};
writeln("after, resizedDom = ", resizedDom);

Indices can be iterated over as tuples.

for idx in twoDimensions do
  write(idx, ", ");

These tuples can also be deconstructed.

for (x,y) in twoDimensions {
  write("(", x, ", ", y, ")", ", ");

Associative domains act like sets.

var stringSet: domain(string); // empty set of strings
stringSet += "a";
stringSet += "b";
stringSet += "c";
stringSet += "a"; // Redundant add "a"
stringSet -= "c"; // Remove "c"

Associative domains can also have a literal syntax

var intSet = {1, 2, 4, 5, 100};

Both ranges and domains can be sliced to produce a range or domain with the intersection of indices.

var rangeA = 1.. ; // range from 1 to infinity
var rangeB =  ..5; // range from negative infinity to 5
var rangeC = rangeA[rangeB]; // resulting range is 1..5
writeln((rangeA, rangeB, rangeC));

var domainA = {1..10, 5..20};
var domainB = {-5..5, 1..10};
var domainC = domainA[domainB];
writeln((domainA, domainB, domainC));


Arrays are similar to those of other languages. Their sizes are defined using domains that represent their indices.

var intArray: [1..10] int;
var intArray2: [{1..10}] int; // equivalent

They can be accessed using either brackets or parentheses

for i in 1..10 do
  intArray[i] = -i;

We cannot access intArray[0] because it exists outside of the index set, {1..10}, we defined it to have. intArray[11] is illegal for the same reason.

var realDomain: domain(2) = {1..5,1..7};
var realArray: [realDomain] real;
var realArray2: [1..5,1..7] real;   // equivalent
var realArray3: [{1..5,1..7}] real; // equivalent

for i in 1..5 {
  for j in realDomain.dim(2) {   // Only use the 2nd dimension of the domain
    realArray[i,j] = -1.61803 * i + 0.5 * j;  // Access using index list
    var idx: 2*int = (i,j);                   // Note: 'index' is a keyword
    realArray[idx] = - realArray[(i,j)];      // Index using tuples

Arrays have domains as members, and can be iterated over as normal.

for idx in realArray.domain {  // Again, idx is a 2*int tuple
  realArray[idx] = 1 / realArray[idx[1], idx[2]]; // Access by tuple and list


The values of an array can also be iterated directly.

var rSum: real = 0;
for value in realArray {
  rSum += value; // Read a value
  value = rSum;  // Write a value
writeln(rSum, "\n", realArray);

Associative arrays (dictionaries) can be created using associative domains.

var dictDomain: domain(string) = { "one", "two" };
var dict: [dictDomain] int = ["one" => 1, "two" => 2];
dict["three"] = 3; // Adds 'three' to 'dictDomain' implicitly
for key in dictDomain.sorted() do

Arrays can be assigned to each other in a few different ways. These arrays will be used in the example.

var thisArray : [0..5] int = [0,1,2,3,4,5];
var thatArray : [0..5] int;

First, simply assign one to the other. This copies thisArray into thatArray, instead of just creating a reference. Therefore, modifying thisArray does not also modify thatArray.

thatArray = thisArray;
thatArray[1] = -1;
writeln((thisArray, thatArray));

Assign a slice from one array to a slice (of the same size) in the other.

thatArray[4..5] = thisArray[1..2];
writeln((thisArray, thatArray));

Operations can also be promoted to work on arrays. 'thisPlusThat' is also an array.

var thisPlusThat = thisArray + thatArray;

Moving on, arrays and loops can also be expressions, where the loop body's expression is the result of each iteration.

var arrayFromLoop = for i in 1..10 do i;

An expression can result in nothing, such as when filtering with an if-expression.

var evensOrFives = for i in 1..10 do if (i % 2 == 0 || i % 5 == 0) then i;


Array expressions can also be written with a bracket notation. Note: this syntax uses the forall parallel concept discussed later.

var evensOrFivesAgain = [i in 1..10] if (i % 2 == 0 || i % 5 == 0) then i;

They can also be written over the values of the array.

arrayFromLoop = [value in arrayFromLoop] value + 1;


Chapel procedures have similar syntax functions in other languages.

proc fibonacci(n : int) : int {
  if n <= 1 then return n;
  return fibonacci(n-1) + fibonacci(n-2);

Input parameters can be untyped to create a generic procedure.

proc doublePrint(thing): void {
  write(thing, " ", thing, "\n");

The return type can be inferred, as long as the compiler can figure it out.

proc addThree(n) {
  return n + 3;


It is also possible to take a variable number of parameters.

proc maxOf(x ...?k) {
  // x refers to a tuple of one type, with k elements
  var maximum = x[1];
  for i in 2..k do maximum = if maximum < x[i] then x[i] else maximum;
  return maximum;
writeln(maxOf(1, -10, 189, -9071982, 5, 17, 20001, 42));

Procedures can have default parameter values, and the parameters can be named in the call, even out of order.

proc defaultsProc(x: int, y: real = 1.2634): (int,real) {
  return (x,y);

writeln(defaultsProc(x=12, y=5.432));
writeln(defaultsProc(y=9.876, x=13));

The ? operator is called the query operator, and is used to take undetermined values like tuple or array sizes and generic types. For example, taking arrays as parameters. The query operator is used to determine the domain of A. This is uesful for defining the return type, though it's not required.

proc invertArray(A: [?D] int): [D] int{
  for a in A do a = -a;
  return A;


We can query the type of arguments to generic procedures. Here we define a procedure that takes two arguments of the same type, yet we don't define what that type is.

proc genericProc(arg1 : ?valueType, arg2 : valueType): void {
  select(valueType) {
    when int do writeln(arg1, " and ", arg2, " are ints");
    when real do writeln(arg1, " and ", arg2, " are reals");
    otherwise writeln(arg1, " and ", arg2, " are somethings!");

genericProc(1, 2);
genericProc(1.2, 2.3);
genericProc(1.0+2.0i, 3.0+4.0i);

We can also enforce a form of polymorphism with the where clause This allows the compiler to decide which function to use. Note: That means that all information needs to be known at compile-time. The param modifier on the arg is used to enforce this constraint.

proc whereProc(param N : int): void
 where (N > 0) {
  writeln("N is greater than 0");

proc whereProc(param N : int): void
 where (N < 0) {
  writeln("N is less than 0");


whereProc(0) would result in a compiler error because there are no functions that satisfy the where clause's condition. We could have defined a whereProc without a where clause that would then have served as a catch all for all the other cases (of which there is only one).

where clauses can also be used to constrain based on argument type.

proc whereType(x: ?t) where t == int {
  writeln("Inside 'int' version of 'whereType': ", x);

proc whereType(x: ?t) {
  writeln("Inside general version of 'whereType': ", x);



Intent modifiers on the arguments convey how those arguments are passed to the procedure.

  • in: copy arg in, but not out
  • out: copy arg out, but not in
  • inout: copy arg in, copy arg out
  • ref: pass arg by reference
proc intentsProc(in inarg, out outarg, inout inoutarg, ref refarg) {
  writeln("Inside Before: ", (inarg, outarg, inoutarg, refarg));
  inarg = inarg + 100;
  outarg = outarg + 100;
  inoutarg = inoutarg + 100;
  refarg = refarg + 100;
  writeln("Inside After: ", (inarg, outarg, inoutarg, refarg));

var inVar: int = 1;
var outVar: int = 2;
var inoutVar: int = 3;
var refVar: int = 4;
writeln("Outside Before: ", (inVar, outVar, inoutVar, refVar));
intentsProc(inVar, outVar, inoutVar, refVar);
writeln("Outside After: ", (inVar, outVar, inoutVar, refVar));

Similarly, we can define intents on the return type. refElement returns a reference to an element of array. This makes more practical sense for class methods where references to elements in a data-structure are returned via a method or iterator.

proc refElement(array : [?D] ?T, idx) ref : T {
  return array[idx];

var myChangingArray : [1..5] int = [1,2,3,4,5];
ref refToElem = refElement(myChangingArray, 5); // store reference to element in ref variable
refToElem = -2; // modify reference which modifies actual value in array

Operator Definitions

Chapel allows for operators to be overloaded. We can define the unary operators: + - ! ~ and the binary operators: + - * / % ** == <= >= < > << >> & | ˆ by += -= *= /= %= **= &= |= ˆ= <<= >>= <=>

Boolean exclusive or operator.

proc ^(left : bool, right : bool): bool {
  return (left || right) && !(left && right);

writeln(true  ^ true);
writeln(false ^ true);
writeln(true  ^ false);
writeln(false ^ false);

Define a * operator on any two types that returns a tuple of those types.

proc *(left : ?ltype, right : ?rtype): (ltype, rtype) {
  writeln("\tIn our '*' overload!");
  return (left, right);

writeln(1 * "a"); // Uses our ``*`` operator.
writeln(1 * 2);   // Uses the default ``*`` operator.

Note: You could break everything if you get careless with your overloads. This here will break everything. Don't do it.

proc +(left: int, right: int): int {
  return left - right;


Iterators are sisters to the procedure, and almost everything about procedures also applies to iterators. However, instead of returning a single value, iterators may yield multiple values to a loop.

This is useful when a complicated set or order of iterations is needed, as it allows the code defining the iterations to be separate from the loop body.

iter oddsThenEvens(N: int): int {
  for i in 1..N by 2 do
    yield i; // yield values instead of returning.
  for i in 2..N by 2 do
    yield i;

for i in oddsThenEvens(10) do write(i, ", ");

Iterators can also yield conditionally, the result of which can be nothing

iter absolutelyNothing(N): int {
  for i in 1..N {
    if N < i { // Always false
      yield i;     // Yield statement never happens

for i in absolutelyNothing(10) {
  writeln("Woa there! absolutelyNothing yielded ", i);

We can zipper together two or more iterators (who have the same number of iterations) using zip() to create a single zipped iterator, where each iteration of the zipped iterator yields a tuple of one value yielded from each iterator.

for (positive, negative) in zip(1..5, -5..-1) do
  writeln((positive, negative));

Zipper iteration is quite important in the assignment of arrays, slices of arrays, and array/loop expressions.

var fromThatArray : [1..#5] int = [1,2,3,4,5];
var toThisArray : [100..#5] int;

Some zipper operations implement other operations. The first statement and the loop are equivalent.

toThisArray = fromThatArray;
for (i,j) in zip(toThisArray.domain, fromThatArray.domain) {
  toThisArray[i] = fromThatArray[j];

These two chunks are also equivalent.

toThisArray = [j in -100..#5] j;

for (i, j) in zip(toThisArray.domain, -100..#5) {
  toThisArray[i] = j;

This is very important in understanding why this statement exhibits a runtime error.

var iterArray : [1..10] int = [i in 1..10] if (i % 2 == 1) then i;

Even though the domain of the array and the loop-expression are the same size, the body of the expression can be thought of as an iterator. Because iterators can yield nothing, that iterator yields a different number of things than the domain of the array or loop, which is not allowed.


Classes are similar to those in C++ and Java, allocated on the heap.

class MyClass {

Member variables

var memberInt : int;
var memberBool : bool = true;

Explicitly defined initializer. We also get the compiler-generated initializer, with one argument per field. Note that soon there will be no compiler-generated initializer when we define any initializer(s) explicitly.

proc MyClass(val : real) {
  this.memberInt = ceil(val): int;

Explicitly defined deinitializer. If we did not write one, we would get the compiler-generated deinitializer, which has an empty body.

proc deinit() {
  writeln("MyClass deinitializer called ", (this.memberInt, this.memberBool));

Class methods.

  proc setMemberInt(val: int) {
    this.memberInt = val;

  proc setMemberBool(val: bool) {
    this.memberBool = val;

  proc getMemberInt(): int{
    return this.memberInt;

  proc getMemberBool(): bool {
    return this.memberBool;
} // end MyClass

Call compiler-generated initializer, using default value for memberBool.

var myObject = new MyClass(10);
    myObject = new MyClass(memberInt = 10); // Equivalent

Same, but provide a memberBool value explicitly.

var myDiffObject = new MyClass(-1, true);
    myDiffObject = new MyClass(memberInt = -1,
                                memberBool = true); // Equivalent

Call the initializer we wrote.

var myOtherObject = new MyClass(1.95);
    myOtherObject = new MyClass(val = 1.95); // Equivalent

We can define an operator on our class as well, but the definition has to be outside the class definition.

proc +(A : MyClass, B : MyClass) : MyClass {
  return new MyClass(memberInt = A.getMemberInt() + B.getMemberInt(),
                      memberBool = A.getMemberBool() || B.getMemberBool());

var plusObject = myObject + myDiffObject;


delete myObject;
delete myDiffObject;
delete myOtherObject;
delete plusObject;

Classes can inherit from one or more parent classes

class MyChildClass : MyClass {
  var memberComplex: complex;

Here's an example of generic classes.

class GenericClass {
  type classType;
  var classDomain: domain(1);
  var classArray: [classDomain] classType;

Explicit constructor.

proc GenericClass(type classType, elements : int) {
  this.classDomain = {1..#elements};

Copy constructor. Note: We still have to put the type as an argument, but we can default to the type of the other object using the query (?) operator. Further, we can take advantage of this to allow our copy constructor to copy classes of different types and cast on the fly.

proc GenericClass(other : GenericClass(?otherType),
                   type classType = otherType) {
  this.classDomain = other.classDomain;
  // Copy and cast
  for idx in this.classDomain do this[idx] = other[idx] : classType;

Define bracket notation on a GenericClass object so it can behave like a normal array i.e. objVar[i] or objVar(i)

proc this(i : int) ref : classType {
  return this.classArray[i];

Define an implicit iterator for the class to yield values from the array to a loop i.e. for i in objVar do ...

  iter these() ref : classType {
    for i in this.classDomain do
      yield this[i];
} // end GenericClass

We can assign to the member array of the object using the bracket notation that we defined.

var realList = new GenericClass(real, 10);
for i in realList.classDomain do realList[i] = i + 1.0;

We can iterate over the values in our list with the iterator we defined.

for value in realList do write(value, ", ");

Make a copy of realList using the copy constructor.

var copyList = new GenericClass(realList);
for value in copyList do write(value, ", ");

Make a copy of realList and change the type, also using the copy constructor.

var copyNewTypeList = new GenericClass(realList, int);
for value in copyNewTypeList do write(value, ", ");


Modules are Chapel's way of managing name spaces. The files containing these modules do not need to be named after the modules (as in Java), but files implicitly name modules. For example, this file implicitly names the learnChapelInYMinutes module

module OurModule {

We can use modules inside of other modules. Time is one of the standard modules.

use Time;

We'll use this procedure in the parallelism section.

proc countdown(seconds: int) {
  for i in 1..seconds by -1 {

It is possible to create arbitrarily deep module nests. i.e. submodules of OurModule

  module ChildModule {
    proc foo() {

  module SiblingModule {
    proc foo() {
} // end OurModule

Using OurModule also uses all the modules it uses. Since OurModule uses Time, we also use Time.

use OurModule;

At this point we have not used ChildModule or SiblingModule so their symbols (i.e. foo) are not available to us. However, the module names are available, and we can explicitly call foo() through them.


Now we use ChildModule, enabling unqualified calls.

use ChildModule;


In other languages, parallelism is typically done with complicated libraries and strange class structure hierarchies. Chapel has it baked right into the language.

We can declare a main procedure, but all the code above main still gets executed.

proc main() {

A begin statement will spin the body of that statement off into one new task. A sync statement will ensure that the progress of the main task will not progress until the children have synced back up.

sync {
  begin { // Start of new task's body
    var a = 0;
    for i in 1..1000 do a += 1;
    writeln("Done: ", a);
  } // End of new tasks body
  writeln("spun off a task!");
writeln("Back together");

proc printFibb(n: int) {
  writeln("fibonacci(",n,") = ", fibonacci(n));

A cobegin statement will spin each statement of the body into one new task. Notice here that the prints from each statement may happen in any order.

cobegin {
  printFibb(20); // new task
  printFibb(10); // new task
  printFibb(5);  // new task
    // This is a nested statement body and thus is a single statement
    // to the parent statement, executed by a single task.
    writeln("this gets");
    writeln("executed as");
    writeln("a whole");

A coforall loop will create a new task for EACH iteration. Again we see that prints happen in any order. NOTE: coforall should be used only for creating tasks! Using it to iterating over a structure is very a bad idea!

var num_tasks = 10; // Number of tasks we want
coforall taskID in 1..#num_tasks {
  writeln("Hello from task# ", taskID);

forall loops are another parallel loop, but only create a smaller number of tasks, specifically --dataParTasksPerLocale= number of tasks.

forall i in 1..100 {
  write(i, ", ");

Here we see that there are sections that are in order, followed by a section that would not follow (e.g. 1, 2, 3, 7, 8, 9, 4, 5, 6,). This is because each task is taking on a chunk of the range 1..10 (1..3, 4..6, or 7..9) doing that chunk serially, but each task happens in parallel. Your results may depend on your machine and configuration

For both the forall and coforall loops, the execution of the parent task will not continue until all the children sync up.

forall loops are particularly useful for parallel iteration over arrays. Lets run an experiment to see how much faster a parallel loop is

use Time; // Import the Time module to use Timer objects
var timer: Timer;
var myBigArray: [{1..4000,1..4000}] real; // Large array we will write into

Serial Experiment:

timer.start(); // Start timer
for (x,y) in myBigArray.domain { // Serial iteration
  myBigArray[x,y] = (x:real) / (y:real);
timer.stop(); // Stop timer
writeln("Serial: ", timer.elapsed()); // Print elapsed time
timer.clear(); // Clear timer for parallel loop

Parallel Experiment:

timer.start(); // start timer
forall (x,y) in myBigArray.domain { // Parallel iteration
  myBigArray[x,y] = (x:real) / (y:real);
timer.stop(); // Stop timer
writeln("Parallel: ", timer.elapsed()); // Print elapsed time

You may have noticed that (depending on how many cores you have) the parallel loop went faster than the serial loop.

The bracket style loop-expression described much earlier implicitly uses a forall loop.

[val in myBigArray] val = 1 / val; // Parallel operation

Atomic variables, common to many languages, are ones whose operations occur uninterrupted. Multiple threads can therefore modify atomic variables and can know that their values are safe. Chapel atomic variables can be of type bool, int, uint, and real.

var uranium: atomic int;
uranium.write(238);      // atomically write a variable
writeln(uranium.read()); // atomically read a variable

Atomic operations are described as functions, so you can define your own.

uranium.sub(3); // atomically subtract a variable

var replaceWith = 239;
var was = uranium.exchange(replaceWith);
writeln("uranium was ", was, " but is now ", replaceWith);

var isEqualTo = 235;
if uranium.compareExchange(isEqualTo, replaceWith) {
  writeln("uranium was equal to ", isEqualTo,
           " so replaced value with ", replaceWith);
} else {
  writeln("uranium was not equal to ", isEqualTo,
           " so value stays the same...  whatever it was");

sync {
  begin { // Reader task
    writeln("Reader: waiting for uranium to be ", isEqualTo);
    writeln("Reader: uranium was set (by someone) to ", isEqualTo);

  begin { // Writer task
    writeln("Writer: will set uranium to the value ", isEqualTo, " in...");

sync variables have two states: empty and full. If you read an empty variable or write a full variable, you are waited until the variable is full or empty again.

var someSyncVar$: sync int; // varName$ is a convention not a law.
sync {
  begin { // Reader task
    writeln("Reader: waiting to read.");
    var read_sync = someSyncVar$;
    writeln("Reader: value is ", read_sync);

  begin { // Writer task
    writeln("Writer: will write in...");
    someSyncVar$ = 123;

single vars can only be written once. A read on an unwritten single results in a wait, but when the variable has a value it can be read indefinitely.

var someSingleVar$: single int; // varName$ is a convention not a law.
sync {
  begin { // Reader task
    writeln("Reader: waiting to read.");
    for i in 1..5 {
      var read_single = someSingleVar$;
      writeln("Reader: iteration ", i,", and the value is ", read_single);

  begin { // Writer task
    writeln("Writer: will write in...");
    someSingleVar$ = 5; // first and only write ever.

Heres an example using atomics and a sync variable to create a count-down mutex (also known as a multiplexer).

var count: atomic int; // our counter
var lock$: sync bool;   // the mutex lock

count.write(2);       // Only let two tasks in at a time.
lock$.writeXF(true);  // Set lock$ to full (unlocked)
// Note: The value doesnt actually matter, just the state
// (full:unlocked / empty:locked)
// Also, writeXF() fills (F) the sync var regardless of its state (X)

coforall task in 1..#5 { // Generate tasks
  // Create a barrier
  do {
    lock$;                 // Read lock$ (wait)
  } while (count.read() < 1); // Keep waiting until a spot opens up

  count.sub(1);          // decrement the counter
  lock$.writeXF(true); // Set lock$ to full (signal)

  // Actual 'work'
  writeln("Task #", task, " doing work.");

  count.add(1);        // Increment the counter
  lock$.writeXF(true); // Set lock$ to full (signal)

We can define the operations + * & | ^ && || min max minloc maxloc over an entire array using scans and reductions. Reductions apply the operation over the entire array and result in a scalar value.

var listOfValues: [1..10] int = [15,57,354,36,45,15,456,8,678,2];
var sumOfValues = + reduce listOfValues;
var maxValue = max reduce listOfValues; // 'max' give just max value

maxloc gives max value and index of the max value. Note: We have to zip the array and domain together with the zip iterator.

var (theMaxValue, idxOfMax) = maxloc reduce zip(listOfValues,

writeln((sumOfValues, maxValue, idxOfMax, listOfValues[idxOfMax]));

Scans apply the operation incrementally and return an array with the values of the operation at that index as it progressed through the array from array.domain.low to array.domain.high.

  var runningSumOfValues = + scan listOfValues;
  var maxScan = max scan listOfValues;
} // end main()