Tasks and Parallel Computing

Tasks

Task(func)

Create a Task (i.e. thread, or coroutine) to execute the given function (which must be callable with no arguments). The task exits when this function returns.

yieldto(task, arg = nothing)

Switch to the given task. The first time a task is switched to, the task’s function is called with no arguments. On subsequent switches, arg is returned from the task’s last call to yieldto. This is a low-level call that only switches tasks, not considering states or scheduling in any way. Its use is discouraged.

current_task()

Get the currently running Task.

istaskdone(task) → Bool

Tell whether a task has exited.

istaskstarted(task) → Bool

Tell whether a task has started executing.

consume(task, values...)

Receive the next value passed to produce by the specified task. Additional arguments may be passed, to be returned from the last produce call in the producer.

produce(value)

Send the given value to the last consume call, switching to the consumer task. If the next consume call passes any values, they are returned by produce.

yield()

Switch to the scheduler to allow another scheduled task to run. A task that calls this function is still runnable, and will be restarted immediately if there are no other runnable tasks.

task_local_storage(symbol)

Look up the value of a symbol in the current task’s task-local storage.

task_local_storage(symbol, value)

Assign a value to a symbol in the current task’s task-local storage.

task_local_storage(body, symbol, value)

Call the function body with a modified task-local storage, in which value is assigned to symbol; the previous value of symbol, or lack thereof, is restored afterwards. Useful for emulating dynamic scoping.

Condition()

Create an edge-triggered event source that tasks can wait for. Tasks that call wait on a Condition are suspended and queued. Tasks are woken up when notify is later called on the Condition. Edge triggering means that only tasks waiting at the time notify is called can be woken up. For level-triggered notifications, you must keep extra state to keep track of whether a notification has happened. The Channel type does this, and so can be used for level-triggered events.

notify(condition, val=nothing; all=true, error=false)

Wake up tasks waiting for a condition, passing them val. If all is true (the default), all waiting tasks are woken, otherwise only one is. If error is true, the passed value is raised as an exception in the woken tasks.

schedule(t::Task, [val]; error=false)

Add a task to the scheduler’s queue. This causes the task to run constantly when the system is otherwise idle, unless the task performs a blocking operation such as wait.

If a second argument is provided, it will be passed to the task (via the return value of yieldto) when it runs again. If error is true, the value is raised as an exception in the woken task.

@schedule()

Wrap an expression in a Task and add it to the local machine’s scheduler queue.

@task()

Wrap an expression in a Task without executing it, and return the Task. This only creates a task, and does not run it.

sleep(seconds)

Block the current task for a specified number of seconds. The minimum sleep time is 1 millisecond or input of 0.001.

ReentrantLock()

Creates a reentrant lock. The same task can acquire the lock as many times as required. Each lock must be matched with an unlock.

lock(l::ReentrantLock)

Associates l with the current task. If l is already locked by a different task, waits for it to become available. The same task can acquire the lock multiple times. Each “lock” must be matched by an “unlock”

unlock(l::ReentrantLock)

Releases ownership of the lock by the current task. If the lock had been acquired before, it just decrements an internal counter and returns immediately.

Channel{T}(sz::Int)

Constructs a Channel that can hold a maximum of sz objects of type T. put! calls on a full channel block till an object is removed with take!.

Other constructors:

  • Channel() - equivalent to Channel{Any}(32)
  • Channel(sz::Int) equivalent to Channel{Any}(sz)

General Parallel Computing Support

addprocs(n::Integer; exeflags=``) → List of process identifiers

Launches workers using the in-built LocalManager which only launches workers on the local host. This can be used to take advantage of multiple cores. addprocs(4) will add 4 processes on the local machine.

addprocs() → List of process identifiers

Equivalent to addprocs(CPU_CORES)

Note that workers do not run a .juliarc.jl startup script, nor do they synchronize their global state (such as global variables, new method definitions, and loaded modules) with any of the other running processes.

addprocs(machines; keyword_args...) → List of process identifiers

Add processes on remote machines via SSH. Requires julia to be installed in the same location on each node, or to be available via a shared file system.

machines is a vector of machine specifications. Worker are started for each specification.

A machine specification is either a string machine_spec or a tuple - (machine_spec, count).

machine_spec is a string of the form [user@]host[:port] [bind_addr[:port]]. user defaults to current user, port to the standard ssh port. If [bind_addr[:port]] is specified, other workers will connect to this worker at the specified bind_addr and port.

count is the number of workers to be launched on the specified host. If specified as :auto it will launch as many workers as the number of cores on the specific host.

Keyword arguments:

  • tunnel: if true then SSH tunneling will be used to connect to the worker from the master process. Default is false.
  • sshflags: specifies additional ssh options, e.g.
sshflags=`-i /home/foo/bar.pem`
  • max_parallel: specifies the maximum number of workers connected to in parallel at a host. Defaults to 10.
  • dir: specifies the working directory on the workers. Defaults to the host’s current directory (as found by pwd())
  • exename: name of the julia executable. Defaults to "$JULIA_HOME/julia" or "$JULIA_HOME/julia-debug" as the case may be.
  • exeflags: additional flags passed to the worker processes.
  • topology: Specifies how the workers connect to each other. Sending a message between unconnected workers results in an error.
  • topology=:all_to_all : All processes are connected to each other. This is the default.
  • topology=:master_slave : Only the driver process, i.e. pid 1 connects to the workers. The workers do not connect to each other.
  • topology=:custom : The launch method of the cluster manager specifes the connection topology via fields ident and connect_idents in WorkerConfig. A worker with a cluster manager identity ident will connect to all workers specified in connect_idents.

Environment variables :

If the master process fails to establish a connection with a newly launched worker within 60.0 seconds, the worker treats it a fatal situation and terminates. This timeout can be controlled via environment variable JULIA_WORKER_TIMEOUT. The value of JULIA_WORKER_TIMEOUT on the master process, specifies the number of seconds a newly launched worker waits for connection establishment.

addprocs(manager::ClusterManager; kwargs...) → List of process identifiers

Launches worker processes via the specified cluster manager.

For example Beowulf clusters are supported via a custom cluster manager implemented in package ClusterManagers.

The number of seconds a newly launched worker waits for connection establishment from the master can be specified via variable JULIA_WORKER_TIMEOUT in the worker process’s environment. Relevant only when using TCP/IP as transport.

nprocs()

Get the number of available processes.

nworkers()

Get the number of available worker processes. This is one less than nprocs(). Equal to nprocs() if nprocs() == 1.

procs()

Returns a list of all process identifiers.

workers()

Returns a list of all worker process identifiers.

rmprocs(pids...)

Removes the specified workers.

interrupt([pids...])

Interrupt the current executing task on the specified workers. This is equivalent to pressing Ctrl-C on the local machine. If no arguments are given, all workers are interrupted.

myid()

Get the id of the current process.

pmap(f, lsts...; err_retry=true, err_stop=false, pids=workers())

Transform collections lsts by applying f to each element in parallel. (Note that f must be made available to all worker processes; see Code Availability and Loading Packages for details.) If nprocs() > 1, the calling process will be dedicated to assigning tasks. All other available processes will be used as parallel workers, or on the processes specified by pids.

If err_retry is true, it retries a failed application of f on a different worker. If err_stop is true, it takes precedence over the value of err_retry and pmap stops execution on the first error.

remotecall(id, func, args...)

Call a function asynchronously on the given arguments on the specified process. Returns a RemoteRef.

wait([x])

Block the current task until some event occurs, depending on the type of the argument:

  • RemoteRef: Wait for a value to become available for the specified remote reference.
  • Channel: Wait for a value to be appended to the channel.
  • Condition: Wait for notify on a condition.
  • Process: Wait for a process or process chain to exit. The exitcode field of a process can be used to determine success or failure.
  • Task: Wait for a Task to finish, returning its result value. If the task fails with an exception, the exception is propagated (re-thrown in the task that called wait).
  • RawFD: Wait for changes on a file descriptor (see poll_fd for keyword arguments and return code)

If no argument is passed, the task blocks for an undefined period. If the task’s state is set to :waiting, it can only be restarted by an explicit call to schedule or yieldto. If the task’s state is :runnable, it might be restarted unpredictably.

Often wait is called within a while loop to ensure a waited-for condition is met before proceeding.

fetch(x)

Waits and fetches a value from x depending on the type of x. Does not remove the item fetched:

  • RemoteRef: Wait for and get the value of a remote reference. If the remote value is an exception, throws a RemoteException which captures the remote exception and backtrace.
  • Channel : Wait for and get the first available item from the channel.
remotecall_wait(id, func, args...)

Perform wait(remotecall(...)) in one message.

remotecall_fetch(id, func, args...)

Perform fetch(remotecall(...)) in one message. Any remote exceptions are captured in a RemoteException and thrown.

put!(RemoteRef, value)

Store a value to a remote reference. Implements “shared queue of length 1” semantics: if a value is already present, blocks until the value is removed with take!. Returns its first argument.

put!(Channel, value)

Appends an item to the channel. Blocks if the channel is full.

take!(RemoteRef)

Fetch the value of a remote reference, removing it so that the reference is empty again.

take!(Channel)

Removes and returns a value from a Channel. Blocks till data is available.

isready(r::RemoteRef)

Determine whether a RemoteRef has a value stored to it. Note that this function can cause race conditions, since by the time you receive its result it may no longer be true. It is recommended that this function only be used on a RemoteRef that is assigned once.

If the argument RemoteRef is owned by a different node, this call will block to wait for the answer. It is recommended to wait for r in a separate task instead, or to use a local RemoteRef as a proxy:

rr = RemoteRef()
@async put!(rr, remotecall_fetch(p, long_computation))
isready(rr)  # will not block
close(Channel)

Closes a channel. An exception is thrown by:

  • put! on a closed channel.
  • take! and fetch on an empty, closed channel.
RemoteRef()

Make an uninitialized remote reference on the local machine.

RemoteRef(n)

Make an uninitialized remote reference on process n.

timedwait(testcb::Function, secs::Float64; pollint::Float64=0.1)

Waits till testcb returns true or for secs seconds, whichever is earlier. testcb is polled every pollint seconds.

@spawn()

Creates a closure around an expression and runs it on an automatically-chosen process, returning a RemoteRef to the result.

@spawnat()

Accepts two arguments, p and an expression. A closure is created around the expression and run asynchronously on process p. Returns a RemoteRef to the result.

@fetch()

Equivalent to fetch(@spawn expr).

@fetchfrom()

Equivalent to fetch(@spawnat p expr).

@async()

Like @schedule, @async wraps an expression in a Task and adds it to the local machine’s scheduler queue. Additionally it adds the task to the set of items that the nearest enclosing @sync waits for. @async also wraps the expression in a let x=x, y=y, ... block to create a new scope with copies of all variables referenced in the expression.

@sync()

Wait until all dynamically-enclosed uses of @async, @spawn, @spawnat and @parallel are complete. All exceptions thrown by enclosed async operations are collected and thrown as a CompositeException.

@parallel()

A parallel for loop of the form :

@parallel [reducer] for var = range
    body
end

The specified range is partitioned and locally executed across all workers. In case an optional reducer function is specified, @parallel performs local reductions on each worker with a final reduction on the calling process.

Note that without a reducer function, @parallel executes asynchronously, i.e. it spawns independent tasks on all available workers and returns immediately without waiting for completion. To wait for completion, prefix the call with @sync, like :

@sync @parallel for var = range
    body
end
@everywhere()

Execute an expression on all processes. Errors on any of the processes are collected into a CompositeException and thrown.

Shared Arrays

SharedArray(T::Type, dims::NTuple; init=false, pids=Int[])

Construct a SharedArray of a bitstype T and size dims across the processes specified by pids - all of which have to be on the same host.

If pids is left unspecified, the shared array will be mapped across all processes on the current host, including the master. But, localindexes and indexpids will only refer to worker processes. This facilitates work distribution code to use workers for actual computation with the master process acting as a driver.

If an init function of the type initfn(S::SharedArray) is specified, it is called on all the participating workers.

SharedArray(filename::AbstractString, T::Type, dims::NTuple, [offset=0]; mode=nothing, init=false, pids=Int[])

Construct a SharedArray backed by the file filename, with element type T (must be a bitstype) and size dims, across the processes specified by pids - all of which have to be on the same host. This file is mmapped into the host memory, with the following consequences:

  • The array data must be represented in binary format (e.g., an ASCII format like CSV cannot be supported)
  • Any changes you make to the array values (e.g., A[3] = 0) will also change the values on disk

If pids is left unspecified, the shared array will be mapped across all processes on the current host, including the master. But, localindexes and indexpids will only refer to worker processes. This facilitates work distribution code to use workers for actual computation with the master process acting as a driver.

mode must be one of "r", "r+", "w+", or "a+", and defaults to "r+" if the file specified by filename already exists, or "w+" if not. If an init function of the type initfn(S::SharedArray) is specified, it is called on all the participating workers. You cannot specify an init function if the file is not writable.

offset allows you to skip the specified number of bytes at the beginning of the file.

procs(S::SharedArray)

Get the vector of processes that have mapped the shared array

sdata(S::SharedArray)

Returns the actual Array object backing S

indexpids(S::SharedArray)

Returns the index of the current worker into the pids vector, i.e., the list of workers mapping the SharedArray

localindexes(S::SharedArray)

Returns a range describing the “default” indexes to be handled by the current process. This range should be interpreted in the sense of linear indexing, i.e., as a sub-range of 1:length(S). In multi-process contexts, returns an empty range in the parent process (or any process for which indexpids returns 0).

It’s worth emphasizing that localindexes exists purely as a convenience, and you can partition work on the array among workers any way you wish. For a SharedArray, all indexes should be equally fast for each worker process.

Cluster Manager Interface

This interface provides a mechanism to launch and manage Julia workers on different cluster environments. LocalManager, for launching additional workers on the same host and SSHManager, for launching on remote hosts via ssh are present in Base. TCP/IP sockets are used to connect and transport messages between processes. It is possible for Cluster Managers to provide a different transport.

launch(manager::FooManager, params::Dict, launched::Vector{WorkerConfig}, launch_ntfy::Condition)

Implemented by cluster managers. For every Julia worker launched by this function, it should append a WorkerConfig entry to launched and notify launch_ntfy. The function MUST exit once all workers, requested by manager have been launched. params is a dictionary of all keyword arguments addprocs was called with.

manage(manager::FooManager, pid::Int, config::WorkerConfig. op::Symbol)

Implemented by cluster managers. It is called on the master process, during a worker’s lifetime, with appropriate op values:

  • with :register/:deregister when a worker is added / removed from the Julia worker pool.
  • with :interrupt when interrupt(workers) is called. The ClusterManager should signal the appropriate worker with an interrupt signal.
  • with :finalize for cleanup purposes.
kill(manager::FooManager, pid::Int, config::WorkerConfig)

Implemented by cluster managers. It is called on the master process, by rmprocs. It should cause the remote worker specified by pid to exit. Base.kill(manager::ClusterManager.....) executes a remote exit() on pid

init_worker(manager::FooManager)

Called by cluster managers implementing custom transports. It initializes a newly launched process as a worker. Command line argument --worker has the effect of initializing a process as a worker using TCP/IP sockets for transport.

connect(manager::FooManager, pid::Int, config::WorkerConfig) -> (instrm::AsyncStream, outstrm::AsyncStream)

Implemented by cluster managers using custom transports. It should establish a logical connection to worker with id pid, specified by config and return a pair of AsyncStream objects. Messages from pid to current process will be read off instrm, while messages to be sent to pid will be written to outstrm. The custom transport implementation must ensure that messages are delivered and received completely and in order. Base.connect(manager::ClusterManager.....) sets up TCP/IP socket connections in-between workers.

Base.process_messages(instrm::AsyncStream, outstrm::AsyncStream)

Called by cluster managers using custom transports. It should be called when the custom transport implementation receives the first message from a remote worker. The custom transport must manage a logical connection to the remote worker and provide two AsyncStream objects, one for incoming messages and the other for messages addressed to the remote worker.