Fiber::ExecutionContext::Parallel
Inherits Fiber::ExecutionContext / Reference / Object
Parallel execution context.
Fibers running in the same context run both concurrently and in parallel to each others, in addition to the other fibers running in other execution contexts.
The context internally keeps a number of fiber schedulers, each scheduler runs on a system thread, so multiple schedulers can run in parallel. The fibers are resumable by any scheduler in the context, and can thus move from one system thread to another at any time.
The actual parallelism is dynamic. As the need for parallelism increases, for example more fibers running longer, the more schedulers will start (and thus system threads), as the need decreases, for example not enough fibers, the schedulers will pause themselves and parallelism will decrease.
The parallelism can be as low as 1, in which case the context becomes a concurrent context (no parallelism) until resized.
For example: we can start a parallel context to run consumer fibers, while
the default context produces values. Because the consumer fibers can run in
parallel, we must protect accesses to the shared value variable. Running
the example without Atomic#add would produce a different result every
time!
require "wait_group"
consumers = Fiber::ExecutionContext::Parallel.new("consumers", 8)
channel = Channel(Int32).new(64)
wg = WaitGroup.new(32)
result = Atomic.new(0)
32.times do
consumers.spawn do
while value = channel.receive?
result.add(value)
end
ensure
wg.done
end
end
1024.times { |i| channel.send(i) }
channel.close
# wait for all workers to be done
wg.wait
p result.get # => 523776
NOTE: The Parallel execution context isn't tied to a fixed set of system
threads, and execution can switch to other system threads, for example when
a fiber is blocked on a syscall.
Constructors
Starts a Parallel context with a maximum parallelism. The context
starts with an initial parallelism of zero. It will grow to one when a
fiber is spawned, then the actual parallelism will keep increasing and
decreasing as needed, but will never go past the configured maximum.
DEPRECATED Use Fiber::ExecutionContext::Parallel.new(String, Int32) instead.
Instance methods
The maximum number of schedulers that can be started, aka how many fibers can run in parallel or maximum parallelism of the context.
Appends a String representation of this object which includes its class name, its object address and the values of all instance variables.
class Person
def initialize(@name : String, @age : Int32)
end
end
Person.new("John", 32).inspect # => #<Person:0x10fd31f20 @name="John", @age=32>
Resizes the context to the new maximum parallelism.
The new maximum can grow, in which case more schedulers are created to eventually increase the parallelism.
The new maximum can also shrink, in which case the overflow schedulers are removed and told to shutdown immediately. The actual shutdown is cooperative, so running schedulers won't stop until their current fiber tries to switch to another fiber.