OpenMP.NET

A library for writing OpenMP-style parallel code in .NET. Inspired by the fork-join paradigm of OpenMP, and attempts to replicate the OpenMP programming style as faithfully as possible, though breaking spec at times.

Supported Constructs

Parallel

Given the OpenMP:

#pragma omp parallel
{
    work();
}

OpenMP.NET provides:

OpenMP.Parallel.ParallelRegion(() => {
    work();
});

This function supports the num_threads optional parameter, which sets the number of threads to spawn. The default value is the number of logical threads on the system.

For

Given the OpenMP:

#pragma omp for
for (int i = a, i < b; i++)
{
    work(i);
}

OpenMP.NET provides:

OpenMP.Parallel.For(a, b, i => {
    work(i);
});

This function supports the schedule optional parameter, which sets the parallel scheduler to use. Permissible values are OpenMP.Parallel.Schedule.Static, OpenMP.Parallel.Schedule.Dynamic, and OpenMP.Parallel.Schedule.Guided. The default value is OpenMP.Parallel.Schedule.Static.

This function supports the chunk_size optional parameter, which sets the chunk size for the scheduler to use. The default value is dependent on the scheduler and is not documented, as it may change from version to version.

The behavior of OpenMP.Parallel.For is undefined if not used within a ParallelRegion.

Parallel For

Given the OpenMP:

#pragma omp parallel for
for (int i = a, i < b; i++)
{
    work(i);
}

OpenMP.NET provides:

OpenMP.Parallel.ParallelFor(a, b, i => {
    work(i);
});

This function supports all of the optional parameters of ParallelRegion and For, and is merely a wrapper around those two functions for conciseness.

For with Reduction

Given the OpenMP:

type local = c;

#pragma omp for reduction(op:local)
for (int i = a; i < b; i++)
{
    local `op` d;
}

OpenMP.NET provides:

type local = c;

OpenMP.Parallel.ForReduction(a, b, op, ref local, (ref type local, int i) => {
    local `op` d;
});

op is a value provided by the OpenMP.Operations enum, which supports the values Add, Subtract, Multiply, BinaryAnd, BinaryOr, BinaryXor, BooleanAnd, BooleanOr, Min, and Max.

This function supports all of the optional parameters of For.

Parallel For with Reduction

Given the OpenMP:

type local = c;

#pragma omp parallel for reduction(op:local)
for (int i = a; i < b; i++)
{
    local `op` d;
}

OpenMP.NET provides:

type local = c;

OpenMP.Parallel.ParallelForReduction(a, b, op, ref local, (ref type local, int i) => {
    local `op` d;
});

This function supports all of the optional parameters of ParallelRegion and ForReduction, and is merely a wrapper around those two functions for conciseness.

Critical

Given the OpenMP:

#pragma omp critical
{
    work();
}

OpenMP.NET provides:

OpenMP.Parallel.Critical(id, () => {
    work();
});

This function requires an id parameter, which is used as a unique identifier for a particular critical region. If multiple critical regions are present in the code, they should each have a unique id. The id should likely be a const int or an integer literal.

Barrier

Given the OpenMP:

#pragma omp barrier

OpenMP.NET provides:

OpenMP.Parallel.Barrier();

Master

Given the OpenMP:

#pragma omp master
{
    work();
}

OpenMP.NET provides:

OpenMP.Parallel.Master(() => {
    work();
});

Master's behavior is left undefined if used outside of a For.

Single

Given the OpenMP:

#pragma omp single
{
    work();
}

OpenMP.NET provides:

OpenMP.Parallel.Single(id, () => {
    work();
});

The id parameter provided should follow the same guidelines as specified in Critical.

The behavior of a single region is as follows: the first thread in a team to reach a given Single region will "own" the Single for the duration of the ParallelRegion. On all subsequent encounters, only that first thread will execute the region. All other threads will ignore it.

Single's behavior is left undefined if used outside of a For.

Ordered

Given the OpenMP:

#pragma omp ordered
{
    work();
}

OpenMP.NET provides:

OpenMP.Parallel.Ordered(id, () => {
    work();
});

The id parameter provided should follow the same guidelines as specified in Critical.

Ordered's behavior is left undefined if used outside of a For.

Atomics

OpenMP atomics are implemented as follows:

#pragma omp atomic
a op b;

where op is some supported operator.

OpenMP.NET supports a subset of this for the int, uint, long, and ulong types. The only implemented atomic operations are a += b, a &= b, a |= b, ++a, and --a. a -= b is implemented, but for signed types only, due to restrictions interfacting with C#'s Interlocked class.

The following table documents the supported atomics:

Operation	OpenMP.NET function
`a += b`	`OpenMP.Atomic.Add(ref a, b)`
`a -= b`	`OpenMP.Atomic.Sub(ref a, b)`
`a &= b`	`OpenMP.Atomic.And(ref a, b)`
`a \|= b`	`OpenMP.Atomic.Or(ref a, b)`
`++a`	`OpenMP.Atomic.Inc(ref a)`
`--a`	`OpenMP.Atomic.Dec(ref a)`

For atomic operations like compare-exchange, we recommend interfacting directly with System.Threading.Interlocked. For non-supported atomic operations or types, we recommend using OpenMP.Parallel.Critical. This is more of a limitation of the underlying hardware than anything.

Locks

OpenMP.NET supports OpenMP-style locks. It is recommended to use C#'s native lock keyword where possible for performance. However, this API is provided to those who want the familiarity of OpenMP locks.

OpenMP.NET supports the OpenMP.Lock object, which is the replacement for omp_lock_t. omp_init_lock and omp_destroy_lock are not implemented. Instead, users should instantiate the OpenMP.Lock object using the new keyword.

OpenMP.NET provides the following functions:

<omp.h> function	OpenMP.NET function	Comments
omp_set_lock(lock)	OpenMP.Locking.Set(lock)	Halt the current thread until the lock is obtained
omp_unset_lock(lock)	OpenMP.Locking.Unset(lock)	Free the current lock, making it available for other threads
omp_test_lock(lock)	OpenMP.Locking.Test(lock)	Attempt to obtain a lock without blocking, returns true if locking is successful

Supported Functions

OpenMP.NET provides an analog of the following functions:

<omp.h> function	OpenMP.NET function	Comments
omp_get_num_procs()	OpenMP.Parallel.GetNumProcs()	Returns the number of logical threads on the system
omp_get_num_threads()	OpenMP.Parallel.GetNumThreads()	Returns the number of active threads in the current region
omp_set_num_threads(int)	OpenMP.Parallel.SetNumThreads(int)	Sets the number of threads for the next parallel region to use
omp_get_thread_num()	OpenMP.Parallel.GetThreadNum()	Gets the ID of the current thread
omp_get_max_threads()	OpenMP.Parallel.GetMaxThreads()	Gets the maximum number of threads the runtime may use in the next region
omp_in_parallel()	OpenMP.Parallel.InParallel()	Returns true if called from within a parallel region
omp_set_dynamic(int)	OpenMP.Parallel.SetDynamic()	Tells the runtime to dynamically adjust the number of threads, can disable by calling SetNumThreads
omp_get_dynamic()	OpenMP.Parallel.GetDynamic()	Returns true if the runtime can dynamically adjust the number of threads
omp_set_nested(int)	OpenMP.Parallel.SetNested(int)	Returns a NotImplementedException
omp_get_nested()	OpenMP.Parallel.GetNested()	Returns false
omp_get_wtime()	OpenMP.Parallel.GetWTime()	Returns the number of seconds since the Unix Epoch as a double

digitalmystic/OpenMP.NET

OpenMP.NET

Supported Constructs

Parallel

For

Parallel For

For with Reduction

Parallel For with Reduction

Critical

Barrier

Master

Single

Ordered

Atomics

Locks

Supported Functions