Chapter68:ParallelLINQ(PLINQ)

Section68.1:Simpleexample

ThisexampleshowshowPLINQcanbeusedtocalculatetheevennumbersbetween1and10,000usingmultiple threads. Note that the resulting list will won’t be ordered!

varsequence=Enumerable.Range(1,10000);

varevenNumbers=sequence.AsParallel()

.Where(x=>x%2==0)

.ToList();

//evenNumbers={4,26,28,30,…}

//Orderwillvarywithdifferentruns

Section68.2:WithDegreeOfParallelism

The degree of parallelism is the maximum number of concurrently executing tasks that will be used to process the query.

varsequence=Enumerable.Range(1,10000);

varevenNumbers=sequence.AsParallel()

.WithDegreeOfParallelism(4)

.Where(x=>x%2==0);

Section68.3:AsOrdered

ThisexampleshowshowPLINQcanbeusedtocalculatetheevennumbersbetween1and10,000usingmultiple threads.Orderwillbemaintainedintheresultinglist,howeverkeepinmindthatAsOrderedmayhurtperformance foralargenumbersofelements,soun-orderedprocessingispreferredwhenpossible.

varsequence=Enumerable.Range(1,10000);

varevenNumbers=sequence.AsParallel()

.AsOrdered()

.Where(x=>x%2==0)

.ToList();

//evenNumbers={2,4,6,8,…,10000}

Section68.4:AsUnordered

Ordered sequences may hurt performance when dealing with a large number of elements. To mitigate this, it’s possible to call AsUnorderedwhen the sequence order is no longer necessary.

varsequence=Enumerable.Range(1,10000).Select(x=>-1*x);//-1,-2,…

varevenNumbers=sequence.AsParallel()

.OrderBy(x=>x)

.Take(5000)

.AsUnordered()

.Where(x=>x%2==0)//Thislinewon’tbeaffectedbyordering

.ToList();