Creating a custom partitioning strategy

Partitioning is another important concept in data parallelism. To achieve parallelism in the source collection, it needs to be partitioned into smaller sections called ranges or chunks, which can be concurrently accessed by various threads. Without partitioning, the loop will execute serially. Partitioners can be classified into two categories and we can create custom partitioners as well. These categories are as follows:

• Range partitioning
• Chunk partitioning

Let's discuss these in detail.

Range partitioning

This type of partitioning is primarily used with collections where the length is known in advance. As the name suggests, every thread gets a range of elements to process or...