Granularity (parallel computing)

In <a href="/facts/Parallel_computing/nQzcpzQt">parallel computing</a>, granularity (or <a href="/facts/Granulation/FWSEm4ne">grain</a> size) of a <a href="/facts/Task_(computing)/ghdT0IkM">task</a> is a measure of the amount of work (or <a href="/facts/Computation/qLsELosq">computation</a>) which is performed by that task.
Another definition of granularity takes into account the communication <a href="/facts/Overhead_(computing)/e2jgPrHr">overhead</a> between multiple <a href="/facts/Processor_(computing)/z3wPljrI">processors</a> or processing elements. It defines granularity as the ratio of computation time to communication time, wherein computation time is the time required to perform the computation of a task and communication time is the time required to exchange data between processors.
If Tcomp is the computation time and Tcomm denotes the communication time, then the granularity G of a task can be calculated as:

G
        =
        
          
            
              T
              
                
                  c
                  o
                  m
                  p
                
              
            
            
              T
              
                
                  c
                  o
                  m
                  m
                
              
            
          
        
      
    
    {\displaystyle G={\frac {T_{\mathrm {comp} }}{T_{\mathrm {comm} }}}}

Granularity is usually measured in terms of the number of <a href="/facts/Instruction_set_architecture/8p8vwUeE">instructions</a> which are <a href="/facts/Execution_(computing)/1fyfGMFm">executed</a> in a particular task. Alternately, granularity can also be specified in terms of the execution time of a program, combining the computation time and communication time.

Granularity (parallel computing) open-in-new

Granularity (parallel computing)