Haskell Parallel Performance

Question

The challenge is to speed up the summation using parMap or parListChunk or better. It works slower with parallelism code.

Edit: Facepalm .. I forgot how to properly execute the application.

Remember to add kernels with

./myHaskellApp paramaters +RTS -N4 -sstderr 

Where N 4 is the number of cores.

Answer 1

Always make sure that you are actually using -threaded optimized with -O2 , and that you are using a reasonable number of cores (e.g. -N4 ). Also check the statistics of garbage collectors .