is intel icpc openmp slower than icc openmp -


i doing 3d simulation of diffusion-reaction model using finite difference. system has on 8 million nodes in size. solve problem, both icc + openmp , icpc + openmp have been used. result, icc + openmp turns out 3-fold faster of icpc-openmp. huge difference in computational time confusing. asking if icpc-openmp slower icc-openmp. more information:

  1. intel compiler version: 15.0.1 20141023
  2. all reaction equations solved using c functions (not class objects).
  3. makefile options:

    a) icc+openmp:

    cc = icc cflags = -g -wall -ilib -o3 -openmp ldflags= -lz  main: $(patsubst %.c,%.o,$(wildcard lib/*.c)) 

    b) icpc+openmp:

    cxx=icpc   cflags = -wall -ilib -o3 -openmp -std=c++11   cxxflags = -wall -ilib -o3 -openmp -std=c++11  ldflags= -lz main: $(patsubst %.cpp,%.o,$(wildcard lib/*.cpp)) 

thank you!


Comments

Popular posts from this blog

android - How to save instance state of selected radiobutton on menu -

python 3 IndexError: list index out of range -

IF statement in MySQL trigger -