Autovectorization

This simple code demonstrates autovectorization for a vector addition routine.

Building

There are two makefiles in this repo, one for a gcc build and another for an icc build. Ensure that you have the intel compiler loaded into your environment:

setpkgs -a intel_cluster_studio_compiler 

To build run:

make -f makefile.gcc

or

make -f makefile.icc

to delete the executable you would then type:

make clean -f makefile.gcc

or

make clean -f makefile.icc

Running

The program accepts one command line argument, the length of the vector. For example:

./vec_add_icc 1000

Exercises

1. Build the binary with and without vectorization enabled using both icc and gcc. Building with level-three optimization will ensure autovectorization is run for both compilers. Time the execution of both versions of the binary for a vector length of 1000000000.

2. Compare the walltime with versions of the binary that are built without vectorization enabled. Using optimization level zero will ensure that vectorization is disabled.

3. Move the actual vector addition step (c = a + b) into a second for loop. With level three optimization enabled, how does the autovectorization behavior compare (read reporting carefully)? How does the walltime compare?

4. Revert back to the original code. Now start the loop at i=1 (ignore the first element for the vector addition) and add this line to the end of the loop body:

c[i] = c[i-1] + 82.3;

What changes? Why?