"Repeated" benchmarks in pthread-wmm

[hernanponcedeleon]

Half of the benchmarks in this folder are tagged *.opt and half *.oepc.
According to the README these are two optimisations to minimise the amount of instrumentation.
These benchmarks were probably generated automatically and thus there were always two versions.
However, after inspection, MANY of the benchmarks only differ in the pthread_id, e.g.

diff safe010_power.opt.c safe010_power.oepc.c
209,212c209,212
<   pthread_t t2011;
<   pthread_create(&t2011, NULL, P0, NULL);
<   pthread_t t2012;
<   pthread_create(&t2012, NULL, P1, NULL);
---
>   pthread_t t2009;
>   pthread_create(&t2009, NULL, P0, NULL);
>   pthread_t t2010;
>   pthread_create(&t2010, NULL, P1, NULL);

and thus such benchmarks can be considered duplicated.
I'm planning to make a PR removing these duplications, but before going to the pain of this work, I wanted to be sure this would be the actions to be taken.