The main takeaway from this slide is that relaxing W->R ordering constraint allows us to improve performance by up to 30% on some applications due to the potential to interleave memory accesses to hide latency, but the result is a system that is no longer sequentially consistent anymore.
The main takeaway from this slide is that relaxing W->R ordering constraint allows us to improve performance by up to 30% on some applications due to the potential to interleave memory accesses to hide latency, but the result is a system that is no longer sequentially consistent anymore.