I had a similar realization when I read on STM performance[1], where almost all benchmarks showed considerable slowdowns for the parallel+stm implementations wh

I had a similar realization when I read on STM performance[1], where almost all benchmarks showed considerable slowdowns for the parallel+stm implementations when compared against sequential solutions. [1]: http://queue.acm.org/detail.cfm?id=1454466 http://queue.acm.org/detail.cfm?id=1454466