Time and again data privacy and computer security are severely threatened by software vulnerabilities. With more and more computing devices getting network connectivity, the exploitation of vulnerable programs by remote users has become a ubiquitous issue. While tremendous effort is carried out to counter software exploitation, no single approach developed so far has been able to satisfactorily provide solid security along with efficient performance. Taint analysis, which is one of the approaches to track information flow to counter program exploits, has shown promise. In this thesis work, we propose a novel implementation of fine-grained dynamic vulnerability detection by parallelizing the actual computation and taint computation and tapping the power of idle cores on multicore machines to mitigate performance overheads. We propose a paradigm of secure and efficient computing using many cores and use binary rewriting to empower a program with parallelized taint monitoring capability. The challenge lies in minimizing the thread synchronization. We demonstrate the effectiveness of our approach in protecting against various attacks while offering an order of magnitude performance improvement compared to state of the art approaches.