The D0 detector underwent an upgrade to its silicon vertex detector and triggering systems during the transition from Run IIa to Run IIb to maximize its ability to fully exploit Run II at the Fermilab Tevatron. This thesis describes improvements made to the tracking and vertexing algorithms used by the high level trigger in both Run IIa and Run IIb, as well as a search for resonant di-J/{psi} states using both Run IIa and Run IIb data. Improvements made to the tracking and vertexing algorithms during Run IIa included the optimization of the existing tracking software to reduce overall processing time and the certification and testing of a new software release. Upgrades made to the high level trigger for Run IIb included the development of a new tracking algorithm and the inclusion of the new Layer 0 silicon detector into the existing software. The integration of Layer 0 into the high level trigger has led to an improvement in the overall impact parameter resolution for tracks of {approx}50%. The development of a new parameterization method for finding the error associated to the impact parameter of tracks returned by the high level tracking algorithm, in association with the inclusion of Layer …

We perform a search for {nu}{sub {mu}} {yields} {nu}{sub e} oscillations, a process which would manifest a nonzero value of the {theta}{sub 13} mixing angle, in the MINOS long-baseline neutrino oscillation experiment. The analysis consists of searching for an excess of {nu}{sub e} charged-current candidate events over the predicted backgrounds, made mostly of neutral-current events with high electromagnetic content. A novel technique to select electron neutrino events is developed, which achieves an improved separation between the signal and the backgrounds, and which consequently yields a better reach in {theta}{sub 13}. The backgrounds are predicted in the Far Detector from Near Detector measurements. An excess is observed in the Far Detector data over the predicted backgrounds, which is consistent with the background-only hypothesis at 1.2 standard deviations.

Concurrency levels in large-scale, distributed-memory supercomputers are rising exponentially. Modern machines may contain 100,000 or more microprocessor cores, and the largest of these, IBM's Blue Gene/L, contains over 200,000 cores. Future systems are expected to support millions of concurrent tasks. In this dissertation, we focus on efficient techniques for measuring and analyzing the performance of applications running on very large parallel machines. Tuning the performance of large-scale applications can be a subtle and time-consuming task because application developers must measure and interpret data from many independent processes. While the volume of the raw data scales linearly with the number of tasks in the running system, the number of tasks is growing exponentially, and data for even small systems quickly becomes unmanageable. Transporting performance data from so many processes over a network can perturb application performance and make measurements inaccurate, and storing such data would require a prohibitive amount of space. Moreover, even if it were stored, analyzing the data would be extremely time-consuming. In this dissertation, we present novel methods for reducing performance data volume. The first draws on multi-scale wavelet techniques from signal processing to compress systemwide, time-varying load-balance data. The second uses statistical sampling to select a small …