In case anyone here would be interested, we did some instrumentation on Mothur with OpenMP and saw orders of magnitude speed-up on cluter command and others. Here is the link to the proceeding paper: http://dl.acm.org/citation.cfm?id=2616505
BTW, XSEDE (successor of TeraGrid) resources are open to university researchers, and currently the approval rate for a start-up proposal (10,000 ~ 100,000 cpu hours) is close to 100%. You’re welcome to take advantage of national computational resources. Here is more details on the allocation process: https://portal.xsede.org/allocations-overview