1. ALPS: Software Framework for Scheduling Parallel Computations with Application to Parallel Space-Time Adaptive Processing Kyusoon Lee and Adam W. Bojanczyk Cornell University Ithaca, NY, 14850 {KL224,AWB8}@cornell.edu High Performance Embedded Computing (HPEC) Workshop 18 – 20 September 2007

2. /* This is a.cpp file of the partially adaptive STAP methods that the user had in mind. This is optimized for minimum latency or maximum throughput on a given parallel machine. */ #include... Various STAP heuristics constructed from ALPS building blocks Modeling execution times of basic building blocks under various configurations Find optimal configuration that minimizes overall makespan or maximizes throughput ALPS CodeGen User’s idea on Parallel Computations (e.g. STAP) Task Graph ALPS Benchmarker & Modeler ALPS Scheduler ALPS Library ALPS Software Framework

3. Building execution time models using incremental weighted least-squares Measurements are not quite reliable Number of variables in the model is large Difficulties 00.10.20.30.40.50.6 0.7 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Relative errors from ordinary least-square Relative errors from weighted least-square 10% better 20% better 30% better 40% better 7.5% error from INV-weighted least-square 26.6% error from ordinary least-square

4. Scheduling tree-shaped task graphs on parallel computers exploiting mixed-parallelism Communication cost not negligible Arbitrary computation costs Trade-off between task-parallelism and data-parallelism Difficulties 0 10 20 30 Data- parallelism only Task- parallelism preferred Our scheduling algorithm Makespan (sec) estimatedachieved 0 0.05 0.1 Data- parallelism only Task- parallelism preferred Our scheduling algorithm Throughput (1/sec) estimated achieved