Search

The many-to-many assignment problem (M- MAP), and the CPU/FPGA scheduling problem are two correlated issues in the field of combinatorial optimization. The framework for Kernighan Lin-driven logarithmic barrier approach (KD-LBA) is made to solve the many-to-many assignment problem. KD-LBA is a deterministic technique, which initiates to achieve the globally optimal solutions for MMAP using logarithmic barrier function-based gradient descent technique. Then, the obtained solution is optimized further using the Kernighan Lin-based local search method. Successive Kernighan Lin-driven logarithmic barrier approach (Successive KD-LBA) is also proposed to sort the issue of scheduling in a CPU/FPGA heterogeneous system. It solves the CPU/ FPGA scheduling problem by transforming it into an MMAP. KD-LBA outperforms the state-of-art methods in terms of convergence speed for MMAPs with a group size greater than 40. Successive KD-LBA presents novel scheduling solutions for the CPU/FPGA scheduling problem as compared to the existing works, regarding average makespan and computation time.

We present DI-MMAP, a high-performance runtime that memory-maps large external data sets into an application's address space and shows significantly better performance than the Linux mmap system call. Our implementation is particularly effective when used with high performance locally attached Flash arrays on highly concurrent, latency-tolerant data-intensive HPC applications. We describe the kernel module and show performance results on a benchmark test suite, a new bioinformatics metagenomic classification application, and on a level-asynchronous Breadth-First Search (BFS) graph traversal algorithm. Using DI-MMAP, the metagenomics classification application performs up to 4× better than standard Linux mmap. A fully external memory configuration of BFS executes up to 7.44× faster than traditional mmap. Finally, we demonstrate that DI-MMAP shows scalable out-of-core performance for BFS traversal in main memory constrained scenarios. Such scalable memory constrained performance would allow a system with a fixed amount of memory to solve a larger problem as well as provide memory QoS guarantees for systems running multiple data-intensive applications.