Daniel Cordeiro

TL;DR: NucoLB is introduced, a topology-aware load balancer that focuses on redistributing work while reducing communication costs among and within compute nodes and takes the asymmetric memory access costs present on NUMA multi-core compute nodes, the interconnection network overheads, and the application communication patterns into account in its balancing decisions.

TL;DR: This work presents a simple and intuitive BSP-based model for predicting the CUDA application execution times on GPUs, based on the number of computations and memory accesses of the GPU, with additional information on cache usage obtained from profiling.

TL;DR: This paper proposes a flexible and extensible middleware solution that addresses the challenges present in the large-scale deployment of service compositions and describes the middleware architecture and implementation, and presents experimental results demonstrating the feasibility of the approach.

TL;DR: It is proved that selfish behavior imposes a lower bound of 2 on the approximation ratio for the global makespan, and various approximation algorithms of ratio 2 are presented which validate selfishness restrictions.