Conference Program

All sessions are scheduled according to the Brazilian Daylight Saving Time (GMT-2).

Complete program in PDF version for download here.

SBAC-PAD papers with abstracts here.

SBAC-PAD Sessions

Sunday, October 18. 3 pm - 4 pm

ACC1: Applications and algorithms using hardware accelerators

  • Towards Seismic Wave Modeling on Heterogeneous Many-core Architectures using Task-based Runtime System
  • Optimized Parallel Label Propagation based Community Detection on the Intel® Xeon Phi™ Architecture

Monday, October 19. 10:30 am - 11:30 am

ACC2: Applications and algorithms using hardware accelerators

  • GPU-accelerated High-speed Eye Pupil Tracking System
  • Efficient Irregular Wavefront Propagation Algorithms on Intel Xeon Phi

Monday, October 19. 2 pm - 4 pm

Arch: New architectures and hardware mechanisms to improve performance

  • Performance and Energy Efficient Hardware-based Scheduler for Symmetric/Asymmetric CMPs
  • Analysis and Optimization of Engines for Dynamically Typed Languages
  • Memory Centric Computation (mc2) for Large-scale Graph Processing
  • Progressive Codesign of an Architecture and Compiler using a Proxy Application

Monday, October 19. 5:30 pm - 7 pm

Mem: Memory systems and optimizations

  • Tidy Cache: Improving Data Placement in Die-stacked DRAM Caches
  • Unifying Router Power Gating with Data Placement for Energy-Efficient NoC
  • i-MIRROR: A Software Managed Die-Stacked DRAM-Based Memory Subsystem

Tuesday, October 20. 10:30 am - 11:30 pm

Code Opt: Code optimization

  • Function Call Fusion
  • OpenCL Kernel Fusion for GPU, Xeon Phi and CPU

Tuesday, October 20. 2 pm - 4 pm

Eval: System characterization and performance evaluation

  • WattWatcher: Fine-Grained Power Estimation For Emerging Workloads
  • Performance Characterization of Modern Databases on Out-of-order CPUs
  • Cloud Services Evaluation through QoE: a Methodological Approach
  • Non-stationary Simulation of Computer Systems and Dynamic Performance Evaluation: a Concern-based Approach and Case Study on Cloud Computing
  • Serialization Management for Best-Effort Hardware Transactional Memory: A key for performance

Tuesday, October 20. 5:30 pm - 7 pm

FT & Storage: Fault tolerance and cloud storage

  • Exploring Energy-Consistency Trade-off in Cassandra Cloud Storage System
  • COMET: Client-Oriented Metadata Service for Highly Available Distributed File Systems
  • A Fault-Tolerance Protocol for Parallel Applications with Communication Imbalance

Wednesday, October 21. 10:30 am - 11:30 am

Sched & VM-1: Scheduling and virtual machines

  • Comparison of Static and Runtime Resource Allocation Strategies for Matrix Multiplication
  • Device-Sensitive Framework for Handling Heterogeneous Asymmetric Clusters Efficiently

Wednesday, October 21. 2 pm - 3:30 pm

Sched & VM-2: Scheduling and virtual machines

  • Evaluating the Impact of Memory Allocation and Swap for Vertical Memory Elasticity in VMs
  • Quantum Virtual Machine: a Scalable Model to Optimize Energy Savings and Resource Management
  • A Programming Interface for Overload Control in Staged Event-based Architectures