11th Nordic Workshop on Multi-Core Computing (MCC2018)

Objective

The objective of MCC is to bring together Nordic researchers and practitioners from academia and industry to present and discuss recent work in the area of multi-core computing. This year's edition is hosted by Chalmers University of Technology (Gothenburg, Sweden).

Scope

The scope of the workshop is both hardware and software aspects of multi-core computing, including design and development as well as practical usage of systems. The topics of interest include, but is not limited to, the following:

• Architecture of multi-core processors, GPUs, accelerators, heterogeneous systems, memory systems, interconnects and on-chip networks
• Parallel programming models, languages, environments
• Parallel algorithms and applications
• Compiler optimizations and techniques for multi-core systems
• Hardware/software design trade-offs in multi-core systems
• Operating system, middleware, and run-time system support for multi-core systems
• Correctness and performance analysis of parallel hardware and software
• Tools and methods for development and evaluation of multi-core systems

Venue Info

Chalmers University of Technology

Department of Computer Science and Engineering

EDIT Building, Room ED, Floor 5

SE-41296, Göteborg

Instructions for reaching the venue

Program

Day1 - Nov 29th

Start Time - Type (Duration in mins.)

8:45 - Registration opens

9:00 - Welcome (15) Madhavan Manivannan and Miquel Pericàs

9:15 - Keynote1 (45+15)

Stefanos Kaxiras, Uppsala University, Keynote Title: Consistency models and their implications on core architecture

Abstract: The memory consistency model has important implications on the microarchitecture of out-of-order cores. In this talk we discuss recent advances and in particular non-speculative reordering of loads and non-speculative store coalescing and show how such advances can change the way we design cores.

10:15 - Coffee Break (15)

10:30 - Session1 (60)

Futhark: Purely Functional GPU-Programming with Nested Parallelism and In-Place Array Updates. Troels Henriksen, Niels Gustav Westphal Serup, Martin Elsman, Fritz Henglein and Cosmin Oancea

RVSDG: An Intermediate Representation for the Multi-Core Era. Nico Reissmann, Jan Christian Meyer and Magnus Sjalander

Towards global optimization of operand transfer fusion in heterogeneous computing, Christoph Kessler

A Register File Organization to Support Variable Floating-Point Precision in GPUs, Alexandra Angerd, Per Stenström and Erik Sintorn

11:30 - Break (15)

11:45 - Session2 (60)

Odd-ECC: On-demand DRAM Error Correcting Codes, Alirad Malek, Evangelos Vasilakis, Vassilis Papaefstathiou, Pedro Trancoso and Ioannis Sourdis

FreewayNoC: a DDR NoC with Pipeline Bypassing, Ahsen Ejaz, Vassilis Papaefstathiou and Ioannis Sourdis

Ghost Loads: What is the Cost of Invisible Speculation?, Christos Sakalis, Mehdi Alipour, Alberto Ros, Stefanos Kaxiras, Alexandra Jimborean and Magnus Själander

Energy Efficiency Platform Characterization for Heterogeneous Multicore Architectures, Hergys Rexha and Sebastien Lafond

12:45 - Lunch (75)

14:00 - Keynote2 (45+15)

Hampus Linander and Carlos Galvez, Zenuity, Keynote Title: Artificial neural networks for autonomous driving

Abstract: Artificial neural networks have recently enabled robust real-time scene perception in complex environments using affordable hardware. This provides a critical step towards an accurate representation of the surroundings necessary for an autonomous vehicle.A crucial ingredient in these advances is the parallel processing paradigm and its realization in modern GPUs, both to efficiently train and deploy neural networks. More recently there has also been an increase in the availability of dedicated accelerators that enable higher performance and lower power consumption. We will review the hardware, software and data that is used to train, evaluate and deploy these algorithms in an industrial application with focus on scene perception for autonomous driving.

15:00 - Break (15)

15:15 - Session3 (60)

Load Balancing Domain Decompositions of a Lattice-Boltzmann Proxy Application, Jan Christian Meyer, Janusa Ragunathan, Jørgen Valstad and Nico Reissmann

Exploring and Analyzing the Real Impact of Modern On-package Memory on HPC Scientific Kernels, Ang Li, Weifeng Liu, Mads R. B. Kristensen, Brian Vinter, Hao Wang, Kaixi Hou, Andres Marquez and Shuaiwen Leon Song

Elastic Places: An Adaptive Resource Manager for Scalable and Portable Performance, Miquel Pericas

QoS-Driven Coordinated Management of Resources to Save Energy in Multicore Systems, Mehrzad Nejat, Miquel Pericas and Per Stenstrom

16:15 - Coffee Break (15)

16:30 - Session4 (60)

How Lock-free Data Structures Perform in Dynamic Environments: Models and Analyses, Aras Atalar, Paul Renaud-Goud and Philippas Tsigas

Enhancing Transactional Memory with Software-Prefetching, Marina Shimchenko, Ruben Titos-Gil, Ricardo Fernandez-Pascual, Manuel E. Acacio, Stefanos Kaxiras, Alberto Ros and Alexandra Jimborean

A Simple and Practical Concurrent Non-blocking Unbounded Graph with Linearizable Reachability Queries, Bapi Chatterjee, Sathya Peri, Muktikanta Sa, Nandini Singhal

Single Window Stream Aggregation using Reconfigurable Hardware, Prajith Ramakrishnan Geethakumari, Vincenzo Gulisano, Bo Joel Svensson, Pedro Trancoso and Ioannis Sourdis

17:30 - End Remarks

18:00 - Dinner

Day2 - Nov 30th

Start Time - Type (Duration in mins.)

9:15 - Keynote (45+15)

Sven Karlsson, Ericsson, Keynote Title: The Device Program at Ericsson Research -- Why It Matters to You

Abstract: Connected devices remain one of the most challenging domains within computer system research. It involves trade-offs in competing metrics such as energy efficiency, size and performance. Ericsson Research is currently building up a new research group with a charter to be at the forefront of mobile device research. The group uses a broad approach with activities in several key areas. In this talk, I will outline the effort and discuss some relevant areas we find interesting. I will also introduce some of the main challenges we engage and some of the projects tackling said challenges. What we do matters to you!

10:15 - Coffee Break (15)

10:30 - Session5 (60)

HSTREAM: A directive-based language extension for heterogeneous stream computing, Suejb Memeti and Sabri Pllana

Hybrid CPU-GPU execution support in the skeleton programming framework SkePU, Tomas Öhberg, August Ernstsson and Christoph Kessler

Viper: Communication-Layer Determinism and Scaling in Low-Latency Stream Processing, Ivan Walulya, Yiannis Nikolakopoulos, Vincenzo Gulisano, Marina Papatriantafilou and Philippas Tsigas

Designing Compact Convolutional Neural Network for Embedded Stereo Vision Systems, Mohammad Loni, Amin Majd, Abdolah Loni, Masoud Daneshtalab, Mikael Sjödin and Elena Troubitsyna

11:30 - Break (15)

11:45 - Session6 (60)

PHOENIX: Efficient Computation in Memory, Mats Rimborg, Pedro Trancoso and Gunnar Carlstedt

Decoupled Fused Cache: fusing a decoupled LLC with a DRAM cache, Evangelos Vasilakis, Vassilis Papaefstathiou, Pedro Trancoso and Ioannis Sourdis

Generating Domain Specific Manycore Architectures from Dataflow Programs, Suleyman Savas, Zain Ul-Abdin and Tomas Nordström

Adaptive Cache Warming for Faster Simulations, Gustaf Borgström, Andreas Sembrant and David Black-Schaffer

12:45 - Closing Remarks and Lunch

Call For Papers

You can find the call for papers here.

Important Dates

Full Paper Submission: October 15th, 2018

Author Notification: November 2nd, 2018

Registration Deadline: November 26th, 2018

MCC Workshop: November 29th - 30th, 2018

Program Committee

Magnus Själander, NTNU

Christoph Kessler, LiU

Troels Henriksen, U. Copenhagen

Håkan Grahn, BTH

Fuad Tabba, ARM

Johan Lilius, Åbo

Eric Berg, Ericsson

Sven Karlsson, Ericsson

Alberto Ros, U. Murcia

Magnus Jahre, NTNU

Angelos Arelakis, Zeropoint

Ivan Walulya, Chalmers

Håkan Zeffer, Intel

Helmut Neukirchen, U. Iceland

Aras Atalar, Chalmers

Weifeng Liu, NTNU

Christian Schulte, KTH

Mafijul Islam, Volvo

Gregory Vaumourin, Uppsala

David Black-Schaffer, Uppsala

Organizers

Miquel Pericàs, Chalmers

Madhavan Manivannan, Chalmers

Submission Info

There are two types of papers eligible for submission. The first type is original research work and the second type is work already published in 2017 or later. Participants submitting original work are asked to send an electronic version of the paper that does not exceed four pages using the ACM proceedings format, http://www.acm.org/publications/proceedings-template, to https://easychair.org/conferences/?conf=mcc2018. The same URL is to be used should you want to present an already published paper as described above. In that case, you need to clearly specify that the paper is already published and where the paper has been published.