News

[PhD] Virtual machine for debugging high performance computing

The RMOD team of INRIA Lille has an open position for a PhD student.

I. Context

Language virtual machines (such as the Java virtual machine) are deployed everywhere. They can be found on mobiles, on web browsers, on server applications and embedded devices. They implement many features and allow applications to be packaged once and deployed on many different platforms. They provide debugging facilities for the program they run.

Besides, in the context of high performance computing, a single program is deployed on many different machines with different hardwares (CPUs, GPUs, TPUs, FPGAs, ...). Tracking down bugs in this context is tough: all the different machines communicate asynchronously with a nondeterministic order of execution due to side-effects such as latency. Debugging this kind of problem is today close to impossible.

In the RMoD team in Inria Lille, we have skills in virtual machine implementation and application debugging. We built a virtual machine with one of the most powerful debugging engines. Our virtual machine has no support, so far, for vectorised instructions which are really important for high performance computing.

II. Problem & research directions

How to extend the language virtual machine to solve easily and efficiently bugs happening in high performance computing problems in a heterogenous cluster of machines?

The Ph.D will be split in 3 phases:

  1. How to support vector instructions in a language virtual machine with full debugging support? The student will focus on Intel SIMD instructions (AVX, AVX-512) and most likely will design a debuggable DSL on top of the virtual machine using those instructions.
  2. How to integrate vector instructions with the eco-system (frameworks, development tools, virtual machine)? The student would need to implement a library or framework (Matrix library ?) to prove that it is possible to use the DSL implemented in part 1). In terms of development tools, some enhancements are required for them to support vector instructions. Regarding the VM, multiple problems arise (How to align vector objects? How to deal with 256/512 bits of raw data on stack? How to support vector instructions efficiently in JIT compilation?)
  3. Can we efficiently debug high performance computing programs by launching part of the execution on our virtual machine while keeping decent performance ?

The student will need to adapt an existing high performance computing framework to compile to our virtual machine and build the debugging toolchain to be able to debug the part of the program run on our virtual machine. Performance will be slower than normal execution, but it has to stay practical (an hour computation can be down to a night of computation to help solving a problem, but not to months of computation).

Application

To apply, please send us :

  • a CV,
  • a copy of your Master diploma
  • a copy of your Master thesis
  • 2 (two) reference letters, with the contact details of the referents

The application materials should be sent by email to C. Béra <bera.clement@inria.fr> and S. Ducasse <stephane.ducasse@inria.fr>.

Email subject must start with : [PhD-RMoD-VM-2018].

Links

Posted by admin at 21 February 2018, 4:16 pm link

[JOB] 3 Year Starting Research Positions

Inria opended a call for starting researchers. This is a three year contract, with in total 8 research positions (for both starting and advanced combined).

Requirements:

  • Have a French doctorate or an equivalent international degree
  • From 2 to 8 years of experience in research (after the thesis)
  • No age or nationality conditions.
  • Deadline: April 3rd, 2018
  • Before writing their research program, the applicants are strongly encouraged to contact the teams or project(s)-teams leaders concerned by their applications.

More information on the official website.

MISSION

Join one of our two lines of research: Tools for Software Maintenance or New Programming Language Constructs for Software Evolution.

THE REGION

Lille is strategically located in Northern Europe with train connections to Paris (1h), London (1h30) and Brussels (30 min.) The Nord Pas de Calais region is a dynamic area in both industry and research for computer science with companies that are european leaders in their branch. Lille, the capital of northern France is an attractive metropolis of over 1 million inhabitants. Cost of living in Lille is lower than that of very big cities such as Paris and London.

THE TEAM

RMoD is an Inria research project team. It consists of 5 permant researchers (Inria CR and DR, as well as University Assistant Professors) and Engineers, postdocs and students (Master and PhD). RMoD is an international team with around 20 members from six countries.

LINKS:

APPLICATION

Please contact the head of the research team: Stephane.Ducasse@inria.fr or one of the following group members: Nicolas.Anquetil@inria.fr, Marcus.Denker@inria.fr

Posted by admin at 19 February 2018, 9:15 pm link

[JOB] 3 Year Advanced Research Position

Inria opended a call for experienced researchers. This is a three year contract, with in total 8 research positions (for both starting and advanced combined).

Requirements:

  • Have a French doctorate or an equivalent international degree
  • Be able to prove at least eight years of research and/or post-doctoral academic or industrial experience;
  • No age or nationality conditions.
  • Deadline: April 3rd, 2018
  • Before writing their research program, the applicants are strongly encouraged to contact the teams or project(s)-teams leaders concerned by their applications.

More information on the official website.

MISSION

Join one of our two lines of research: Tools for Software Maintenance or New Programming Language Constructs for Software Evolution.

THE REGION

Lille is strategically located in Northern Europe with train connections to Paris (1h), London (1h30) and Brussels (30 min.) The Nord Pas de Calais region is a dynamic area in both industry and research for computer science with companies that are european leaders in their branch. Lille, the capital of northern France is an attractive metropolis of over 1 million inhabitants. Cost of living in Lille is lower than that of very big cities such as Paris and London.

THE TEAM

RMoD is an Inria research project team. It consists of 5 permant researchers (Inria CR and DR, as well as University Assistant Professors) and Engineers, postdocs and students (Master and PhD). RMoD is an international team with around 20 members from six countries.

LINKS:

APPLICATION

Please contact the head of the research team: Stephane.Ducasse@inria.fr or one of the following group members: Nicolas.Anquetil@inria.fr, Marcus.Denker@inria.fr

Posted by admin at 17 February 2018, 12:30 pm link

Talk: Challenges of High-performance Language Virtual Machines

In the Inria Lille Series of "30 minutes of Science", Clément Bera will give a talk:

Title: Challenges of High-performance Language Virtual Machines

Abstract:

Language virtual machines are now everywhere, from server to mobile applications through web pages. The main ones are developed and maintained by large industrial groups (Oracle and the JVM, Microsoft and .Net, Google and V8, etc.). The talk introduces the different components of a virtual machine (Interpreter, JIT compiler, Memory Manager) and the key challenges to build a good one. Then, the talk details the current and incoming research going-on at RMoD around this topic.

  • When? 7 Feb 2018, 13:30
  • Where? Inria Lille, Building A
Posted by admin at 6 February 2018, 2:47 pm link

[JOB][PhD] Infrastructure and Language Kernels for IoT Systems

The RMOD team of INRIA Lille and the CAR theme of IMT Lille Douai have an open position for a PhD student on Infrastructure and language kernels for IoT Systems.

Description

Over the last years, the RMOD team of INRIA Lille and the CAR theme of IMT Lille Douai have been working together on creating tiny language core. For example, Guillermo Polito demonstrated in his PhD a fully reflective kernel that fits into 80 kb of memory and that it is possible to have hyper specialized kernels down to 11 kb. We have also worked on remote debugging (PhD of N. Papoulias) and dynamic code updates (PhD of P. Tesone) of such kernels. All of these works are prototyped in Pharo. More recently, RMOD have been working on advanced probes mechanisms (M. Denker) and a solid remote debugging infrastructure (D. Kudriashov). The goal of this PhD is to revisit the architecture around such mini-kernels for building IoT applications.

The plan is to:

  • improve the tools: edit, compile, debug, deploy and update such kernels on IoT devices. Learning how to debug remotely and dynamically update such IoT systems using the TelePharo/PharoThings environment. This task will be in cooperation with M. Denker and D. Kudriashov on remote debugging for IoT and G. Polito for the kernel edition and tooling.
  • define some language extensions to manage groups of IoT devices to program them at once. Managing hundreds or even thousands of IoT devices is a challenging task. We want to explore different solutions to help deploying and updating groups of IoT devices using some registration mechanism in a cloud server or some groups/roles based approaches for example. Note that ZweiDenker GmbH is interested in collaboration on the IoT cloud management infrastructure.
  • explore how to express the architecture of IoT applications. We would like to explore also how to express IoT architectures and what are the abstractions that should be offered to developers such as expressing event-driven architectures with declarative ECA (Event-Conditions-Actions) rules. But we will study a couple of typical IoT applications.
  • dynamically update an IoT application. An IoT application needs to adapt itself because unreachable or faulty devices or the diminution of available bandwith. We would like that the whole application can reconfigure itself in such situations as Guillaume Grondin proposes it in its PhD.
  • explore Lightweight Virtual Machines. Virtual machines in the IoT context are very powerful for incremental deployment or dynamic updates. In this task, we would like to investigate what is the minimal memory consumption that we can reach for a VM usable for IoT.

Application

To apply, please send us :

  • a CV,
  • a copy of your Master diploma
  • a copy of your Master thesis
  • 2 (two) reference letters, with the contact details of the referents
  • links to videos of demos of your experiments and/or simulations

The application materials should be sent by email to Prof. S. Ducasse stephane.ducasse@inria.fr. Email subject must start with : [PhD-RMoD-CAR-2018].

Bibliography

  • Guillermo Polito, Stephane Ducasse, N Bouraqadi, L Fabresse, M Mattone. Virtualization Support for Dynamic Core Library Update. Onward !, Oct 2015, Pittsburg, USA.
  • Guillermo Polito. Virtualization Support for Application Runtime Virtualization and Extension. Ph.D. Thesis 2015. Co-delivree par l’Universite de Lille et l’Ecole des Mines de Douai.
  • Extended results of Tornado : A Run-Fail-Grow approach for Dynamic Application Tayloring. Ecole des mines de Douai, France. 50p, July 2014
  • Nick Papoulias, Noury Bouraqadi, Luc Fabresse, Stephane Ducasse and Marcus Denker, Mercury: Properties and Design of a Remote Debugging Solution using Reflection, Journal of Object Technology, 14, 1 :1-36, 2015

Links

Posted by admin at 25 January 2018, 11:54 am link
<< 1 2 3 4 5 6 7 8 9 10 >>