Author Archives: Sohan Lal

Ben Juurlink presents LPGPU2 research results at ScalPerf workshop

Ben Juurlink has been invited to the ScalPerf workshop to present recent research results. The ScalPerf’17 workshop is held in Bertinoro, Italy from September 17 to September 22 and this 15th edition focuses on “Storage and Memory Issues in Computing Systems.” Ben Juurlink will present the work “E²MC: Entropy Encoding Based Memory Compression for GPUs” which has been carried out in the context of the LPGPU2 project and which has been previously presented at the IPDPS conference.

PEGPUM 2018 workshop proposal accepted

The 6th LPGPU2 Workshop on Power-Efficient GPU and Many-core Computing has been accepted by the HiPEAC 2018 organization committee as a full day workshop. The workshop will consists of a mix of (a) presentations of short papers submitted in response to the call for papers and reviewed by the organizers, (b) presentations of LPGPU2 consortium members of project results, (c) demonstrations and short tutorials by LPGPU2 consortium members on how to use the tools developed in the LPGPU2 project, and (d) one invited talk by a key person from the mobile or high-performance GPU industry. The organizers and PC members are Ben Juurlink from TU Berlin, Germany, Jan Lucas from TU Berlin, Germany, Martyn Bliss from Samsung Research UK, Georgios Keramides from Think Silicon Ltd, Greece, Henk Corporaal from TU Eindhoven, Netherlands, Paul Keir from University of the West of Scotland, UK, and Ana Lucia Varbanescu from the University of Amsterdam, The Netherlands. The HiPEAC 2018 conference will take place in Manchester, UK.

Ben Juurlink Senior Member of the ACM

Prof. Ben Juurlink, coordinator of the LPGPU2 project, has been accepted by the ACM Senior Member Committee as Senior Member of the ACM. His endorsers were Prof. Per Stenström of Chalmers University of Technology, Dr. Georgios Keramides – CTO of Think Silicon – and Dr. Alex Ramirez of Google. Successful candidates for Senior Member must have demonstrated performance that sets them apart from their peers. In general, this will be reflected in one or more of the following:

  • Technical leadership: project/engineering leadership, research leadership, education leadership, management, etc.

  • Technical contributions: publications in refereed journals or conference proceedings, textbooks, success product engineering/development, patents, standards, etc.

  • Professional contributions: service to professional societies, review committees, conference committees, standards committees, etc.

LPGPU2 at ISC High Performance 2017

Jan Lucas from TUB represented LPGPU2 at ISC High Performance conference which was held from 18th June – 22nd June at Frankfurt, Germany.  Jan demonstrated the power measurement testbed developed by TUB for LPGPU2 and gave an update of the research performed by the consortium at the HiPEAC booth on 19th of June.

Read more »

TUB at IPDPS 2017

Sohan Lal from TU Berlin attended 31st IEEE International Parallel & Distributed Processing Symposium (IPDPS) which was held from 29th May to 02 June at Orlando, Florida, USA. Sohan presented paper “E²MC: Entropy Encoding Based Memory Compression for GPUs” which is authored by Sohan Lal, Jan Lucas and Ben Juurlink.

Read more »

Getting to Know…TUB

TU Berlin is participating in the LPGPU2 project through the Embedded Systems Architectures (Architektur eingebetteter Systeme, AES) group of the School of Electrical Engineering and Computer Science. The AES groups investigates and teaches the field of computer architecture, ranging from low-power embedded systems to massively parallel high-performance systems. Its current research focuses include multi-/many-core architectures and GPUs, development and optimization of highly scalable parallel applications, and (architecture support for) parallel programming models. More information about the AES group is available at http://www.aes.tu-berlin.de/en.

TUB LPGPU2 team (from left to right): Nadjib Mammeri, Sohan Lal, Ben Juurlink, Jana Pilz, Jan Lucas, Biao Wang.

Read more »

TUB Paper Accepted at IPDPS 2017

The paper “E²MC: Entropy Encoding based Memory Compression for GPUs” by Sohan Lal, Jan  Lucas and Ben Juurlink has been accepted for publication at 31st IEEE International Parallel & Distributed Processing Symposium (IPDPS) to be held in Orlando, Florida USA from May 29 – June 2, 2017. This paper proposes an entropy encoding based memory compression technique for GPUs. The proposed compression technique addressed the key challenges of probability estimation, choosing an appropriate symbol length for encoding, and decompression with low latency. It achieves higher compression ratio and performance gain compared to state of the art.

Figure: Speedup with increased memory bandwidth

Read more »

By continuing to use the site, you agree to the use of cookies. more information

The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.

Close