Image © TUB 2017
The LPGPU2 consortium had another very successful face-to-face meeting in Berlin on October 25 and October 26, 2016.
Most of the time was spent reviewing presentations from all beneficiaries on their current status and upcoming plans and deliverables. It also provided a welcome opportunity for everyone to meet the new members of the LPGPU2 team.
Present at the meeting were Ben Juurlink Jan Lucas, Sohan Lal and Nadjib Mammeri from TUB, Simon Brand, Luke Iwanski, and Mehdi Goli from Codeplay, Georgios Keramidas, and Ignacio Aransay from Think Silicon, Prashant Sharma, Albert Saa-Garriga, and Martyn Bliss from Samsung, and Mauricio Alvarez-Mesa from Spin Digital.
The LPGPU2 project was mentioned twice as a project example par excellence at the H2020 Info Day on December 2, 2016 (https://ec.europa.eu/digital-single-market/en/news/h2020-info-day-factories-future-12-ict-5-and-ict-31-ict-innovation-manufacturing-smes-i4ms).
It was first mentioned in the presentation by Sandro d’Elia from the European Commission, who coordinates the call on “Customised and low energy computing (including Low power processor technologies)” and who is also Project Officer of the LPGPU2 project. Additionally LPGPU2 was mentioned as an exemplary project in the presentation by Prof. Koen de Bosschere from the University of Gent, who is the coordinator of the HiPEAC network (www.hipeac.net), the European network of excellence in the area of high-performance and embedded computer architecture and compilation.
Biagio Cosenza who was present for TU Berlin (the project coordinators for the LPGPU2 project) found he had an easier time introducing himself to the public due to the visibility given to the project by the previous references.
Check out https://youtu.be/Et__c9pa5TE for demonstrations of Codeplay’s VisionCpp library, which has been funded in part by LPGPU2. VisionCpp allows computer vision application developers to write all of their code in C++ and have it aggressively optimised at compile time through kernel fusion and expression templates. Codeplay’s implementation of SYCL – ComputeCpp – enables acceleration of the application on GPU devices, giving you both expressibility in the source code and real-time performance. See the project on GitHub if you would like to give it a try:
Everyone check out the LPGPU2 YouTube channel at: https://www.youtube.com/channel/UCDVYThdvn8gwNj0FdzpH1hw
We are now live and featuring a video of the power simulator developed by TUB in the LPGPU project:
Power measurement testbed designed by TUB to accurately measure the power consumption of discrete GPUs under EU funded LPGPU project. The testbed was used to validate the GPU power simulator which was also developed under LPGPU project.
More content coming soon!
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 »
Prof. Ben Juurlink, Professor for Embedded Systems Architectures, TU Berlin gave an update on the progress of LPGPU2 at the 2016 PEGPUM workshop / HiPEAC 2017. For more details please see: lpgpu2-pegpum-public
After a bustling CES in Las Vegas, we are heading to nother exciting event with the 4Y4N/MWC in Barcelona!
IoT has arrived in different forms. Smart connected devices, either worn on the wrist or as embedded and fixed installed solutions in your smart-home are here to stay! Whether for fitness, health, security, productivity or entertainment, many of those applications require a graphically rich display experience and ultra-low power consumption.
At MWC we will demonstrate the first Ultra-Low power FPGA-GPU, VR / Image processing applications, and our latest SDK API development tool platform.
LPGPU2 will be attending the 12th annual HiPEAC conference which will be held on the 23rd-25th January at the Waterfront Congress Centre in Stockholm, Sweden. Representing LPGPU2 at this years conference will be Prof. Ben Juurlink, Professor for Embedded Systems Architectures, TU Berlin. Prof Ben Juurlink will be taking part in the PEGPUM: Power-Efficient GPU and heterogeneous Multi-/Many-core Computing workshop on the 24th January – 10.00am – 17.30pm in Room 27, Waterfront Congress Centre giving an update on the LPGPU2 Project.
Read more »
Machine learning is being adopted by more and more companies for a wide range of artificial intelligence applications such as image recognition and self-driving vehicles. Machine learning involves running large data sets through a software program in order to train the software to build its intelligence. One of the biggest challenges facing developers building software that incorporates machine learning is in processing the huge amounts of data required for training their programs. By using GPUs, developers can vastly increase the amount of data that can be processed whilst also reducing the overall power consumption by running calculations in parallel.
Read more »
Spin Digital Video Technologies GmbH demonstrated at IBC 2016 an end-to-end HDR HEVC/H.265 software solution for ultra-high definition video (4K, 8K, and beyond). A demonstration was presented at the Amsterdam RAI at Hall 1 Booth 1.F11. Using the HDR-enabled video codec it is possible to encode, decode, and display high quality HDR video using a software solution running on PC platforms.
Read more »