E2Data project

As part of the upcoming H2020 E2Data project, we are interested in exploiting heterogeneous hardware resources on Big Data stacks. Our main interest is in sharing resources for researching and developing novel tool and components for dynamic acceleration of data analytics.

We expect to contribute knowledge and insights regarding dynamic compilation and reconfiguration of data analytics code executed on top of managed runtime environments. E2Data proposes an end-to-end solution for Big Data deployments that fully exploits and advances the state-of-the-art in infrastructure services by delivering significant performance increase with much less cloud resources. E2Data provides a new Big Data software paradigm of achieving the maximum resource utilisation for heterogeneous cloud deployments without affecting current Big Data programming norms (i.e. no code changes in the original source). This will be achieved by co-designing the Big Data stack and applications with novel compiler and run-time components for dynamic and intelligent acceleration based on each individual scenario.

OPERA project (non-active)

Share results from the OPERA project which main achievements are on Computing continuum, Low Power Scalable Form Factor Data Center, workload decomposition, Ultra Low Power Devices and FPGA acceleration for CNN.

Contribution will be on participating in the alliance to share ideas for technologies improvement on heterogeneous architecture, worklopad decompostion and resources allocation on Heterogeneous Low Power infrastructure, Computing Continuum architecture, Offloading.

ISMB Research Center

Istituto Superiore Mario Boella is a Research and Innovation Center in the Information and Communication Technologies (ICT) domain.  On the alliance main fields of contribution will on applied research on resources allocation, Low Power computing and communications, Heterogeneous Low Power architectures, CNN on FPGA accelerators, Cloud computing, sharing projects results and skills for next generation on heterogeneous hardware and software architectures.

Research center shares projects results and skills, competences and future research work in H2020.

Teseo S.p.A

EIFFAGE globally is heavily investing in development of new technologies that can improve industry operations and people’s life. Teseo, as the Italian subsidiary of EIFFAGE Group, has several direct interests in the Alliance, including contribution to the growth of the heterogeneity market leveraging on the experience made during the decades and in a H2020 project on Ultra low power and low power heterogeneous architectures called OPERA.

The contributions expected to the Alliance are to give expertise on the following aspects: hardware/firmware/software design or co-design, hardware/firmware/software integrations and RF/EMC Testing

Think Silicon S.A.

Think Silicon S.A. is specialized in designing and developing ultra low-power IP semiconductor cores in the field of computer graphics. Our product portfolio includes multicore Graphics Processors and Display controllers. We are currently designing a heterogeneous GPU targeting mainly the IoT and Wearable market. Apart from the GPU HW, we are developing the GPU libraries and drivers as well as profiling tools for multicore GPUs.

In the field of GPUs, the programming APIs are mostly defined by the Khronos group. We are working in developing the low level implementation (driver) of OpenGL and Vulkan API assuming a heterogeneous multicore GPU. Our role in the Alliance is mainly from the end-user perspective. We would like to get a better understating of the proposed technology, apply this technology in our GPU HW and profiling tools and providing feedback to the alliance as GPU HW vendor. Our GPU profiling tools are being developed as part of an EU-funded project called LPGPU2 (http://lpgpu.org/), from which we will establish an active communication link between this alliance and the LPGPU2 project.

IUMA-ULPGC Institute for Applied Microelectronics, University of Las Palmas de Gran Canaria

IUMA-ULPGC is a research center linked to Hipeac network, and a established and certified NVIDIA CUDA Research Center in Europe. It is also a reference center for ADA 2012 Run Time Environment supporting Ada Core NY and Paris (IUMA also supported RTE of ADA 95 and Ada 2005). IUMA works in GPU, Hardware Accelerators, and MPSoC, including many-cores.

IUMA can contribute tools and technologies in the areas of parallel languages, parallel programming, GPU programming, and design space exploration for HW-SW accelerators. IUMA has expertise in ASIC and FPGA and GPU-FPGA implementations of HW-SW accelerators following optimised design flows from High Level and System Level ESL in various target IC technologies. IUMA can also share expertise in the applied field of Multiscalar and Hiperspectral Image Processing, Machine Learning, Classification, and Data Science and Analytics problems. Other application fields include High Speed Network processors for Deep Packet Inspection and Cibersecurity. IUMA-ULPGC is working on heterogeneous embedded systems in various national and international projects (including ESA projects) and they are very interested to widen their approaches and cooperation. Together with space and air drone embedded applications IUMA also contributes work for submarine embededd developments in aquaculture problems and marine glider requirements for onboard navigation electronics.

Certios B.V.

Certios is interested in sharing ideas, tap into a potentially very interesting international network, and start up new research projects with other members.

Certios is contributing with technical know-how, specifically bringing in our energy efficiency models, knowledge, and experience.

WINGS ICT Solutions

WINGS ICT Solutions is interested in the Alliance, as we do have the experience and expertise to actively contribute to different phases of applications for H-HW&SW.

WINGS ICT Solutions aims at contributing to the respective discussions and activities for the development of future technologies for applications using heterogeneous hardware and software.

Barcelona Supercomputing Center

The Barcelona Supercomputing Center (BSC) is the Spanish national supercomputing facility which mission is to do research, develop and manage information technologies to facilitate scientific progress. The Computer Sciences Department at BSC aims at influencing in the way of computers are build and program as well as adapting current hardware and software technologies to make efficient use of supercomputing infrastructures.

BSC will be an active participant of the Heterogeneity Alliance collaborating in the alliance working groups, and it will contribute to the catalog with the integrated version of COMPSs and OmpSs programming models and runtimes. As consequence of the collaboration with the Alliance members, we expect to identify new use cases for extending and developing innovative features for the BSC programming models and increasing the BSC programming model adoption. Due to the BSC experience in programming models, it can potentially lead the area of programming models for heterogeneous parallel architectures proposing activities to find synergies with other programming models, runtimes, and other middleware components. These activities can include the integration of programming model runtimes and middleware components to leverage complementary functionalities or to support new platforms, or the identification of common features to submit joint proposals to standardization bodies such as OpenMP.

The University of Manchester

University of Manchester is interested in joining the Heterogeneity Alliance to disseminate their results in the domain across diverse projects and also to foster collaboration with other projects and organisations working on the same field.

Their contribution will be mainly focused on the heterogeneous execution of JVMs.