No abstract

Abstract-High-performance computing (HPC) is recognized as one of the pillars for further advance of science, industry, medicine, and education. Current HPC systems are being developed to overcome emerging challenges in order to reach Exascale level of performance, which is expected by the year 2020. The much larger embedded and mobile market allows for rapid development of IP blocks, and provides more flexibility in designing an application-specific SoC, un turn giving possibility in balancing performance, energy-efficiency and cost. In the Mont-Blanc project, we advocate for HPC systems be built from such commodity IP blocks, currently used in embedded and mobile SoCs.As a first demonstrator of such approach, we present the MontBlanc prototype; the first HPC system built with commodity SoCs, memories, and NICs from the embedded and mobile domain, and offthe-shelf HPC networking, storage, cooling and integration solutions. We present the system's architecture, and evaluation including both performance and energy efficiency. Further, we compare the system's abilities against a production level supercomputer. At the end, we discuss parallel scalability, and estimate the maximum scalability point of this approach across a set of applications.

International audienceThe evolution of manycore sytems, forecasted to feature hundreds of cores by the end of the decade calls for efficient solutions for design space exploration and debugging. Among the relevant existing solutions the well-known gem5 simu-lator provides a rich architecture description framework. However , these features come at the price of prohibitive simulation time that limits the scope of possible explorations to configurations made of tens of cores. To address this limitation, this paper proposes a novel trace-driven simulation approach for efficient exploration of manycore architectures

Scalability and programmability are important issues in large homogeneous MPSoCs. Such architectures often rely on explicit message-passing among processors, each of which possessing a local private memory. This paper presents a low-overhead hardware/software distributed shared memory approach that makes such architectures multithreading-capable. The proposed solution is implemented into an open-source message-passing MPSoC through developing a POSIX-like thread API, which shows excellent scalability using application kernels used for benchmarking in shared-memory systems. This approach efficiently draws strengths from the on-chip distributed private memory that opens the way to exposing the multithreading programmability/capabilities of that component as a generalpurpose accelerator.

Dense linear algebra libraries, such as BLAS and LAPACK, provide a relevant collection of numerical tools for many scientific and engineering applications. While there exist high performance implementations of the BLAS (and LAPACK) functionality for many current multi-threaded architectures, the adaption of these libraries for asymmetric multicore processors (AMPs) is still pending. In this paper we address this challenge by developing an asymmetry-aware implementation of the BLAS, based on the BLIS framework, and tailored for AMPs equipped with two types of cores: fast/power hungry versus slow/energy efficient. For this purpose, we integrate coarsegrain and fine-grain parallelization strategies into the library routines which, respectively, dynamically distribute the workload between the two core types and statically repartition this work among the cores of the same type.Our results on an ARM R big.LITTLE TM processor embedded in the Exynos 5422 SoC, using the asymmetry-aware version of the BLAS and a plain migration of the legacy version of LAPACK, experimentally assess the benefits, limitations, and potential of this approach.