Scaling to larger systems, with current levels of reliability, requires cost-effective methods to mitigate hardware failures. One of the main causes of hardware failure is an uncorrected error in memory, which terminates the current job and wastes all computation since the last checkpoint. This paper presents the first adaptive method for triggering uncorrected error mitigation. It uses a prediction approach that considers the likelihood of an uncorrected error and its current potential cost. The method is based on reinforcement learning, and the only user-defined parameters are the mitigation cost and whether the job can be restarted from a mitigation point. We evaluate our method using classical machine learning metrics together with a cost-benefit analysis, which compares the cost of mitigation actions with the benefits from mitigating some of the errors. On two years of production logs from the MareNostrum supercomputer, our method reduces lost compute time by 54% compared with no mitigation and is just 6% below the optimal Oracle method. All source code is open source.

It is widely accepted that cosmic rays are a plausible cause of DRAM errors in high-performance computing (HPC) systems, and various studies suggest that they could explain some aspects of the observed DRAM error behavior. However, this phenomenon is insufficiently studied in production environments. We analyze the correlations between cosmic rays and DRAM errors on two HPC clusters: a production supercomputer with server-class DDR3-1600 and a prototype with LPDDR3-1600 and no hardware error correction. Our error logs cover 2000 billion MB-hours for the MareNostrum 3 supercomputer and 135 million MB-hours for the Mont-Blanc prototype. Our analysis combines quantitative analysis, formal statistical methods and machine learning. We detect no indications that cosmic rays have any influence on the DRAM errors. To understand whether the findings are specific to systems under study, located at 100 meters above the sea level, the analysis should be repeated on other HPC clusters, especially the ones located on higher altitudes. Also, analysis can (and should) be applied to revisit and extend numerous previous studies which use cosmic rays as a hypothetical explanation for some aspects of the observed DRAM error behaviors.

Clusters of emerging technologies are appearing with more and more frequency in HPC. After years of skepticism, data-centers are adopting them as production systems thanks to several geopolitical and technological factors. The most honorable example is the Fugaku supercomputer, powered by the latest Fujitsu A64FX CPU. Which is the behavior of mature HPC codes on such emerging technology clusters? Which performance will obtain scientists when running their HPC applications “as is” on these clusters? This paper presents the evaluation of CTE-Arm, a Fugaku-like system, including both fine-tuned micro-benchmarks and five scientific applications run without prior fine-tuning: Alya, NEMO, Gromacs, OpenIFS, and WRF. Results show that while micro-architectural benchmarks show performance as expected, the performance obtained running HPC applications not tuned for a specific architecture are between $2\times $ and $4\times $ slower compared with a standard Intel-based HPC system. Therefore further effort is needed to improve tools (e.g., compilers) and system software (e.g., MPI libraries) to ease applications deployment and improve their performance.

This paper presents and evaluates a method to predict DRAM uncorrected errors, a leading cause of hardware failures in large-scale HPC clusters. The method uses a random forest classifier, which was trained and evaluated using error logs from two years of production of the MareNostrum 3 supercomputer. By enabling the system to take measures to mitigate node failures, our method reduces lost compute time by up to 57%, a net saving of 21,000 node-hours per year. We release all source code as open source. We also discuss and clarify aspects of methodology that are essential for a DRAM prediction method to be useful in practice. We explain why standard evaluation metrics, such as precision and recall, are insufficient, and base the evaluation on a cost-benefit analysis. This methodology can help ensure that any DRAM error predictor is clear from training bias and has a clear cost-benefit calculation.

In this paper we propose two decentralized medium access control (MAC) schemes to coordinate the access of multiple transmitters adopting a Multi-Packet Reception (MPR) physical (PHY) layer. The proposed MAC schemes operate over a MPR-based PHY-layer and are divided in two stages. In the first stage the nodes indicate their willingness to transmit, while in the second one they jointly transmit. The MAC schemes differ by considering a Single-Packet Reception (SPR) or a MPR PHY-layer in the first stage. We analyze the throughput achieved by each MAC in a theoretical way. Several results obtained through simulation are also presented, which validate the numerical results obtained with the theoretical model. The results demonstrate the effectiveness of the proposed MAC schemes, and allow us to draw valuable conclusions about the design of distributed MPR systems.

In recent schemes proposed to the physical layer of wireless systems the receiver node has the capability of simultaneously receiving multiple transmissions, which is commonly referred to as a Multi-Packet Reception (MPR) scheme. Assuming that the transmitters are randomly located in the far-field region and a generic signal-to-interference-plus-noise ratio (SINR) threshold-based criterion is used as the packet capture model, we derive the probability of successful reception of a packet considering path loss, shadowing and fading effects. Our approach relies on the characterization of the aggregate interference caused by the transmitters, which is used in the capture model in a simplified way. The probability of successful reception of a packet is then used to approximate the average number of packets simultaneously received. The proposed analysis is evaluated for different scenarios and compared with several results obtained through simulation. The simplicity of the model, as well as its accuracy, makes it a useful tool to assist the design of future medium access control mechanisms for MPR wireless systems.

This work addresses the target tracking problem that makes use of combined measurements, namely received signal strength (RSS) and angle of arrival (AoA). By linearizing the measurement models and incorporating the prior knowledge obtained from target state transition model, we show that the application of the Kalman filter (KF) to the considered tracking problem is straightforward. Then, an extension of the linearization approach to the case where the target transmit power is not known is introduced and applied to the measurement model to obtain an estimate of the transmit power. By taking advantage of this estimated value, we show that the proposed KF algorithm can easily be generalized to the case of unknown transmit power. Our simulation results confirm the efficacy of the proposed algorithms in comparison with the existing one, as well as the robustness of the proposed approach to not knowing the transmit power. Finally, the supremacy of using the Bayesian approach in comparison with the classical one which disregards the prior knowledge information is also validated through computer simulations.