An official website of the United States government
—Exascale computing enables unprecedented, detailed and coupled scientific simulations which generate data on the order of tens of petabytes. Due to large data volumes, lossy compressors become indispensable as they enable better compression ratios and runtime performance than lossless compressors. Moreover, as (high-performance computing) HPC systems grow larger, they draw power on the scale of tens of megawatts. Data motion is expensive in time and energy. Therefore, optimizing compressor and data I/O power usage is an important step in reducing energy consumption to meet sustainable computing goals and stay within limited power budgets. In this paper, we explore efficient power consumption gains for the SZ and ZFP lossy compressors and data writing on a cloud HPC system while varying the CPU frequency, scientific data sets, and system architecture. Using this power consumption data, we construct a power model for lossy compression and present a tuning methodology that reduces energy overhead of lossy compressors and data writing on HPC systems by 14.3% on average. We apply our model and find 6.5 kJs, or 13%, of savings on average for 512GB I/O. Therefore, utilizing our model results in more energy efficient lossy data compression and I/O.
more »
« less
Market Mechanism-Based User-in-the-Loop Scalable Power Oversubscription for HPC Systems
Significant power consumption is one of the major challenges for current and future high-performance computing (HPC) systems. All the while, HPC systems generally remain power underutilized, making them a great candidate for applying power oversubscription to reclaim unused capacity. However, an oversubscribed HPC system may occasionally get overloaded. In this paper, we propose MPR (Market-based Power Reduction), a scalable market-based approach where users actively participate in reducing the HPC system’s power consumption to mitigate overloads. In MPR, HPC users bid to supply, in exchange for incentives, the resource reduction required for handling the overloads. Using several real-world trace-based simulations, we extensively evaluate MPR and show that, by participating in MPR, users always receive more rewards than the cost of performance loss. At the same time, the HPC manager enjoys orders of magnitude more resource gain than her incentive payoff to the users. We also demonstrate the real-world effectiveness of MPR on a prototype system.
more »
« less
- PAR ID:
- 10410652
- Date Published:
- Journal Name:
- 2023 IEEE International Symposium on High-Performance Computer Architecture (HPCA)
- Page Range / eLocation ID:
- 485 to 498
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Federated learning is a novel paradigm allowing the training of a global machine-learning model on distributed devices. It shares model parameters instead of private raw data during the entire model training process. While federated learning enables machine learning processes to take place collaboratively on Internet of Things (IoT) devices, compared to data centers, IoT devices with limited resource budgets typically have less security protection and are more vulnerable to potential thermal stress. Current research on the evaluation of federated learning is mainly based on the simulation of multi-clients/processes on a single machine/device. However, there is a gap in understanding the performance of federated learning under thermal stress in real-world distributed low-power heterogeneous IoT devices. Our previous work was among the first to evaluate the performance of federated learning under thermal stress on real-world IoT-based distributed systems. In this paper, we extended our work to a larger scale of heterogeneous real-world IoT-based distributed systems to further evaluate the performance of federated learning under thermal stress. To the best of our knowledge, the presented work is among the first to evaluate the performance of federated learning under thermal stress on real-world heterogeneous IoT-based systems. We conducted comprehensive experiments using the MNIST dataset and various performance metrics, including training time, CPU and GPU utilization rate, temperature, and power consumption. We varied the proportion of clients under thermal stress in each group of experiments and systematically quantified the effectiveness and real-world impact of thermal stress on the low-end heterogeneous IoT-based federated learning system. We added 67% more training epochs and 50% more clients compared with our previous work. The experimental results demonstrate that thermal stress is still effective on IoT-based federated learning systems as the entire global model and device performance degrade when even a small ratio of IoT devices are being impacted. Experimental results have also shown that the more influenced client under thermal stress within the federated learning system (FLS) tends to have a more major impact on the performance of FLS under thermal stress.more » « less
-
High performance computing (HPC) system runs compute-intensive parallel applications requiring large number of nodes. An HPC system consists of heterogeneous computer architecture nodes, including CPUs, GPUs, field programmable gate arrays (FPGAs), etc. Power capping is a method to improve parallel application performance subject to variable power constraints. In this paper, we propose a parallel application power and performance prediction simulator. We present prediction model to predict application power and performance for unknown power-capping values considering heterogeneous computing architecture. We develop a job scheduling simulator based on parallel discrete-event simulation engine. The simulator includes a power and performance prediction model, as well as a resource allocation model. Based on real-life measurements and trace data, we show the applicability of our proposed prediction model and simulator.more » « less
-
Accurate prediction of parallel application performance in HPC systems is essential for efficient resource allocation and system design. Classical performance models estimate of speedup based on theoretical assumptions, but their applicability is limited by parameter estimation, data acquisition, and real-world system issues such as latency and network congestion. This paper describes performance prediction using classical performance models boosted by a trainable machine learning framework. Domain-informed machine-learning models estimate the overhead of an application for a given problem size and resource configuration as a coefficient of the estimated speedup provided by performance laws. We evaluate this approach on two HPC mini-applications and two full applications with varying patterns of computation and communication and also evaluate the prediction accuracy on runs with varying processors-per-node configurations. Our results show that this method significantly improves the accuracy of performance predictions over standard analytical models and black-box regressors, while remaining robust even with limited training data.more » « less
-
null (Ed.)With the recent advances in both machine learning and embedded systems research, the demand to deploy computational models for real-time execution on edge devices has increased substantially. Without deploying computational models on edge devices, the frequent transmission of sensor data to the cloud results in rapid battery draining due to the energy consumption of wireless data transmission. This rapid power dissipation leads to a considerable reduction in the battery lifetime of the system, therefore jeopardizing the real-world utility of smart devices. It is well-established that for difficult machine learning tasks, models with higher performance often require more computation power and thus are not power-efficient choices for deployment on edge devices. However, the trade-offs between performance and power consumption are not well studied. While numerous methods (e.g., model compression) have been developed to obtain an optimal model, these methods focus on improving the efficiency of a single model. In an entirely new direction, we introduce an effective method to find a combination of multiple models that are optimal in terms of power-efficiency and performance by solving an optimization problem in which both performance and power consumption are taken into account. Experimental results demonstrate that on the ImageNet dataset, we can achieve a 20% energy reduction with only 0.3% accuracy drop compared to Squeeze-and-Excitation Networks. Compared to a pruned convolutional neural network for human activity recognition, while consuming 1.7% less energy, our proposed policy achieves 1.3% higher accuracy.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/HPCA56546.2023.10071006
