An official website of the United States government
Scalable methods for optical transmission performance prediction using machine learning (ML) are studied in metro reconfigurable optical add-drop multiplexer (ROADM) networks. A cascaded learning framework is introduced to encompass the use of cascaded component models for end-to-end (E2E) optical path prediction augmented with different combinations of E2E performance data and models. Additional E2E optical path data and models are used to reduce the prediction error accumulation in the cascade. Off-line training (pre-trained prior to deployment) and transfer learning are used for component-level erbium-doped fiber amplifier (EDFA) gain models to ensure scalability. Considering channel power prediction, we show that the data collection process of the pre-trained EDFA model can be reduced to only 5% of the original training set using transfer learning. We evaluate the proposed method under three different topologies with field deployed fibers and achieve a mean absolute error of 0.16 dB with a single (one-shot) E2E measurement on the deployed 6-span system with 12 EDFAs.
more »
« less
Empirical Evaluation of ML Models for Per-Job Power Prediction
Sustainability has become a critical focus area across the technology industry, most notably in cloud data centers. In such shared-use computing environments, there is a need to account for the power consumption of individual users. Prior work on power prediction of individual user jobs in shared environments has often focused on workloads that stress a single resource, such as CPU or DRAM. These works typically employ a specific machine learning (ML) model to train and test on the target workload for high accuracy. However, modern workloads in data centers can stress multiple resources simultaneously, and cannot be assumed to always be available for training. This paper empirically evaluates the performance of various ML models under different model settings and training data assumptions for the per-job power prediction problem using a range of workloads. Our evaluation results provide key insights into the efficacy of different ML models. For example, we find that linear ML models suffer from poor prediction accuracy (as much as 25% prediction error), especially for unseen workloads. Conversely, non-linear models, specifically XGBoost and xRandom Forest, provide reasonable accuracy (7–9% error). We also find that data-normalization and the power-prediction model formulation affect the accuracy of individual ML models in different ways.
more »
« less
- PAR ID:
- 10525131
- Publisher / Repository:
- ACM
- Date Published:
- ISBN:
- 9798400704451
- Page Range / eLocation ID:
- 181 to 188
- Subject(s) / Keyword(s):
- Hardware -> Power and energy Power estimation and optimization Enterprise level and data centers power issues Sustainability per-job power prediction ML models co-executed workloads
- Format(s):
- Medium: X
- Location:
- London United Kingdom
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract One compelling vision of the future of materials discovery and design involves the use of machine learning (ML) models to predict materials properties and then rapidly find materials tailored for specific applications. However, realizing this vision requires both providing detailed uncertainty quantification (model prediction errors and domain of applicability) and making models readily usable. At present, it is common practice in the community to assess ML model performance only in terms of prediction accuracy (e.g. mean absolute error), while neglecting detailed uncertainty quantification and robust model accessibility and usability. Here, we demonstrate a practical method for realizing both uncertainty and accessibility features with a large set of models. We develop random forest ML models for 33 materials properties spanning an array of data sources (computational and experimental) and property types (electrical, mechanical, thermodynamic, etc). All models have calibrated ensemble error bars to quantify prediction uncertainty and domain of applicability guidance enabled by kernel-density-estimate-based feature distance measures. All data and models are publicly hosted on the Garden-AI infrastructure, which provides an easy-to-use, persistent interface for model dissemination that permits models to be invoked with only a few lines of Python code. We demonstrate the power of this approach by using our models to conduct a fully ML-based materials discovery exercise to search for new stable, highly active perovskite oxide catalyst materials.more » « less
-
Modern Internet of Things (IoT) applications, from contextual sensing to voice assistants, rely on ML-based training and serving systems using pre-trained models to render predictions. However, real-world IoT environments are diverse, with rich IoT sensors and need ML models to be personalized for each setting using relatively less training data. Most existing general-purpose ML systems are optimized for specific and dedicated hardware resources and do not adapt to changing resources and different IoT application requirements. To address this gap, we propose MLIoT, an end-to-end Machine Learning System tailored towards supporting the entire lifecycle of IoT applications. MLIoT adapts to different IoT data sources, IoT tasks, and compute resources by automatically training, optimizing, and serving models based on expressive applicationspecific policies. MLIoT also adapts to changes in IoT environments or compute resources by enabling re-training, and updating models served on the fly while maintaining accuracy and performance. Our evaluation across a set of benchmarks show that MLIoT can handle multiple IoT tasks, each with individual requirements, in a scalable manner while maintaining high accuracy and performance. We compare MLIoT with two state-of-the-art hand-tuned systems and a commercial ML system showing that MLIoT improves accuracy from 50% - 75% while reducing or maintaining latency.more » « less
-
null (Ed.)GPUs are a key enabler of the revolution in machine learning and high-performance computing, functioning as de facto co-processors to accelerate large-scale computation. As the programming stack and tool support have matured, GPUs have also become accessible to programmers, who may lack detailed knowledge of the underlying architecture and fail to fully leverage the GPU’s computation power. GEVO (Gpu optimization using EVOlutionary computation) is a tool for automatically discovering optimization opportunities and tuning the performance of GPU kernels in the LLVM representation. GEVO uses population-based search to find edits to GPU code compiled to LLVM-IR and improves performance on desired criteria while retaining required functionality. We demonstrate that GEVO improves the execution time of general-purpose GPU programs and machine learning (ML) models on NVIDIA Tesla P100. For the Rodinia benchmarks, GEVO improves GPU kernel runtime performance by an average of 49.48% and by as much as 412% over the fully compiler-optimized baseline. If kernel output accuracy is relaxed to tolerate up to 1% error, GEVO can find kernel variants that outperform the baseline by an average of 51.08%. For the ML workloads, GEVO achieves kernel performance improvement for SVM on the MNIST handwriting recognition (3.24×) and the a9a income prediction (2.93×) datasets with no loss of model accuracy. GEVO achieves 1.79× kernel performance improvement on image classification using ResNet18/CIFAR-10, with less than 1% model accuracy reduction.more » « less
-
CNNs are increasingly deployed across different hardware, dynamic environments, and low-power embedded devices. This has led to the design and training of CNN architectures with the goal of maximizing accuracy subject to such variable deployment constraints. As the number of deployment scenarios grows, there is a need to find scalable solutions to design and train specialized CNNs. Once-for-all training has emerged as a scalable approach that jointly co-trains many models (subnets) at once with a constant training cost and finds specialized CNNs later. The scalability is achieved by training the full model and simultaneously reducing it to smaller subnets that share model weights (weight-shared shrinking). However, existing once-for-all training approaches incur huge training costs reaching 1200 GPU hours. We argue this is because they either start the process of shrinking the full model too early or too late. Hence, we propose Delayed Epsilon-Shrinking (DepS) that starts the process of shrinking the full model when it is partially trained, which leads to training cost improvement and better in-place knowledge distillation to smaller models. The proposed approach also consists of novel heuristics that dynamically adjust subnet learning rates incrementally, leading to improved weight-shared knowledge distillation from larger to smaller subnets as well. As a result, DepS outperforms state-of-the-art once-for-all training techniques across different datasets including CIFAR10/100, ImageNet-100, and ImageNet-1k on accuracy and cost. It achieves higher ImageNet-1k top1 accuracy or the same accuracy with 1.3x reduction in FLOPs and 2.5x drop in training cost (GPU*hrs).more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1145/3629527.3651418
