skip to main content


Title: RLScheduler: An Automated HPC Batch Job Scheduler Using Reinforcement Learning
Today’s high-performance computing (HPC) platforms are still dominated by batch jobs. Accordingly, effective batch job scheduling is crucial to obtain high system efficiency. Existing HPC batch job schedulers typically leverage heuristic priority functions to prioritize and schedule jobs. But, once configured and deployed by the experts, such priority function scan hardly adapt to the changes of job loads, optimization goals, or system settings, potentially leading to degraded system efficiency when changes occur. To address this fundamental issue, we present RLScheduler, an automated HPC batch job scheduler built on reinforcement learning. RLScheduler relies on minimal manual interventions or expert knowledge, but can learn high-quality scheduling policies via its own continuous ‘trial and error’. We introduce a new kernel-based neural network structure and trajectory filtering mechanism in RLScheduler to improve and stabilize the learning process. Through extensive evaluations,we confirm that RLScheduler can learn high-quality scheduling policies towards various workloads and various optimization goals with relatively low computation cost. Moreover, we show that the learned models perform stably even when applied to unseen workloads, making them practical for production use.  more » « less
Award ID(s):
1817089
NSF-PAR ID:
10196073
Author(s) / Creator(s):
; ; ; ;
Date Published:
Journal Name:
SC'20: The International Conference for High Performance Computing, Networking, Storage, and Analysis 2020
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Compute heterogeneity is increasingly gaining prominence in modern datacenters due to the addition of accelerators like GPUs and FPGAs. We observe that datacenter schedulers are agnostic of these emerging accelerators, especially their resource utilization footprints, and thus, not well equipped to dynamically provision them based on the application needs. We observe that the state-of-the-art datacenter schedulers fail to provide fine-grained resource guarantees for latency-sensitive tasks that are GPU-bound. Specifically for GPUs, this results in resource fragmentation and interference leading to poor utilization of allocated GPU resources. Furthermore, GPUs exhibit highly linear energy efficiency with respect to utilization and hence proactive management of these resources is essential to keep the operational costs low while ensuring the end-to-end Quality of Service (QoS) in case of user-facing queries.Towards addressing the GPU orchestration problem, we build Knots, a GPU-aware resource orchestration layer and integrate it with the Kubernetes container orchestrator to build Kube- Knots. Kube-Knots can dynamically harvest spare compute cycles through dynamic container orchestration enabling co-location of latency-critical and batch workloads together while improving the overall resource utilization. We design and evaluate two GPU-based scheduling techniques to schedule datacenter-scale workloads through Kube-Knots on a ten node GPU cluster. Our proposed Correlation Based Prediction (CBP) and Peak Prediction (PP) schemes together improves both average and 99 th percentile cluster-wide GPU utilization by up to 80% in case of HPC workloads. In addition, CBP+PP improves the average job completion times (JCT) of deep learning workloads by up to 36% when compared to state-of-the-art schedulers. This leads to 33% cluster-wide energy savings on an average for three different workloads compared to state-of-the-art GPU-agnostic schedulers. Further, the proposed PP scheduler guarantees the end-to-end QoS for latency-critical queries by reducing QoS violations by up to 53% when compared to state-of-the-art GPU schedulers. 
    more » « less
  2. Due to a growing interest in deep learning applications [5], compute-intensive and long-running (hours to days) training jobs have become a significant component of datacenter workloads. A large fraction of these jobs is often exploratory, with the goal of determining the best model structure (e.g., the number of layers and channels in a convolutional neural network), hyperparameters (e.g., the learning rate), and data augmentation strategies for the target application. Notably, training jobs are often terminated early if their learning metrics (e.g., training and validation accuracy) are not converging, with only a few completing successfully. For this motivating application, we consider the problem of scheduling a set of jobs that can be terminated at predetermined checkpoints with known probabilities estimated from historical data. We prove that, in order to minimize the time to complete the first K successful jobs on a single server, optimal scheduling does not require preemption (even when preemption overhead is negligible) and provide an optimal policy; advantages of this policy are quantified through simulation. Related Work. While job scheduling has been investigated extensively in many scenarios (see [6] and [2] for a survey of recent result), most policies require that the cost of waiting times of each job be known at scheduling time; in contrast, in our setting the scheduler does not know which job will be the K-th successful job, and sojourn times of subsequent jobs do not contribute to the target metric. For example, [4, 3] minimize makespan (i.e., the time to complete all jobs) for known execution times and waiting time costs; similarly, Gittins index [1] and SR rank [7] minimize expected sojourn time of all jobs, i.e., both successfully completed jobs and jobs terminated early. Unfortunately, scheduling policies not distinguishing between these two types of jobs may favor jobs where the next stage is short and leads to early termination with high probability, which is an undesirable outcome in our applications of interest. 
    more » « less
  3. High performance computing (HPC) is undergoing significant changes. The emerging HPC applications comprise both compute- and data-intensive applications. To meet the intense I/O demand from emerging data-intensive applications, burst buffers are deployed in production systems. Existing HPC schedulers are mainly CPU-centric. The extreme heterogeneity of hardware devices, combined with workload changes, forces the schedulers to consider multiple resources (e.g., burst buffers) beyond CPUs, in decision making. In this study, we present a multi-resource scheduling scheme named BBSched that schedules user jobs based on not only their CPU requirements, but also other schedulable resources such as burst buffer. BBSched formulates the scheduling problem into a multi-objective optimization (MOO) problem and rapidly solves the problem using a multi-objective genetic algorithm. The multiple solutions generated by BBSched enables system managers to explore potential tradeoffs among various resources, and therefore obtains better utilization of all the resources. The trace-driven simulations with real system workloads demonstrate that BBSched improves scheduling performance by up to 41% compared to existing methods, indicating that explicitly optimizing multiple resources beyond CPUs is essential for HPC scheduling. 
    more » « less
  4. Obeid, I. ; Selesnik, I. ; Picone, J. (Ed.)
    The Neuronix high-performance computing cluster allows us to conduct extensive machine learning experiments on big data [1]. This heterogeneous cluster uses innovative scheduling technology, Slurm [2], that manages a network of CPUs and graphics processing units (GPUs). The GPU farm consists of a variety of processors ranging from low-end consumer grade devices such as the Nvidia GTX 970 to higher-end devices such as the GeForce RTX 2080. These GPUs are essential to our research since they allow extremely compute-intensive deep learning tasks to be executed on massive data resources such as the TUH EEG Corpus [2]. We use TensorFlow [3] as the core machine learning library for our deep learning systems, and routinely employ multiple GPUs to accelerate the training process. Reproducible results are essential to machine learning research. Reproducibility in this context means the ability to replicate an existing experiment – performance metrics such as error rates should be identical and floating-point calculations should match closely. Three examples of ways we typically expect an experiment to be replicable are: (1) The same job run on the same processor should produce the same results each time it is run. (2) A job run on a CPU and GPU should produce identical results. (3) A job should produce comparable results if the data is presented in a different order. System optimization requires an ability to directly compare error rates for algorithms evaluated under comparable operating conditions. However, it is a difficult task to exactly reproduce the results for large, complex deep learning systems that often require more than a trillion calculations per experiment [5]. This is a fairly well-known issue and one we will explore in this poster. Researchers must be able to replicate results on a specific data set to establish the integrity of an implementation. They can then use that implementation as a baseline for comparison purposes. A lack of reproducibility makes it very difficult to debug algorithms and validate changes to the system. Equally important, since many results in deep learning research are dependent on the order in which the system is exposed to the data, the specific processors used, and even the order in which those processors are accessed, it becomes a challenging problem to compare two algorithms since each system must be individually optimized for a specific data set or processor. This is extremely time-consuming for algorithm research in which a single run often taxes a computing environment to its limits. Well-known techniques such as cross-validation [5,6] can be used to mitigate these effects, but this is also computationally expensive. These issues are further compounded by the fact that most deep learning algorithms are susceptible to the way computational noise propagates through the system. GPUs are particularly notorious for this because, in a clustered environment, it becomes more difficult to control which processors are used at various points in time. Another equally frustrating issue is that upgrades to the deep learning package, such as the transition from TensorFlow v1.9 to v1.13, can also result in large fluctuations in error rates when re-running the same experiment. Since TensorFlow is constantly updating functions to support GPU use, maintaining an historical archive of experimental results that can be used to calibrate algorithm research is quite a challenge. This makes it very difficult to optimize the system or select the best configurations. The overall impact of all of these issues described above is significant as error rates can fluctuate by as much as 25% due to these types of computational issues. Cross-validation is one technique used to mitigate this, but that is expensive since you need to do multiple runs over the data, which further taxes a computing infrastructure already running at max capacity. GPUs are preferred when training a large network since these systems train at least two orders of magnitude faster than CPUs [7]. Large-scale experiments are simply not feasible without using GPUs. However, there is a tradeoff to gain this performance. Since all our GPUs use the NVIDIA CUDA® Deep Neural Network library (cuDNN) [8], a GPU-accelerated library of primitives for deep neural networks, it adds an element of randomness into the experiment. When a GPU is used to train a network in TensorFlow, it automatically searches for a cuDNN implementation. NVIDIA’s cuDNN implementation provides algorithms that increase the performance and help the model train quicker, but they are non-deterministic algorithms [9,10]. Since our networks have many complex layers, there is no easy way to avoid this randomness. Instead of comparing each epoch, we compare the average performance of the experiment because it gives us a hint of how our model is performing per experiment, and if the changes we make are efficient. In this poster, we will discuss a variety of issues related to reproducibility and introduce ways we mitigate these effects. For example, TensorFlow uses a random number generator (RNG) which is not seeded by default. TensorFlow determines the initialization point and how certain functions execute using the RNG. The solution for this is seeding all the necessary components before training the model. This forces TensorFlow to use the same initialization point and sets how certain layers work (e.g., dropout layers). However, seeding all the RNGs will not guarantee a controlled experiment. Other variables can affect the outcome of the experiment such as training using GPUs, allowing multi-threading on CPUs, using certain layers, etc. To mitigate our problems with reproducibility, we first make sure that the data is processed in the same order during training. Therefore, we save the data from the last experiment and to make sure the newer experiment follows the same order. If we allow the data to be shuffled, it can affect the performance due to how the model was exposed to the data. We also specify the float data type to be 32-bit since Python defaults to 64-bit. We try to avoid using 64-bit precision because the numbers produced by a GPU can vary significantly depending on the GPU architecture [11-13]. Controlling precision somewhat reduces differences due to computational noise even though technically it increases the amount of computational noise. We are currently developing more advanced techniques for preserving the efficiency of our training process while also maintaining the ability to reproduce models. In our poster presentation we will demonstrate these issues using some novel visualization tools, present several examples of the extent to which these issues influence research results on electroencephalography (EEG) and digital pathology experiments and introduce new ways to manage such computational issues. 
    more » « less
  5. Grid Engine is a Distributed Resource Manager (DRM), that manages the resources of distributed systems (such as Grid, HPC, or Cloud systems) and executes designated jobs which have requested to occupy or consume those resources. Grid Engine applies scheduling policies to allocate resources for jobs while simultaneously attempting to maintain optimal utilization of all machines in the distributed system. However, due to the complexity of Grid Engine's job submission commands and complicated resource management policies, the number of faulty job submissions in data centers increases with the number of jobs being submitted. To combat the increase in faulty jobs, Grid Engine allows administrators to design and implement Job Submission Verifiers (JSV) to verify jobs before they enter into Grid Engine. In this paper, we will discuss a Job Submission Verifier that was designed and implemented for Univa Grid Engine, a commercial version of Grid Engine, and thoroughly evaluated at the High Performance Computing Center of Texas Tech University. Our newly developed JSV communicates with Univa Grid Engine (UGE) components to verify whether a submitted job should be accepted as is, or modified then accepted, or rejected due to improper requests for resources. It had a substantial positive impact on reducing the number of faulty jobs submitted to UGE by far. For instance, it corrected 28.6% of job submissions and rejected 0.3% of total jobs from September 2018 to February 2019, that may otherwise lead to long or infinite waiting time in the job queue. 
    more » « less