More Like this

1. PPipe: Efficient Video Analytics Serving on Heterogeneous GPU Clusters via Pool-Based Pipeline Parallelism

   Kong, Z Jonny; Xu, Qiang; Hu, Y Charlie (July 2025, USENIX)

   With the rapid innovation of GPUs, heterogeneous GPU clusters in both public clouds and on-premise data centers have become increasingly commonplace. In this paper, we demonstrate how pipeline parallelism, a technique wellstudied for throughput-oriented deep learning model training, can be used effectively for serving latency-bound model inference, e.g., in video analytics systems, on heterogeneous GPU clusters. Our work exploits the synergy between diversity in model layers and diversity in GPU architectures, which results in comparable inference latency for many layers when running on low-class and high-class GPUs. We explore how such overlooked capability of low-class GPUs can be exploited using pipeline parallelism and present a novel inference serving system, PPipe, that employs pool-based pipeline parallelism via an MILP-based control plane and a data plane that performs resource reservation-based adaptive batching. Evaluation results on diverse workloads (18 CNN models) show that PPipe achieves 41.1%–65.5% higher utilization of low-class GPUs while maintaining high utilization of high-class GPUs, leading to 32.2%–75.1% higher serving throughput compared to various baselines. 

   more » « less
2. Kube-Knots: Resource Harvesting through Dynamic Container Orchestration in GPU-based Datacenters

   https://doi.org/10.1109/cluster.2019.8891040  

   Thinakaran, Prashanth; Gunasekaran, Jashwant Raj; Sharma, Bikash; Kandemir, Mahmut Taylan; Das, Chita R. (September 2019, IEEE International Conference on Cluster Computing (CLUSTER))

   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
3. Exploiting ML Task Correlation in the Minimization of Capital Expense for GPU Data Centers

   Subramaniyan, Srinivasan; Wang, Xiaorui (November 2025, The 44th IEEE International Performance Computing and Communications Conference (IPCCC 2025), Austin, Texas, November 2025.)

   Efficiently scheduling ML training tasks in a GPU data center presents a significant research challenge. Existing solutions commonly schedule such tasks based on their demanded GPU utilization, but simply assume that the GPU utilization of each task can be approximated as a constant number (e.g., by using the peak value), even though the ML training tasks commonly have their GPU utilization varying significantly over time. Using a constant number to schedule tasks can result in an overestimation of the needed GPU count and, therefore, a high capital expense for GPU purchases. To address this, we design CorrGPU, a correlation-aware GPU scheduling algorithm that considers the utilization correlation among different tasks to minimize the number of needed GPUs in a data center. CorrGPU is designed based on a key observation from the analysis of real ML traces that different tasks do not have their GPU utilization peak at exactly the same time. As a result, if the correlations among tasks are considered in scheduling, more tasks can be scheduled onto the same GPUs, without extending the training duration beyond the desired due time. For a GPU data center to be constructed based on an estimated ML workload, CorrGPU can help the operators purchase a smaller number of GPUs, thus minimizing their capital expense. Our hardware testbed results demonstrate CorrGPU’s potential to reduce the number of GPUs needed. Our simulation results on real-world ML traces also show that CorrGPU outperforms several state-of-the-art solutions by reducing capital expense by 20.88%. This work was published in the 44th IEEE International Performance Computing and Communications Conference (IPCCC 2025) in November 2025. Our paper received the Best Paper Runner-up Award from IPCCC. 

   more » « less
4. Latency-guaranteed Co-location of Inference and Training for Reducing Data Center Expenses

   Chen, Guoyu; Subramaniyan, Srinivasan; Wang, Xiaorui (July 2024, IEEE)

   Today's data centers often need to run various machine learning (ML) applications with stringent SLO (Service-Level Objective) requirements, such as inference latency. To that end, data centers prefer to 1) over-provision the number of servers used for inference processing and 2) isolate them from other servers that run ML training, despite both use GPUs extensively, to minimize possible competition of computing resources. Those practices result in a low GPU utilization and thus a high capital expense. Hence, if training and inference jobs can be safely co-located on the same GPUs with explicit SLO guarantees, data centers could flexibly run fewer training jobs when an inference burst arrives and run more afterwards to increase GPU utilization, reducing their capital expenses. In this paper, we propose GPUColo, a two-tier co-location solution that provides explicit ML inference SLO guarantees for co-located GPUs. In the outer tier, we exploit GPU spatial sharing to dynamically adjust the percentage of active GPU threads allocated to spatially co-located inference and training processes, so that the inference latency can be guaranteed. Because spatial sharing can introduce considerable overheads and thus cannot be conducted at a fine time granularity, we design an inner tier that puts training jobs into periodic sleep, so that the inference jobs can quickly get more GPU resources for more prompt latency control. Our hardware testbed results show that GPUColo can precisely control the inference latency to the desired SLO, while maximizing the throughput of the training jobs co-located on the same GPUs. Our large-scale simulation with a 57-day real-world data center trace (6500 GPUs) also demonstrates that GPUColo enables latency-guaranteed inference and training co-location. Consequently, it allows 74.9% of GPUs to be saved for a much lower capital expense. 

   more » « less
5. Closing the Gap: Improving the Accuracy of gem5’s GPU Models

   Vishnu Ramadas; Daniel Kouchekinia; Ndubuisi Osuji; Matthew D. Sinclair (June 2023, 5th gem5 Users' Workshop)

   In recent years, we have been enhancing and updating gem5’s GPU support, including enhanced gem5’s GPU support to enable running ML workloads. Moreover, we created, validated, and released a Docker image with the proper software and libraries needed to run AMD’s GCN3 and Vega GPU models in gem5. With this container, users can run the gem5 GPU model, as well as build the ROCm applications that they want to run in the GPU model, out of the box without needing to properly install the appropriate ROCm software and libraries. Additionally, we updated gem5 to make it easier to reproduce results, including releasing support for a number of GPU workloads in gem5-resources and enabling continuous integration testing for a variety of GPU workloads. Current gem5 support focuses on Carrizo- and Vega-class GPUs. Unfortunately, these models do not always provide high accuracy relative to the equivalent ”real” GPUs. This leads to a mismatch in expectations: when prototyping new optimizations in gem5 users may draw the wrong conclusions about the efficacy of proposed optimizations if gem5’s GPU models do not provide high fidelity. Accordingly, to help bridge this divide, we design a series of micro-benchmarks designed expose the latencies, bandwidths, and sizes of a variety of GPU components on real GPUs. By iteratively applying fixes and improvements to gem’s GPU model, we significantly improve its fidelity relative to real AMD GPUs. 

   more » « less