The current state of practice in supercomputer resource allocation places jobs from different users on disjoint nodes both in terms of time and space. While this approach largely guarantees that jobs from different users do not degrade one another's performance, it does so at high cost to system throughput and energy efficiency. This focused study presents job striping, a technique that significantly increases performance over the current allocation mechanism by colocating pairs of jobs from different users on a shared set of nodes. To evaluate the potential of job striping in large-scale environments, the experiments are run at the scale of 128 nodes on the state-of-the-art Gordon supercomputer. Across all pairings of 1024 process network-attached storage parallel benchmarks, job striping increases mean throughput by 26% and mean energy efficiency by 22%. On pairings of the real applications Gyrokinetic Toroidal Code (GTC), Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS), and MIMD Lattice Computation (MILC) at equal scale, job striping improves average throughput by 12% and mean energy efficiency by 11%. In addition, the study provides a simple set of heuristics for avoiding low performing application pairs. 239 Figure 3. Increase in system throughput (STP) over compact when applying job spreading and striping to the NAS parallel benchmarks and GTC, LAMMPS, and MILC. Figure 3 shows the performance results for the first set of experiments with the NPBs and the second set of experiments with GTC, LAMMPS, and MILC. For the NPBs, the mean performance increase from job spreading is 50%. If one examines striped coschedules of non-identical NPBs, the average performance increase is 26%. If one selects the best running mate other than embarrassingly parallel (EP) for each benchmark, then the average increase in performance is 36%. We choose to exclude EP because EP is minimally contentious. Each EP task's working set fits entirely in the private levels of cache, and EP spends very little time in active communication. Because of these traits, EP universally causes every application that it stripes with to achieve its best striped performance. Thus, for the sake of fairness and realism, we exclude these results from the 'Best' average. For the NPBs, random striping yields about 50% of the performance benefit of job spreading, and striping each job with its best running mate provides 70% of the performance benefit of spreading. This trend continues for real applications as well. For GTC, LAMMPS, and MILC, job spreading increases throughput by 23% and the mean heterogeneous striping and the mean best striping improve performance by 12% and 16%, respectively. PERFORMANCE RESULTS Compact versus spreading versus striping Network-attached storage parallel benchmarksIn this section, we examine the increase in collective throughput and energy efficiency for pairs of striped NPBs. The results are presented in Figure 4. For completeness, we run all pairwise combinations. This includes both heterogeneous pairings ...

Ensuring the quality of service (QoS) for latency-sensitive applications while allowing co-locations of multiple applications on servers is critical for improving server utilization and reducing cost in modern warehouse-scale computers (WSCs). Recent work relies on static profiling to precisely predict the QoS degradation that results from performance interference among co-running applications to increase the number of "safe" co-locations. However, these static profiling techniques have several critical limitations: 1) a priori knowledge of all workloads is required for profiling, 2) it is difficult for the prediction to capture or adapt to phase or load changes of applications, and 3) the prediction technique is limited to only two co-running applications.To address all of these limitations, we present Bubble-Flux, an integrated dynamic interference measurement and online QoS management mechanism to provide accurate QoS control and maximize server utilization. Bubble-Flux uses a Dynamic Bubble to probe servers in real time to measure the instantaneous pressure on the shared hardware resources and precisely predict how the QoS of a latency-sensitive job will be affected by potential co-runners. Once "safe" batch jobs are selected and mapped to a server, Bubble-Flux uses an Online Flux Engine to continuously monitor the QoS of the latency-sensitive application and control the execution of batch jobs to adapt to dynamic input, phase, and load changes to deliver satisfactory QoS. Batch applications remain in a state of flux throughout execution. Our results show that the utilization improvement achieved by Bubble-Flux is up to 2.2x better than the prior static approach.

No abstract

Abstract. Co-location, where multiple jobs share compute nodes in large-scale HPC systems, has been shown to increase aggregate throughput and energy efficiency by 10-20%. However, system operators disallow co-location due to fair-pricing concerns, i.e., a pricing mechanism that considers performance interference from co-running jobs. In the current pricing model, application execution time determines the price, which results in unfair prices paid by the minority of users whose jobs suffer from co-location. This paper presents POPPA, a runtime system that enables fair pricing by delivering precise online interference detection and facilitates the adoption of supercomputers with co-locations. POPPA leverages a novel shutter mechanism -a cyclic, fine-grained interference sampling mechanism to accurately deduce the interference between co-runners -to provide unbiased pricing of jobs that share nodes. POPPA is able to quantify inter-application interference within 4% mean absolute error on a variety of co-located benchmark and real scientific workloads.

Ensuring the quality of service (QoS) for latency-sensitive applications while allowing co-locations of multiple applications on servers is critical for improving server utilization and reducing cost in modern warehouse-scale computers (WSCs). Recent work relies on static profiling to precisely predict the QoS degradation that results from performance interference among co-running applications to increase the number of "safe" co-locations. However, these static profiling techniques have several critical limitations: 1) a priori knowledge of all workloads is required for profiling, 2) it is difficult for the prediction to capture or adapt to phase or load changes of applications, and 3) the prediction technique is limited to only two co-running applications. To address all of these limitations, we present Bubble-Flux , an integrated dynamic interference measurement and online QoS management mechanism to provide accurate QoS control and maximize server utilization. Bubble-Flux uses a Dynamic Bubble to probe servers in real time to measure the instantaneous pressure on the shared hardware resources and precisely predict how the QoS of a latency-sensitive job will be affected by potential co-runners. Once "safe" batch jobs are selected and mapped to a server, Bubble-Flux uses an Online Flux Engine to continuously monitor the QoS of the latency-sensitive application and control the execution of batch jobs to adapt to dynamic input, phase, and load changes to deliver satisfactory QoS. Batch applications remain in a state of flux throughout execution. Our results show that the utilization improvement achieved by Bubble-Flux is up to 2.2x better than the prior static approach.