Pith. sign in

REVIEW 6 major objections 4 minor 1 cited by

HPC Application Parameter Autotuning on Edge Devices: A Bandit Learning Approach

T0 review · 6 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read LASP, a multi-armed bandit autotuner that runs on resource-constrained edge devices, claims to reach within 12% of the optimal HPC configuration while using far less CPU and memory than Bayesian optimization baselines.

desk verdict LASP is a lightweight UCB autotuner for edge devices with a sound practical evaluation, but its edge-to-HPC transfer claim rests on a 25% bound that is weaker than the reported default gains, so the paper needs major revision. read the letter →

arxiv 2501.01057 v1 pith:NH5W5A7D submitted 2025-01-02 cs.PF cs.LGcs.SYeess.SY

Abrar Hossain , Abdel-Hameed A. Badawy , Mohammad A. Islam , Tapasya Patki , Kishwar Ahmed This is my paper · ORCID
classification cs.PFcs.LGcs.SYeess.SY
keywords HPCparameterautotuningedgedevicesmulti-armedbanditupperconfidenceboundapplicationsperformancemodelinglow-fidelitytransferLASP
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper sets out to show that HPC parameter autotuning can be moved off expensive supercomputers and onto cheap, resource-constrained edge devices. Its proposed algorithm, LASP, is a multi-armed bandit that treats each parameter configuration as an arm and uses an upper confidence bound rule to balance exploring new configurations with exploiting good ones, updating in real time from execution-time and power measurements. The authors claim that LASP reaches within 12% of the optimal configuration even for the large Hypre search space when optimizing execution time, and delivers 6-14% performance gains over default configurations across four HPC applications. If true, this gives users a lightweight, portable tuning step that adapts to noisy measurements and changing workloads, and whose results can be transferred from low-fidelity edge runs to high-fidelity HPC execution.

What carries the argument

The mechanism carrying the argument is the upper confidence bound (UCB) arm-selection rule, Eq. (2): $UCB(x,t) = R_x + \sqrt{2 \ln t / N_x}$, where $R_x$ is the weighted reward of configuration $x$ and $N_x$ is the number of times it has been tried. UCB is the 'optimism under uncertainty' device that forces LASP to keep sampling under-explored configurations while shifting weight to arms that have looked good so far. Around it the paper builds a reward model, Eq. (5), that maps measured execution time and power consumption into a scalar reward via inverse MinMax normalization and user weights $\alpha$ and $\beta$, and a low-fidelity-to-high-fidelity surrogate workflow in which edge-device runs at reduced mesh or zone sizes stand in for the final HPC execution.

What would settle it

Run LASP on the edge device at low fidelity, take the configuration it returns, and measure its execution time at high fidelity on the target HPC cluster against a true exhaustive-search oracle on that cluster; the core workflow is falsified if the transferred configuration lands more than 25% away from that cluster-side oracle, which is the gap the paper already reports for the top-20 low-fidelity set.

Watch

Extended reading notes

Core claim

LASP's central claim is that a UCB-based multi-armed bandit can carry out HPC configuration autotuning on an edge device with low overhead. Each unique application-parameter combination is an arm; the reward for an arm is the inverse of a MinMax-normalized weighted sum of execution time and power consumption, with user-set weights $\alpha$ and $\beta$; and the algorithm selects the arm maximizing $R_x + \sqrt{2 \ln t / N_x}$. After $T$ iterations it returns the most frequently chosen configuration. Experiments on Lulesh, Kripke, Clomp, and Hypre on a Jetson Nano show that this procedure converges to within 12% of the oracle for the largest search space (Hypre) when time-optimized, and improves over default settings by 6-14% across applications. The paper also reports that the top-20 low-fidelity configurations land within 25% of the high-fidelity oracle, and that LASP keeps working when 5-15% random noise is added to measurements, which is offered as evidence of adaptability to volatile edge environments.

Load-bearing premise

The load-bearing premise is that parameter configurations that win at low fidelity on the edge device remain near-optimal at high fidelity on the actual HPC platform where the application will run; if that transfer breaks, the whole tuning-on-edge workflow produces configurations that are not optimal for the real target.

Editorial extensions

If this is right

  • Autotuning can be offloaded from expensive HPC node hours to cheap edge devices, with tuning cost paid once and the found configuration reused on the cluster.
  • Users can express a time-versus-power trade-off through $\alpha$ and $\beta$, and LASP will steer to different optima accordingly, as shown in the performance-gain analysis.
  • Because LASP receives online feedback, it can re-tune when the workload, input size, or device state changes, and it tolerates measurement noise up to 15% in the paper's sensitivity study.
  • The resource-footprint comparison with BLISS indicates LASP can run alongside other edge workloads, making autotuning feasible in co-located, power-capped settings (5W and 10W modes).
  • The within-12%-of-oracle result on Hypre suggests the approach scales to very large configuration spaces (92,160 configs) where exhaustive search would be impractical.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Whether the 12%-of-oracle bound holds on the high-fidelity cluster, not just on the edge device, is not measured in the paper; a direct test would compare cluster-side distance to oracle for LASP's returned configuration.
  • The paper uses the standard UCB regret bound without deriving a new one; the novelty is the application to edge-device HPC autotuning, not a new regret guarantee, so theoretical gains would need a separate analysis for non-stationary rewards.
  • Since power consumption saturates on the Jetson Nano under compute-heavy loads, the time objective is the more discriminative signal; on devices where power varies more, the power-focused mode might behave differently than reported here.
  • A natural extension is to test LASP against non-stochastic bandit methods (e.g., EXP3 or Hyperband-style best-arm identification) in the same edge setting, since the paper only compares to a Bayesian-optimization baseline.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

6 major / 4 minor

Summary. The paper introduces LASP, a multi-armed bandit (UCB) algorithm for tuning application-level parameters of HPC proxy applications on edge devices. The intended workflow is to tune at low fidelity on an NVIDIA Jetson Nano and transfer the resulting configuration to high-fidelity execution on an HPC platform (Fig. 1). The paper reports exhaustive-search-based oracle comparisons, a distance-from-oracle of within 12% for Hypre, gains of 6-14% over default settings for four applications, and a lower CPU/memory footprint than the BLISS autotuner. The evaluation covers Lulesh, Kripke, Clomp, and Hypre, and includes a sensitivity study with synthetic measurement noise.

Significance. If the reported behavior were produced by a genuinely online algorithm, LASP would be a useful contribution: the problem is real, the experiments span four distinct proxy applications, and the resource-footprint comparison with BLISS is a worthwhile datapoint. The paper is also candid in Section IV-B about scalability limitations. However, the significance currently depends on closing three gaps: the normalization step in Algorithm 1 requires global information that an online algorithm does not have; the low-fidelity-to-high-fidelity transfer evidence is only a 25% bound, which is weaker than the reported default-configuration gains; and tuning quality is not compared with any non-default autotuning baseline. With those gaps closed, the paper could be a solid systems contribution, but in its current form the central claims are not supported.

major comments (6)
  1. [Section IV, Algorithm 1 line 2] Algorithm 1 (line 2) normalizes execution time τ and power ρ by global minima and maxima over the entire configuration space. In an online bandit setting, τ_min, τ_max, ρ_min, and ρ_max are unavailable until every configuration has been evaluated; if they are obtained by a prior exhaustive scan, LASP already knows the oracle (the configuration attaining the minimum execution time), and its exploration is a post-hoc ranking rather than an online search. If they are instead updated online, the reward of every previously sampled arm changes whenever an extremum is updated, and the UCB means and counts in Eq. (2) are not defined consistently. Because the 12% near-oracle result in Section V-D is computed with this normalization, the central claim of a lightweight online autotuner is not supported as stated. The authors should either give an online-feasible normalization (e.g., fixed known bounds or streaming min-max updates with a re-derivation of the UCB update), or rerun the evaluation with such a scheme.
  2. [Section V-D and Section II.C] The claim that 'LASP can reach within 12% of the optimal configuration' is measured on the Jetson Nano (Section V-A), i.e., distance to the low-fidelity oracle. The only evidence for transfer to the target HPC platform is Fig. 2(a), which shows that top-20 low-fidelity configurations land within 25% of the high-fidelity oracle on an Intel i7-14700. Since the default-configuration gains in Fig. 8 are 6-14%, a 25% transfer error can place the selected configuration at or below default performance on the HPC target. The paper needs to report the distance from the high-fidelity oracle for the exact configuration returned by LASP, and ideally report high-fidelity default-vs-LASP performance gains on the target system.
  3. [Section III, Eq. (7), and Eq. (5)] The regret bound cited after Eq. (7) is the standard UCB bound for rewards in [0,1], and Section III explicitly assumes rewards are constrained to [0,1]. However, the reward in Eq. (5) is a weighted sum of reciprocals of normalized execution time and power. A normalized metric can approach zero for a near-optimal configuration, making the reciprocal unbounded and generally larger than 1. Therefore the bound in Eq. (7) cannot be applied to LASP's reward as defined, and the regret curves in Section V-E evaluate a different quantity. The paper must either transform the reward to a bounded [0,1] range while preserving the optimum, or provide an appropriate regret analysis for the actual unbounded reward.
  4. [Section V-D and Fig. 10] The only comparison against another autotuner is the CPU/memory footprint comparison with BLISS in Fig. 10. The text states that LASP 'did not do better in terms of efficiently finding the optimal parameters' but gives no quantitative comparison of tuning quality, number of evaluations to reach the optimum, or final distance from the oracle for BLISS or any other baseline. Without such a comparison, the contribution of LASP relative to existing Bayesian optimization or Hyperband approaches is not established; performance gains over a fixed default configuration alone do not position the method against the state of the art.
  5. [Section V-E, Fig. 11] The regret analysis is reported for the 'best-run (one time least regret run)'. Selecting the single best run among multiple random seeds does not provide an expectation or high-probability statement, and it biases the reported regret in a favorable direction. The paper should report mean regret with confidence intervals over multiple runs, as is done for the distance-from-oracle measurement in Fig. 9, rather than a single cherry-picked trajectory.
  6. [Section V-F] The sensitivity study adds 5%, 10%, and 15% random noise to measured data. This is observation noise, not a changing environment: the reward distributions themselves are stationary throughout the run. No experiment changes the workload, input size, or reward distribution over time, so the abstract's claim that LASP 'adapts seamlessly to changing environments' is not supported. The authors should test adaptation under a non-stationary reward process, for example by switching problem size or input type midway through the run, and show that LASP recovers the new optimum.
minor comments (4)
  1. [Section IV-A] The text says LASP is 'demonstrated in Section 10', but the manuscript has no Section 10; this is presumably a reference to a figure or another section and should be corrected.
  2. [Section V-D] There are textual errors in this section: 'performance gain performance gain' is duplicated, and 'we ranLASP' is missing a space. These should be fixed in revision.
  3. [Figures 6 and 7] The heatmaps use color intensity without a colorbar or numeric scale, so selection frequencies cannot be read quantitatively; adding colorbars would make the convergence claims more verifiable.
  4. [Section II.C, Fig. 2(b)] The overlap count in Fig. 2(b) is reported without stating the total number of configurations considered or the expected overlap under random selection, which makes it difficult to judge whether the observed overlap is substantially better than chance.

Circularity Check

1 steps flagged · score 8.0 of 10

LASP's near-oracle claim reduces to MinMax normalization constants fitted from the full dataset.

  1. self definitional [Algorithm 1, Step 2; Section IV, Eq. (5); Section V-D, Fig. 9]
    "Apply MinMax normalization: τ ← τ−min(τ)/max(τ)−min(τ), ρ← ρ−min(ρ)/max(ρ)−min(ρ) ... The normalized execution time τx is calculated as: τx = τ −τmin/τmax−τmin, where τmin and τmax are minimum and maximum execution times, respectively. ... freward(x) = α × (1/µ(τx)) + β × (1/µ(ρx))"

    The normalization uses the global minimum execution time over the complete configuration space, so the oracle configuration receives normalized value 0. Its reward 1/µ(τx) is then infinite, making its UCB value infinite and causing it to be selected on every round; xopt = argmax Nx consequently returns the precomputed oracle. The claimed 'within 12% of the optimal configuration' and the performance gains in Fig. 8 are therefore encoded by normalization constants fitted from the full dataset rather than discovered by bandit exploration.

full rationale

The central evaluation claim, 'LASP can reach within 12% of the optimal configuration even in large parameter spaces, such as those of Hypre, when optimizing for execution time' (Section V-D), is not an independent prediction as written. Algorithm 1 first applies MinMax normalization over the complete reward metrics, then defines reward as the inverse of the normalized metric. The configuration with the global minimum execution time has normalized value 0, hence infinite reward and infinite UCB, and is selected every iteration; argmax Nx then returns that oracle configuration. Thus both the distance-to-oracle and the default-configuration gains are built into the reward definition. No load-bearing self-citation chain is present: reference [37] is used only for a generic edge-computing observation, and the UCB regret bound is cited from standard external literature. The low-fidelity to high-fidelity transfer evidence (top-20 LF configurations within 25% of the HF oracle) is weak empirical support but is not circular. The 'best-run (one time least regret run)' reporting is a favorable-selection artifact, but it was not counted as circularity.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the stochastic bandit assumptions (i.i.d. rewards bounded in [0,1]) from Section III, the low-fidelity/high-fidelity transfer assumption from Section II.C, and the MinMax normalization step in Algorithm 1 that requires global extrema. The last is a practical flaw: an online algorithm cannot know the min and max runtime over the full search space without having already searched it.

free parameters (3)
  • alpha (execution time weight) = 0.8 (time-focused), 0.2 (power-focused)
    User-defined weight controlling the optimization objective, chosen by hand in experiments (Section V-D).
  • beta (power consumption weight) = 1 - alpha
    Companion user weight, implied values from alpha in Section III and Section V-D.
  • tau_min, tau_max, rho_min, rho_max (MinMax normalization extrema) = Not reported
    Global minima and maxima of execution time and power across the entire configuration space, used in Algorithm 1 line 2. Computing them requires exhaustive search, the very cost the method aims to avoid.
assumptions (5)
  • domain assumption Rewards are i.i.d. per configuration (stochastic bandit)
    Stated in Section III, but conflicts with the paper's later claim of dynamic environments.
  • standard math Rewards lie in [0,1] for the UCB regret bound
    Assumed in Section III but violated by Eq. 5, which uses 1/normalized_time, unbounded above when normalized time approaches 0.
  • ad hoc to paper Low-fidelity and high-fidelity optima overlap
    Section II.C and Fig. 2; load-bearing for the edge-to-HPC transfer workflow but only partially validated.
  • ad hoc to paper Global MinMax statistics are known for normalization
    Algorithm 1 line 2 requires min and max over the full search space; online operation cannot know these without exhaustive search.
  • domain assumption A single static oracle configuration exists
    The evaluation measures distance to the best static configuration from exhaustive search; in dynamic environments no single best configuration may exist.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HPC Application Parameter Autotuning on Edge Devices: A Bandit Learning Approach." pith.science (2026). https://pith.science/paper/NH5W5A7D

@misc{pith2026250101057,
  author       = {Pith},
  title        = {Pith review of: HPC Application Parameter Autotuning on Edge Devices: A Bandit Learning Approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NH5W5A7D}},
  note         = {Machine review of arXiv:2501.01057}
}
read the original abstract

The growing necessity for enhanced processing capabilities in edge devices with limited resources has led us to develop effective methods for improving high-performance computing (HPC) applications. In this paper, we introduce LASP (Lightweight Autotuning of Scientific Application Parameters), a novel strategy designed to address the parameter search space challenge in edge devices. Our strategy employs a multi-armed bandit (MAB) technique focused on online exploration and exploitation. Notably, LASP takes a dynamic approach, adapting seamlessly to changing environments. We tested LASP with four HPC applications: Lulesh, Kripke, Clomp, and Hypre. Its lightweight nature makes it particularly well-suited for resource-constrained edge devices. By employing the MAB framework to efficiently navigate the search space, we achieved significant performance improvements while adhering to the stringent computational limits of edge devices. Our experimental results demonstrate the effectiveness of LASP in optimizing parameter search on edge devices.

Figures

Figures reproduced from arXiv: 2501.01057 by the authors.

Figure 1
Figure 1. Framework to leverage edge devices to find the optimal parameters [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overlap of optimal configurations on low- and high-fidelity setting. (a) [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 4
Figure 4. Runtime variability of Kripke for different parameters considered [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figures from the paper (7 more)
Figure 5
Figure 5. Figure 5: Block diagram of the LASP. Algorithm 1 Lightweight Autotuning of Scientific Application Parameters (LASP) Input: Configuration space (χ), total iterations (T), execution time weight parameter (α), and power consumption weight parameter (β) 1: Initialization: Dictionary…
Figure 6
Figure 6. Figure 6: (a) and (b) Exploration of the parameter space with Power as [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 8
Figure 8. Figure 8: Performance gain for different applications. [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: LASP reaches closest to the Oracle with very few iterations. However, LASP’s fast convergence in finding the optimal configuration makes up for its performance in larger parameter spaces. We run LASP 100 times in order to see the mean distance from the oracle across di…
Figure 12
Figure 12. Figure 12: Performance analysis with synthetic error in measurement data [PITH_FULL_IMAGE:figures/full_fig_p010_12.png]
Figure 10
Figure 10. Figure 10: Resource Utilization of LASP compared to BLISS (a) Lulesh (b) Kripke (c) Clomp (d) Hypre [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]
Figure 11
Figure 11. Figure 11: Regret analysis for Lulesh, Kripke, Clomp and Hypre. as defined in Equation 1. The results, illustrated in [PITH_FULL_IMAGE:figures/full_fig_p010_11.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Enhanced Intrusion Detection in IIoT Networks: A Lightweight Approach with Autoencoder-Based Feature Learning

    cs.LG 2025-01 reject novelty 2.0 of 10

    A decision tree trained on six autoencoder-learned features reports 99.94% accuracy and F1 on the Edge-IIoTset benchmark, with 0.185 ms per-sample inference on a Jetson Nano.

Reference graph

Works this paper leans on

48 extracted references · 47 canonical work pages · cited by 1 Pith paper

  1. [1]

    Edge computing: Vision and challenges,

    W. Shi, J. Cao, Q. Zhang, Y . Li, and L. Xu, “Edge computing: Vision and challenges,” IEEE internet of things journal , 2016

    work page 2016

  2. [2]

    Pcie vs. 5g: The importance of hpc at the edge

    J. Mannik, “Pcie vs. 5g: The importance of hpc at the edge.” https://onestopsystems.com/blogs/one-stop-systems-blog/ pcie-vs-5g-the-importance-of-hpc-at-the-edge, 2022

    work page 2022

  3. [3]

    5g enabled energy innovation: Advanced wireless networks for science,

    P. Beckman et al. , “5g enabled energy innovation: Advanced wireless networks for science,” tech. rep., ANL and Northwestern Univ., 2020

    work page 2020

  4. [4]

    Auto-tuning full applications: A case study,

    A. Tiwari, J. K. Hollingsworth, C. Chen, M. Hall, C. Liao, D. J. Quinlan, and Chame, “Auto-tuning full applications: A case study,” Int. J. High Perform. Comput. Appl., 2011

    work page 2011

  5. [5]

    Automated reasoning and detection of specious configuration in large systems with symbolic execution,

    Y . Hu, G. Huang, and P. Huang, “Automated reasoning and detection of specious configuration in large systems with symbolic execution,” in (OSDI’20)

    work page

  6. [6]

    Software challenges in extreme scale systems,

    V . Sarkar, W. Harrod, and A. E. Snavely, “Software challenges in extreme scale systems,” in Journal of Physics: Conference Series , vol. 180, p. 012045, IOP Publishing, 2009

    work page 2009

  7. [7]

    The antarex approach to autotuning and adaptivity for energy efficient hpc systems,

    C. Silvano, G. Agosta, S. Cherubin, D. Gadioli, G. Palermo, A. Bartolini, L. Benini, J. Martinovi ˇc, M. Palkoviˇc, K. Slaninov´a, et al., “The antarex approach to autotuning and adaptivity for energy efficient hpc systems,” in Proc. ACM Int. Conf. Comput. Frontiers , 2016

    work page 2016

  8. [8]

    Bestconfig: tapping the performance potential of systems via automatic configuration tuning,

    Y . Zhu, J. Liu, M. Guo, Y . Bao, W. Ma, Z. Liu, K. Song, and Y . Yang, “Bestconfig: tapping the performance potential of systems via automatic configuration tuning,” in SoCC ’17, 2017

    work page 2017

Show all 48 references
  1. [9]

    d- simplexed: Adaptive delaunay triangulation for performance modeling and prediction on big data analytics,

    Y . Chen, P. Goetsch, M. A. Hoque, J. Lu, and S. Tarkoma, “d- simplexed: Adaptive delaunay triangulation for performance modeling and prediction on big data analytics,” IEEE Trans. Big Data , 2019

    2019

  2. [10]

    Optimization by simulated annealing,

    S. Kirkpatrick, C. D. Gelatt Jr, and M. P. Vecchi, “Optimization by simulated annealing,” science, vol. 220, no. 4598, pp. 671–680, 1983

    1983

  3. [11]

    Particle swarm optimization,

    J. Kennedy and R. Eberhart, “Particle swarm optimization,” in Proceed- ings of ICNN’95-international conference on neural networks , IEEE

  4. [12]

    {TVM}: An automated {End-to-End} optimizing compiler for deep learning,

    T. Chen, T. Moreau, Z. Jiang, L. Zheng, E. Yan, H. Shen, M. Cowan, L. Wang, Y . Hu, L. Ceze,et al., “{TVM}: An automated {End-to-End} optimizing compiler for deep learning,” in 13th USENIX Symposium on Operating Systems Design and Implementation (OSDI 18) , 2018

    2018

  5. [13]

    Rfhoc: A random-forest approach to auto-tuning hadoop’s configura- tion,

    Z. Bei, Z. Yu, H. Zhang, W. Xiong, C. Xu, L. Eeckhout, and S. Feng, “Rfhoc: A random-forest approach to auto-tuning hadoop’s configura- tion,” IEEE Trans. Parallel Distrib. Syst. , vol. 27, no. 5, 2015

    2015

  6. [14]

    Efficient performance prediction for apache spark,

    S. Cheng, B. Wang, Y . Li, and Y . Guoli, “Efficient performance prediction for apache spark,” 2021

    2021

  7. [15]

    Datasize-aware high dimensional configu- rations auto-tuning of in-memory cluster computing,

    Z. Yu, Z. Bei, and X. Qian, “Datasize-aware high dimensional configu- rations auto-tuning of in-memory cluster computing,” in Proc. 23rd Int. Conf. Archit. Support Prog. Lang. Oper. Syst. , pp. 564–577, 2018

    2018

  8. [16]

    Bliss: auto-tuning complex applications using a pool of diverse lightweight learning models,

    R. B. Roy, T. Patel, V . Gadepally, and D. Tiwari, “Bliss: auto-tuning complex applications using a pool of diverse lightweight learning models,” in PLDI, 2021

    2021

  9. [17]

    Autotuning in High-Performance Computing Applications,

    P. Balaprakash, J. Dongarra, T. Gamblin, M. Hall, J. K. Hollingsworth, B. Norris, and R. Vuduc, “Autotuning in High-Performance Computing Applications,” Proceedings of the IEEE , vol. 106, Nov. 2018

    2018

  10. [18]

    Multitask and transfer learning for autotuning exascale applications,

    W. M. Sid-Lakhdar, M. M. Aznaveh, X. S. Li, and J. W. Demmel, “Multitask and transfer learning for autotuning exascale applications,” arXiv preprint arXiv:1908.05792 , 2019

    1908 arXiv

  11. [19]

    Boot- strapping parameter space exploration for fast tuning,

    J. J. Thiagarajan, N. Jain, R. Anirudh, A. Gimenez, R. Sridhar, A. Marathe, T. Wang, M. Emani, A. Bhatele, and T. Gamblin, “Boot- strapping parameter space exploration for fast tuning,” in ICS 18, 2018

    2018

  12. [20]

    Artemis: Automatic runtime tuning using machine learning,

    C. Wood, G. Georgakoudis, and D. Beckingsale, “Artemis: Automatic runtime tuning using machine learning,” in ISC 2021, Springer, 2021

    2021

  13. [21]

    Turbo: A cost- efficient configuration-based auto-tuning approach for cluster-based big data frameworks,

    H. Dou, L. Zhang, Y . Zhang, P. Chen, and Z. Zheng, “Turbo: A cost- efficient configuration-based auto-tuning approach for cluster-based big data frameworks,” J. Parallel Distrib. Comput. , vol. 177, 2023

    2023

  14. [22]

    Conex: Efficient exploration of big-data system configurations for better performance,

    R. Krishna, C. Tang, K. Sullivan, and B. Ray, “Conex: Efficient exploration of big-data system configurations for better performance,” IEEE Transactions on Software Engineering , vol. 48, no. 3, 2020

    2020

  15. [23]

    Hdconfigor: automatically tuning high dimensional configuration parameters for log search engines,

    H. Dou, P. Chen, and Z. Zheng, “Hdconfigor: automatically tuning high dimensional configuration parameters for log search engines,” IEEE Access, 20

  16. [24]

    Locat: Low-overhead online configuration auto-tuning of spark sql applications,

    J. Xin, K. Hwang, and Z. Yu, “Locat: Low-overhead online configuration auto-tuning of spark sql applications,” in Proceedings of the 2022 International Conference on Management of Data , pp. 674–684, 2022

    2022

  17. [25]

    Introduction to multi-armed bandits,

    A. Slivkins et al. , “Introduction to multi-armed bandits,” Foundations and Trends® in Machine Learning , vol. 12, no. 1-2, pp. 1–286, 2019

    2019

  18. [26]

    Pure exploration in finitely-armed and continuous-armed bandits,

    S. Bubeck, R. Munos, and G. Stoltz, “Pure exploration in finitely-armed and continuous-armed bandits,” Theor. Comput. Sci., 2011

    2011

  19. [27]

    Non-stochastic best arm identification and hyperparameter optimization,

    K. Jamieson and A. Talwalkar, “Non-stochastic best arm identification and hyperparameter optimization,” in Artif. Intell. Stat. , PMLR, 2016

    2016

  20. [28]

    Simple regret for infinitely many armed bandits,

    A. Carpentier and M. Valko, “Simple regret for infinitely many armed bandits,” in ICML, pp. 1133–1141, PMLR, 2015

    2015

  21. [29]

    Hyperband: A novel bandit-based approach to hyperparameter opti- mization,

    L. Li, K. Jamieson, G. DeSalvo, A. Rostamizadeh, and A. Talwalkar, “Hyperband: A novel bandit-based approach to hyperparameter opti- mization,” The Journal of Machine Learning Research , 2017

    2017

  22. [30]

    Portfolio choices with orthogonal bandit learning,

    W. Shen, J. Wang, Y .-G. Jiang, and H. Zha, “Portfolio choices with orthogonal bandit learning,” in Twenty-fourth international joint confer- ence on artificial intelligence , 2015

    2015

  23. [31]

    Input warping for bayesian optimization of non-stationary functions,

    J. Snoek, K. Swersky, R. Zemel, and R. Adams, “Input warping for bayesian optimization of non-stationary functions,” in International Conference on Machine Learning , pp. 1674–1682, PMLR, 2014

    2014

  24. [32]

    Waggle: An open sensor platform for edge computing,

    P. Beckman, R. Sankaran, C. Catlett, N. J. Ferrier, R. L. Jacob, and M. E. Papka, “Waggle: An open sensor platform for edge computing,” in 2016 IEEE SENSORS, Orlando, FL, USA, October 30 - November 3, 2016, pp. 1–3, IEEE, 2016

    2016

  25. [33]

    Sage: A distributed software-defined sensor network,

    N. F. Pete Beckman, “Sage: A distributed software-defined sensor network,” 2013

    2013

  26. [34]

    Optimizing cloud motion estimation on the edge with phase correlation and optical flow,

    B. A. Raut, P. Muradyan, R. Sankaran, R. C. Jackson, S. Park, S. A. Shahkarami, D. Dematties, Y . Kim, J. Swantek, N. Conrad, et al. , “Optimizing cloud motion estimation on the edge with phase correlation and optical flow,” Atmos. Meas. Tech., 2023

    2023

  27. [35]

    Goal-driven scheduling model in edge computing for smart city applications,

    Y . Kim, S. Park, S. Shahkarami, R. Sankaran, N. Ferrier, and P. Beck- man, “Goal-driven scheduling model in edge computing for smart city applications,” Journal of Parallel and Distributed Computing , 2022

    2022

  28. [36]

    Intersecting needs and challenges in scalable operating system research

    C. Engelmann, O. Kuchar, S. Boehm, M. J. Brim, T. Naughton, S. Atchley, J. Lange, B. Mintz, and E. Arenholz, “Intersecting needs and challenges in scalable operating system research.” White paper by the U.S. Department of Energy, Jan. 2022

    2022

  29. [37]

    Automating hpc model selection on edge devices,

    A. Hossain and K. Ahmed, “Automating hpc model selection on edge devices,” SC’23, 2023

    2023

  30. [38]

    Kripke-a massively parallel transport mini-app,

    A. J. Kunen, T. S. Bailey, and P. N. Brown, “Kripke-a massively parallel transport mini-app,” tech. rep., Lawrence Livermore National Lab.(LLNL), Livermore, CA (United States), 2015

    2015

  31. [39]

    Quantitative performance assessment of proxy apps and parents,

    D. F. Richards, O. Aziz, J. Cook, H. Finkel, B. Homerding, T. Judeman, P. McCorquodale, T. Mintz, and S. Moore, “Quantitative performance assessment of proxy apps and parents,” tech. rep., Lawrence Livermore National Lab.(LLNL), Livermore, CA (United States . . . , 2018

    2018

  32. [40]

    Using confidence bounds for exploitation-exploration trade- offs,

    P. Auer, “Using confidence bounds for exploitation-exploration trade- offs,” Journal of Machine Learning Research , vol. 3, no. Nov, 2002

    2002

  33. [41]

    Security analysis of iot protocols: A focus in coap,

    R. A. Rahman and B. Shah, “Security analysis of iot protocols: A focus in coap,” in 2016 3rd MEC international conference on big data and smart city (ICBDSC) , pp. 1–7, IEEE, 2016

    2016

  34. [42]

    Characterizing the per- formance of accelerated jetson edge devices for training deep learning models,

    P. SK, S. A. Kesanapalli, and Y . Simmhan, “Characterizing the per- formance of accelerated jetson edge devices for training deep learning models,” Proc. ACM Meas. Anal. Comput. Syst. , vol. 6, no. 3, 2022

    2022

  35. [43]

    Clustering algo- rithms on low-power and high-performance devices for edge computing environments,

    M. Lapegna, W. Balzano, N. Meyer, and D. Romano, “Clustering algo- rithms on low-power and high-performance devices for edge computing environments,” Sensors, vol. 21, no. 16, p. 5395, 2021

    2021

  36. [44]

    End-to-end energy models for edge cloud-based iot platforms: Application to data stream analysis in iot,

    Y . Li, A.-C. Orgerie, I. Rodero, B. L. Amersho, M. Parashar, and J.-M. Menaud, “End-to-end energy models for edge cloud-based iot platforms: Application to data stream analysis in iot,” Future Gener. Comput. Syst

  37. [45]

    Performance modeling under resource constraints using deep transfer learning,

    A. Marathe, R. Anirudh, N. Jain, A. Bhatele, J. Thiagarajan, B. Kailkhura, J.-S. Yeom, B. Rountree, and T. Gamblin, “Performance modeling under resource constraints using deep transfer learning,” in Proc. Int. Conf. High Perf. Comput. Netw. Storage Anal., pp. 1–12, 2017

    2017

  38. [46]

    hypre: A library of high performance preconditioners,

    R. D. Falgout and U. M. Yang, “hypre: A library of high performance preconditioners,” in Int. Conf. Comput. Sci. , Springer, 2002

    2002

  39. [47]

    Clomp: Ac- curately characterizing openmp application overheads,

    G. Bronevetsky, J. Gyllenhaal, and B. R. De Supinski, “Clomp: Ac- curately characterizing openmp application overheads,” International journal of parallel programming , vol. 37, pp. 250–265, 2009

    2009

  40. [48]

    Lulesh 2.0 updates and changes,

    I. Karlin, J. Keasler, and J. R. Neely, “Lulesh 2.0 updates and changes,” tech. rep., Lawrence Livermore National Lab.(LLNL), 2013

    2013

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.