With the ever-increasing need for computation of scientific applications, new application domains, and major energy constraints, the landscape of floating-point computation is changing. New floatingpoint representation formats are emerging and there is a need for tools to simulate their impact in legacy codes. In this paper, we propose an automatic tool to evaluate the effect of adapting the floating point precision for each operation over time, which is particularly useful in iterative schemes. We present a backend to emulate any IEEE-754 floating-point operation in lower precision. We tested the numerical errors resilience of our solutions thanks to Monte Carlo Arithmetic and demonstrated the effectiveness of this methodology on YALES2, a large Combustion-CFD HPC code, by achieving 28% to 67% reduction in communication volume by lowering precision.

Summary Current architecture complexity requires fine tuning of compiler and runtime parameters to achieve best performance. Autotuning substantially improves default parameters in many scenarios, but it is a costly process requiring long iterative evaluations. We propose an automatic piecewise autotuner based on CERE (Codelet Extractor and REplayer). CERE decomposes applications into small pieces called codelets: Each codelet maps to a loop or to an OpenMP parallel region and can be replayed as a standalone program. Codelet autotuning achieves better speedups at a lower tuning cost. By grouping codelet invocations with the same performance behavior, CERE reduces the number of loops or OpenMP regions to be evaluated. Moreover, unlike whole‐program tuning, CERE customizes the set of best parameters for each specific OpenMP region or loop. We demonstrate the CERE tuning of compiler optimizations, number of threads, thread affinity, and scheduling policy on both nonuniform memory access and heterogeneous architectures. Over the NAS benchmarks, we achieve an average speedup of 1.08× after tuning. Tuning a codelet is 13× cheaper than whole‐program evaluation and predicts the tuning impact with a 94.7% accuracy. Similarly, exploring thread configurations and scheduling policies for a Black‐Scholes solver on an heterogeneous big.LITTLE architecture is over 40× faster using CERE.

The analysis of brain-imaging data requires complex and often non-linear transformations to support findings on brain function or pathologies. And yet, recent work has shown that variability in the choices that one makes when analyzing data can lead to quantitatively and qualitatively different results, endangering the trust in conclusions. Even within a given method or analytical technique, numerical instabilities could compromise findings. We instrumented a structural-connectome estimation pipeline with Monte Carlo Arithmetic, a technique to introduce random noise in floating-point computations, and evaluated the stability of the derived connectomes, their features, and the impact on a downstream analysis. The stability of results was found to be highly dependent upon which features of the connectomes were evaluated, and ranged from perfectly stable (i.e. no observed variability across executions) to highly unstable (i.e. the results contained no trustworthy significant information). While the extreme range and variability in results presented here could severely hamper our understanding of brain organization in brain-imaging studies, it also leads to an increase in the reliability of datasets. This paper highlights the potential of leveraging the induced variance in estimates of brain connectivity to reduce the bias in networks alongside increasing the robustness of their applications in the detection or classification of individual differences. This paper demonstrates that stability evaluations are necessary for understanding error and bias inherent to scientific computing, and that they should be a component of typical analytical workflows.

No abstract

The analysis of brain-imaging data requires complex processing pipelines to support findings on brain function or pathologies. Recent work has shown that variability in analytical decisions, small amounts of noise, or computational environments can lead to substantial differences in the results, endangering the trust in conclusions. We explored the instability of results by instrumenting a structural connectome estimation pipeline with Monte Carlo Arithmetic to introduce random noise throughout. We evaluated the reliability of the connectomes, the robustness of their features, and the eventual impact on analysis. The stability of results was found to range from perfectly stable (i.e. all digits of data significant) to highly unstable (i.e. 0 − 1 significant digits). This paper highlights the potential of leveraging induced variance in estimates of brain connectivity to reduce the bias in networks without compromising reliability, alongside increasing the robustness and potential upper-bound of their applications in the classification of individual differences. We demonstrate that stability evaluations are necessary for understanding error inherent to brain imaging experiments, and how numerical analysis can be applied to typical analytical workflows both in brain imaging and other domains of computational sciences, as the techniques used were data and context agnostic and globally relevant. Overall, while the extreme variability in results due to analytical instabilities could severely hamper our understanding of brain organization, it also affords us the opportunity to increase the robustness of findings.