Benchmarking
Uncertainty Propagation through Summations and OpenTURNS
System analysis in safety-critical sectors like nuclear energy is frequently constrained by the computational costs of Monte Carlo simulations. Efficiently propagating uncertainty through summations is a basic requirement in design safety analysis. This example application demonstrates system analysis using mathematical methods derived from the OpenTURNS python package [1] but implemented using Signaloid’s UxHw technology. (Signaloid has also created a port of the OpenTURNS Python package, accelerated by our UxHw technology.)
Figure 1. Monte Carlo UQ implementation
When running on the Signaloid Cloud Compute Engine (SCCE), implementations of system safety analysis can replace the traditional approach of sampling followed by sample evaluation with a direct computation on a representation of the probability distribution. Thus, in a single computation, the SCCE computes the same output distribution that traditional Monte Carlo simulations require hundreds of thousands of iterations to achieve.
Figure 2. Signaloid UQ implementation
The summation kernel, taken from OpenTURNS, and running on a single-threaded Signaloid C0Pro-Jupiter-M core, achieves runtimes 1462× faster than an optimized C language Monte-Carlo-based implementation (i.e., the industry standard and status quo) of the same kernel running on an Amazon EC2 R7iz instance. With 95% confidence, a 12.5M-iteration Monte Carlo implementation matches the accuracy of the Signaloid UxHw-based version; if higher confidence is required for the Monte Carlo results, the speedup of the UxHw approach becomes even more pronounced.
Key Performance Indicator
Signaloid Platform Solution
Competing Solution
Signaloid Benefit
Speed for the same uncertainty quantification accuracy.
Run existing non-Monte-Carlo code and use either the Signaloid Compute Engine's automated ingestion of distribution information, or use the Signaloid UxHw API to set program variables as probability distributions.
Run existing Monte Carlo code, or, starting from non-Monte-Carlo code, modify code to implement Monte Carlo sampling, iteration, and aggregation of the results from the Monte Carlo iterations of the computation.
1462x faster execution time than 12.5M-iteration Monte Carlo, while achieving same fidelity of full distribution result.
Plot of output distribution when running on a Signaloid C0Pro-Jupiter-M core that provides the 1462× speedup.
Plot of the output of an 12.5M-iteration Monte Carlo for this use case. Signaloid C0Pro-Jupiter-M core is 1462× faster than this Monte Carlo, while achieving the same or better Wasserstein distance to the ground-truth (20M-iteration) Monte Carlo.
Plot of ground truth (20M-iteration) Monte Carlo.
Benchmarking Methodology
Monte Carlo simulations work by statistical sampling and therefore each multi-iteration Monte Carlo run will result in a slightly different output distribution. By contrast, Signaloid's platform is deterministic and each run produces the same distribution for a given Signaloid C0 core type.
The performance improvements are calculated by comparing Signaloid's platform with a Monte Carlo simulation of a similar quality of distribution. First, we run a large Monte Carlo simulation (about 50M iterations) on an AWS r7iz high-performance AWS instance: We use this distribution result as a baseline or ground truth reference of distribution quality. Then we calculate the performance of Signaloid's technology, and compare it with the performance of a Monte Carlo iteration count where the output distribution's Wasserstein distance (to the output distribution of the ground truth reference) is as accurate as the Signaloid-core-executed algorithm's output distribution, with 95% confidence level.
Performance data based on Fall 2025 release of Signaloid's technology.