Double Precision Research Articles

Relevance of one-loop SMEFT matching in the 2HDM

The two-Higgs doublet model (2HDM) is a well understood alternative to the Standard Model of particle physics. If the new particles included in the 2HDM are at an energy scale much greater than the weak scale, the theory can be matched to the Standard Model effective field theory (SMEFT). We compute for the first time the complete one-loop matching at dimension-6. We compare its numerical impact with that of tree-level matching at dimension-8 by performing a global fit to single Higgs and precision electroweak measurements, and we emphasize the importance of comparing one-loop SMEFT results with corresponding one-loop results in the full 2HDM model. In the SMEFT, we consider the relative importance of both one-loop matching and the inclusion of renormalization group evolution. Our results demonstrate the necessity of studying the impact of various expansions to quantify the uncertainties of the SMEFT matching. Published by the American Physical Society 2024

The features of interlayer fracture of composite materials with a variable layup angle under impact loading

Interlayer defects in structural components made of composite materials are caused by the imperfection of the manufacturing process, complex interactions of the constituents, and the effect of impact loads. The presence of these defects decreases the strength of such components and severely affects residual strength. The paper presents a numerical and experimental study of the behavior of a composite material plate with a variable layup angle under impact loading. Impactor velocities before and after multilayer plate perforation, as well as the dimensions of delamination-type defects, are determined experimentally. The Ansys LS-DYNA software in double precision mode is used to simulate the failure of composite plates under impact loading. It has been found that a significant contribution to the decrease in the impact energy is made by the delamination dimensions depending on the layup angle in the stack. The delamination area is related to the residual velocity of the impactor; namely, the larger the delamination area, the greater the decrease in the impactor velocity.

Acceleration System for Single-Precision Floating-Point Arithmetic Based on IEEE754 Standard

In recent literature, there has been a heightened interest in the algorithms underpinning artificial intelligence. The crux of this research is the integration of these algorithms into embedded systems. Fundamental to these algorithms are arithmetic operations such as addition and multiplication. This study introduces an implementation of single-precision floating-point arithmetic in compliance with the IEEE754 standard. Adopting a modular approach, the system is deconstructed into distinct modules for addition, subtraction, multiplication, and division, each with its own unique implementation. Within an individual module, components such as sign, exponent, and mantissa are distinctly treated. Delving into the mathematical intricacies of the exponent and mantissa allows for the formulation of the subsequent programming. This study proposes a rounding operation to facilitate diverse rounding modes for the data. Advanced techniques such as the over-advancing adder and Zero judgement are employed to enhance the adder's speed, and constructs like Overflow and Error are introduced to evaluate the output data's status in adherence to the IEEE754 standard. Functional tests of the modules were conducted on a designated platform. A timing simulation was performed using ModelSim, corroborating the robust performance of the proposed system.

The Inhibitor: ReLU and Addition-Based Attention for Efficient Transformers (Student Abstract)

To enhance the computational efficiency of quantized Transformers, we replace the dot-product and Softmax-based attention with an alternative mechanism involving addition and ReLU activation only. This side-steps the expansion to double precision often required by matrix multiplication and avoids costly Softmax evaluations but maintains much of the core functionality of conventional dot-product attention. It can enable more efficient execution and support larger quantized Transformer models on resource-constrained hardware or alternative arithmetic systems like homomorphic encryption. Training experiments on four common benchmark tasks show test set prediction scores comparable to those of conventional Transformers with dot-product attention. Our scaling experiments also suggest significant computational savings, both in plaintext and under encryption. In particular, we believe that the ReLU and addition-based attention mechanism introduced in this paper may enable privacy-preserving AI applications operating under homomorphic encryption by avoiding the costly multiplication of encrypted variables.

Abstract 5527: Spatiotemporal analysis of growth factor regulation of osteosarcoma AKT and ERK signaling dynamics in an ex vivo lung metastasis model

Abstract In the present study, we investigated the relationships between spatiotemporal distribution of growth factors in the lung microenvironment, metastatic osteosarcoma cell signaling, and intratumoral gene expression heterogeneity. Osteosarcoma is an aggressive bone malignancy that commonly metastasizes to the lungs, resulting in significant patient mortality. During metastasis, disseminated cancer cells are exposed to microenvironment-derived growth factors that result in signaling pathway activation. Two such pathways are MAPK/ERK and PI3K/AKT, which regulate complex transcriptional programs underlying cell proliferation and survival decisions. In this study, we utilized precision cut lung slices (PCLS) obtained from mice with established osteosarcoma metastases to ex vivo model associations between the spatiotemporal distributions of growth factors in the lung, osteosarcoma signaling dynamics, and intratumoral gene expression heterogeneity, with single cell precision. To accomplish this, we engineered osteosarcoma cells that co-expressed genetically-encoded ERK and AKT sensors, engrafted them into PCLS tissues, then recorded sensor dynamics via live-cell microscopy. These tissues were subsequently fixed and immuno-stained for measurement of growth factor expression in resident lung cells and spatial correlation with observed tumor cell ERK-AKT signaling dynamics. Using this approach, we found that growth factor expression was diffusely increased within areas bordering osteosarcoma metastases. FGF2 and IGF1, which we have previously demonstrated to potently induce ERK signaling in osteosarcoma cells, were both significantly upregulated in these regions. However, despite relatively uniform overexpression of growth factors adjacent to tumors, individual osteosarcoma cells displayed stochastics patterns of ERK and AKT signaling, and downstream gene targets. These findings revealed that upregulation of lung growth factors, and the inherent systems-level properties of osteosarcoma signaling networks, can converge to promote intratumoral signaling and gene expression heterogeneity. Together, our results, using a novel PCLS and live-cell biosensor imaging approach, provide new insight into how spatial and temporal changes in microenvironmental signaling composition are integrated by metastatic osteosarcoma cells to promote heterogeneity. Citation Format: Rawan Makkawi, Ryan D. Roberts, Jeremy Copperman, Alexander E. Davies. Spatiotemporal analysis of growth factor regulation of osteosarcoma AKT and ERK signaling dynamics in an ex vivo lung metastasis model [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5527.

Avoiding breakdown in incomplete factorizations in low precision arithmetic

The emergence of low precision floating-point arithmetic in computer hardware has led to a resurgence of interest in the use of mixed precision numerical linear algebra. For linear systems of equations, there has been renewed enthusiasm for mixed precision variants of iterative refinement. We consider the iterative solution of large sparse systems using incomplete factorization preconditioners. The focus is on the robust computation of such preconditioners in half precision arithmetic and employing them to solve symmetric positive definite systems to higher precision accuracy; however, the proposed ideas can be applied more generally. Even for well-conditioned problems, incomplete factorizations can break down when small entries occur on the diagonal during the factorization. When using half precision arithmetic, overflows are an additional possible source of breakdown. We examine how breakdowns can be avoided and we implement our strategies within new half precision Fortran sparse incomplete Cholesky factorization software. Results are reported for a range of problems from practical applications. These demonstrate that, even for highly ill-conditioned problems, half precision preconditioners can potentially replace double precision preconditioners, although unsurprisingly this may be at the cost of additional iterations of a Krylov solver.

Accelerating SPH-Fatigue Computation by Using Single Precision Program on GPU

The fatigue crack propagation analysis based on the SPH framework “SPH-Fatigue” is accelerated by using the single precision program on GPU. It is found that the single precision program slightly disorders the stress wave but doubles the computational speed. In addition, the computational results of the fatigue crack propagation by the single precision program show almost the same crack shapes as the experimental ones. Accordingly, it is concluded that the use of the single precision program is one of the choices to accelerate the SPH-Fatigue computation.

Security Enabled New Term Weight Measure Technique with Data Driven for Next Generation Mobile Computing Networks

In the field of ASIC and FPGA, Machine Learning (ML) techniques play a major role and become predominant for accurate results for different applications like big data analysis and automotive electronics, and driverless vehicles which are required speed and power savings. Due to increasing the demand for higher accuracy, low power, low area consumption, and higher throughput for the complexity of the designs in the latest technology, the proposed system is fulfilling these demands in ASIC and FPGA domains, reconfigurable hardware architecture has been proposed it consists of an ML-based Support Vector Machine (SVM), high-speed AHB protocol and Floating point (FP) operations and also the system has the flexibility to communicate with I2C and I2S protocols. In order to increase throughput with minimal latency, the proposed architecture with AHB protocol and AHB to APB bridge is incorporated between the fabric dynamically reconfigurable multi-processor (FDPM) and peripherals along with security algorithms using SHA-256bits and AES. In order to perform ML-based applications, the proposed system is incorporated double-precision floating point (DPFP) arithmetic operations. The overall proposed architecture is developed in Verilog HDL and quality checking using the LINT tool and Clock Domain Crossing (CDC) using Spyglass tool and synthesized using DC compiler for ASIC and Vivado Design Suite 2018.1 for FPGA implementation and verification. The entire design is interfaced with the Zynq processor and SDK tool to verify data transfer between hardware and software. The obtained results show the generated custom accelerator is able to compute any complex ML classifiers for a larger amount of data. The obtained results are compared with existing state-of-art results and found that 18 % improvement in throughput, a 21 % improvement in power consumption savings, and a 34 % reduction in latency.

Double Precision is not Necessary for LSQR for Solving Discrete Linear Ill-Posed Problems

Double Precision is not Necessary for LSQR for Solving Discrete Linear Ill-Posed Problems

High capacity double precision image steganography based on chaotic maps

Steganography is the process of hiding confidential information within non-secret multimedia such that the 3rd party cannot distinguish if there is a secret message in it or not. Whereas cryptography is the technique of using mathematical concepts to convert information into unreadable codes via a key. This paper will propose two approaches, lossless and lossy image steganography. Both of them will use cryptography and steganography based on three different chaotic maps to ensure information security. In the cryptography part, two chaotic maps will be used to encrypt the secret information, while in the steganography section, one chaotic map is used to embed the message. The secret information will be concealed in the least significant bits (LSBs) of the double-precision image’s pixels. The double precision image is a high-quality image and can be represented in 64 bits per pixel for grayscale images, leading to a very high redundant bit. Simulation results show a high embedding capacity of 60.938% and 400% for lossless and lossy approaches respectively with a peak signal to noise ratio (PSNR) reach of 69.964 dB. Furthermore, this system is extremely secure due to the use of 3 chaotic maps with key space 2448.

Rationalized diffraction calculations for high accuracy and high speed with few bits.

Diffraction calculations in few-bit formats, such as single-precision floating-point and fixed-point numbers, are important because they yield faster calculations and lower memory usage. However, these methods suffer from low accuracy owing to the loss of trailing digits. Fresnel diffraction is widely known to prevent the loss of trailing digits. However, it can only be used when the paraxial approximation is valid. In this study, a few-bit diffraction calculation method that achieves high accuracy without using any approximation is proposed. The proposed method is derived only by rationalizing the numerator of conventional formulas. Even for scenarios requiring double-precision floating-point numbers using conventional methods, the proposed method exhibits higher accuracy and faster computation time using single-precision floating-point numbers.

Numerical Simulation of Graphene-Nanoplatelet Nanofluid Convection in Millimeter-Sized Automotive Radiator

The depletion of fossil fuels and environmental considerations in transportation sector motivate the researchers to enhance the efficiency and performance of the automotive systems. However, the poor thermal performance of conventional coolant poses a limitation in the development of energy-efficient vehicle due to the cooling constraint. In the present study, a comprehensive numerical study is conducted to scrutinize the convective performance of graphene nanoplatelets (GnP) nanofluid in millimeter-sized automotive radiator, aiming to enhance the understandings on the underlying physical significance of the suspension of graphene-based nanoparticle in water for the performance enhancement of automotive radiator. The temperature-dependent thermophysical properties of GnP-water nanofluid is predicted via existing correlations, while a modified viscosity correlation is developed. ANSYS Fluent is employed in the present numerical simulation to investigate the effects of various pertinent parameters such as Reynolds number, nanoparticles aspect ratio, tube aspect ratio and tube hydraulic diameter on the heat transfer performance of the radiator. Double precision and second-order upwind scheme with inclusion of viscous heating, and convergent criteria of 10˗6 are adopted for the present simulation. It is observed that the convective performance of the radiator is significantly enhanced by increasing Reynolds number and nanoparticle volume fraction while decreasing the aspect ratios of nanoparticle and radiator tube, with an enhancement rate as much as 1816%. Therefore, it is evident that the suspension of GnP intensifies the heat transfer performance of millimeter-sized automotive radiator, which could possibly lead to a more efficient radiator that is smaller and lighter.

Phosphorylation of pyruvate dehydrogenase inversely associates with neuronal activity

For decades, the expression of immediate early genes (IEGs) such as FOS has been the most widely used molecular marker representing neuronal activation. However, to date, there is no equivalent surrogate available for the decrease of neuronal activity. Here, we developed an optogenetic-based biochemical screen in which population neural activities can be controlled by light with single action potential precision, followed by unbiased phosphoproteomic profiling. We identified that the phosphorylation of pyruvate dehydrogenase (pPDH) inversely correlated with the intensity of action potential firing in primary neurons. In invivo mouse models, monoclonal antibody-based pPDH immunostaining detected activity decreases across the brain, which were induced by a wide range of factors including general anesthesia, chemogenetic inhibition, sensory experiences, and natural behaviors. Thus, as an inverse activity marker (IAM) invivo, pPDH can be used together with IEGs or other cell-type markers to profile and identify bi-directional neural dynamics induced by experiences or behaviors.

Nine-Stage Runge–Kutta–Nyström Pairs Sharing Orders Eight and Six

We explore second-order systems of non-stiff initial-value problems (IVPs), particularly those cases where the first derivatives are absent. These types of problems are of significant interest and have applications in various domains, such as astronomy and physics. Runge–Kutta–Nyström (RKN) pairs stand out as highly effective methods of addressing these IVPs. In order to create a pair with eighth and sixth orders, we need to address a certain known set of equations concerning the coefficients. When constructing such pairs for use in double-precision arithmetic, we often need to meet various conditions. Primarily, we aim to maintain small coefficient magnitudes to prevent a loss of accuracy. Nevertheless, in the context of quadruple precision, we can tolerate larger coefficients. This flexibility enables us to establish pairs with eighth and sixth orders that exhibit significantly reduced truncation errors. Traditionally, these pairs are constructed to go through eight stages per step. Here, we propose using nine stages per step. Then we have available more coefficients in order to further reduce truncation errors. As a result, we construct a novel pair that, as anticipated, achieves superior performance compared to equivalent-order pairs in various significant problem scenarios.

Training normalizing flows with computationally intensive target probability distributions

Machine learning techniques, in particular the so-called normalizing flows, are becoming increasingly popular in the context of Monte Carlo simulations as they can effectively approximate target probability distributions. In the case of lattice field theories (LFT) the target distribution is given by the exponential of the action. The common loss function's gradient estimator based on the “reparametrization trick” requires the calculation of the derivative of the action with respect to the fields. This can present a significant computational cost for complicated, non-local actions like e.g. fermionic action in QCD. In this contribution, we propose an estimator for normalizing flows based on the REINFORCE algorithm that avoids this issue. We apply it to two dimensional Schwinger model with Wilson fermions at criticality and show that it is up to ten times faster in terms of the wall-clock time as well as requiring up to 30% less memory than the reparameterization trick estimator. It is also more numerically stable allowing for single precision calculations and the use of half-float tensor cores. We present an in-depth analysis of the origins of those improvements. We believe that these benefits will appear also outside the realm of the LFT, in each case where the target probability distribution is computationally intensive.

Модифицированная формула Герасимова–Капуто

In this work, modified Gerasimov–Caputo formulas were obtained for the first time. The modified formulas take into account the value of the derivative of the function at zero with an order of one less than the order of the derivative under the sign of the Gerasimov–Caputo integral. Without taking into account the new term in the Gerasimov–Caputo formulas, it is not always possible to calculate the fractional derivative on any order interval and for any function. The paper also describes a simple numerical algorithm with the Gaussian quadrature formula, which allows one to calculate the fractional derivative with double precision. tables of fractional derivatives for the sine and cosine functions have been compiled. Moreover, the first half of the tables (in the interval of order (0,1)) and the second half of the tables (in the interval of order (1,2)) were obtained by programs using different algorithms. In the work, an absolute error in calculating the fractional derivative of 10-15 was achieved.

Spin-flip equation-of-motion coupled cluster method with singles, doubles and (full) triples: computational implementation and some pilot applications.

We present our computational implementation of the spin-flip (SF) equation-of-motion (EOM) coupled-cluster (CC) method with singles, doubles, and (full) triples (SDT) within Q-CHEM. The inclusion of triples not only enhances the quantitative accuracy of the SF-EOM-CCSD method but also provides correct qualitative trends in the energy gaps between strongly degenerate states. To assess the accuracy, we compare our SF-EOM-CCSDT results with full configuration interaction (FCI) and complete-active-space self-consistent field second-order (CASSCF-SO) CI benchmarks to study the adiabatic energy gaps in CH2 and NH2+ diradicals, vertical excitation energies in CH radicals and the bond dissociation of the HF molecule. We have implemented SF-EOM-CCSDT using both the conventional double precision (DP) and the single precision (SP) algorithms. The use of SP does not introduce any significant errors in energies and energy gaps, and, due to low cost (relative to DP), turns out to be a promising approach to widen the applicability of EOM-CCSDT to bigger molecules.

Design and experiment of the pneumatic double disc precision seed metering device for Chinese flowering cabbage

Design and experiment of the pneumatic double disc precision seed metering device for Chinese flowering cabbage

Twelve Times Faster yet Accurate: A New State‐Of‐The‐Art in Radiation Schemes via Performance and Spectral Optimization

AbstractRadiation schemes are critical components of Earth system models that need to be both efficient and accurate. Despite the use of approximations such as 1D radiative transfer, radiation can account for a large share of the runtime of expensive climate simulations. Here we seek a new state‐of‐the‐art in speed and accuracy by combining code optimization with improved algorithms. To fully benefit from new spectrally reduced gas optics schemes, we restructure code to avoid short vectorized loops where possible by collapsing the spectral and vertical dimensions. Our main focus is the ecRad radiation scheme, where this requires batching of adjacent cloudy layers, trading some simplicity for improved vectorization and instruction‐level parallelism. When combined with common optimization techniques for serial code and porting widely used two‐stream kernels fully to single precision, we find that ecRad with the TripleClouds solver becomes 12 times faster than the operational radiation scheme in ECMWF's Integrated Forecast System (IFS) cycle 47r3, which uses a less accurate gas optics model (RRMTG) and a more noisy solver (McICA). After applying the spectral reduction and extensive optimizations to the more sophisticated SPARTACUS solver, we find that it’s 2.5 times faster than IFS cy47r3 radiation, making cloud 3D radiative effects affordable to compute in large‐scale models. The code optimization itself gave a threefold speedup for both solvers. While SPARTACUS is still under development, preliminary experiments show slightly improved medium‐range forecasts of 2‐m temperature in the tropics, and in year‐long coupled atmosphere‐ocean simulations the 3D effects warm the surface substantially.

Comparison of two new chaos-based pseudorandom number generators implemented in microcontroller

Chaos-based encryption algorithms uses a pseudorandom number generator (PRNG) to generate chaotic dynamics to carry out the processes to encrypt different types of information. One of the main problems is that some chaotic systems do not present high randomness and uniformity in time series affecting directly the security of the chaos-based cryptosystem. In this work, we proposed two new 2D hyperchaotic maps designed from 1D chaotic maps to use in the design of two new PRNGs for cryptographic applications. The proposed PRNGs were implemented in Renesas RA4M1 32-bit microcontroller for first time in the literature. The chaotic sequences with double-precision floating-point format are also validated with different randomness and security analyses such as the Lyapunov exponent, key space, histograms, autocorrelation, information entropy and the NIST SP800-22 test. In addition, a comparison of the results with similar schemes in the literature about design, security and performance is made. The results achieved in this work can help scholars and designers to improve the effectiveness in randomness of new chaos-based PRNGs for cryptographic applications.