A resource efficient hardware accelerator for Bayesian neural network (BNN) named B2N2, Bernoulli random number based accelerator, is proposed. As networks expand their application into risk sensitive domains where mispredictions may cause serious social and economic losses, evaluating the NN’s confidence on its prediction has emerged as a critical concern. Among many uncertainty evaluation met...

in this paper, by numerical simulation of gas flow, effects of divergence angles on the amplification factor, side force and thrust vector magnitude of a moveable nozzle with supersonic split line have been investigated. the study is conducted for constant exit to throat area ratio and the location of the split line is also constant. the finite volume method with second order upwind scheme for ...

In numerous industries such as drilling, peening, cleaning, etc., a cavitating jet is adopted. However, it challenging to simulate the flow field numerically with accuracy. The of organ pipe cavitation nozzle simulated in this research using RNG k−ε, DES, and LES turbulence models. model can more accurately predict periodic shedding cloud, which basically consistent morphology captured high−spe...

Abstract A computational study of a supersonic flow in the base region and nearest wake cylindrical body moving at speed has been carried out. In this case, “prehistory” was taken into account, i.e. configuration domain as close possible to real one. The use various RANS turbulence models for calculating analyzed. following were considered: 1) SST model; 2) Standard k– ε 3) RNG 4) Reynolds Stre...

To investigate the different influencing characteristics of local scour around submarine pipelines, hydrodynamic and sediment transport two-dimensional models based on Flow-3D are used to numerically simulate pipeline under steady currents. An RNG k-ε turbulence model is applied turbulent flow field pipeline. The instantaneous shear stress bed surface taken as starting transporting conditions s...

Mersenne Twister uniform random number generators are key cores for hardware acceleration of Monte Carlo simulations. In this work, two different architectures are studied: besides the classical tablebased architecture, a new architecture based on a circular buffer and especially targeting FPGAs is proposed. A 30% performance improvement has been obtained when compared to the fastest previous w...

By means of a simple OMNeT++ simulation scenario, which uses parallel streams of random numbers, we want to point out several technical pitfalls when applying different random number generators and using different initialization values for them. We describe shortcomings of the built-in random number generator (RNG) and describe traps when using modern RNGs. Quantitative and qualitative analyses...

This paper analyses the effect of using different random number generators (RNG) in a hardware implementation of Genetic Programming using Field Programmable Gate Arrays. Hardware systems have typically used RNGs based on Logical Feedback Shift Registers or Cellular Automata. Different configurations of these generators are evaluated as well as using a source of true random numbers and a standa...

Random number generation (RNG) is needed for a myriad of applications ranging from secure communication encryption to numerical simulations sports and games. However, generating truly random numbers can be elusive. Pseudorandom bit using computer algorithms provides high rate. Nevertheless, the reliance on predefined makes it deterministic predictable once initial conditions are known. Relying ...