All issues

Although the era of exponential performance growth in computer chips has ended, processor core numbers have reached 16 or more even in general-purpose desktop CPUs. As DRAM throughput is unable to keep pace with this computing power growth, CPU designers need to find ways of lowering memory traffic per instruction. The straightforward way to do this is to reduce the miss rate of the last-level cache. Assuming “non-inclusive cache, inclusive directory” (NCID) scheme already implemented, three ways of reducing the cache miss rate further were studied.

The first is to achieve more uniform usage of cache banks and sets by employing hash-based interleaving and indexing. In the experiments in SPEC CPU2017 refrate tests, even the simplest XOR-based hash functions demonstrated a performance increase of 3.2%, 9.1%, and 8.2% for CPU configurations with 16, 32, and 64 cores and last-level cache banks, comparable to the results of more complex matrix-, division- and CRC-based functions.

The second optimisation is aimed at reducing replication at different cache levels by means of automatically switching to the exclusive scheme when it appears optimal. A known scheme of this type, FLEXclusion, was modified for use in NCID caches and showed an average performance gain of 3.8%, 5.4 %, and 7.9% for 16-, 32-, and 64-core configurations.

The third optimisation is to increase the effective cache capacity using compression. The compression rate of the inexpensive and fast BDI*-HL (Base-Delta-Immediate Modified, Half-Line) algorithm, designed for NCID, was measured, and the respective increase in cache capacity yielded roughly 1% of the average performance increase.

All three optimisations can be combined and demonstrated a performance gain of 7.7%, 16% and 19% for CPU configurations with 16, 32, and 64 cores and banks, respectively.

A promising method of increasing precipitation in arid climates is the method of creating a vertical high-temperature jet seeded by hygroscopic aerosol. Such an installation makes it possible to create artificial clouds with the possibility of precipitation formation in a cloudless atmosphere, unlike traditional methods of artificial precipitation enhancement, which provide for increasing the efficiency of precipitation formation only in natural clouds by seeding them with nuclei of crystallization and condensation. To increase the power of the jet, calcium chloride, carbamide, salt in the form of a coarse aerosol, as well as NaCl/TiO₂ core/shell novel nanopowder, which is capable of condensing much more water vapor than the listed types of aerosols, are added. Dispersed inclusions in the jet are also centers of crystallization and condensation in the created cloud to increase the possibility of precipitation. To simulate convective flows in the atmosphere, a mathematical model of FlowVision large-scale atmospheric flows is used, the solution of the equations of motion, energy and mass transfer is carried out in relative variables. The statement of the problem is divided into two parts: the initial jet model and the FlowVision large-scale atmospheric model. The lower region, where the initial high-speed jet flows, is calculated using a compressible formulation with the solution of the energy equation with respect to the total enthalpy. This division of the problem into two separate subdomains is necessary in order to correctly carry out the numerical calculation of the initial turbulent jet at high velocity (M > 0.3). The main mathematical dependencies of the model are given. Numerical experiments were carried out using the presented model, experimental data from field tests of the installation for creating artificial clouds were taken for the initial data. A good agreement with the experiment is obtained: in 55% of the calculations carried out, the value of the vertical velocity at a height of 400 m (more than 2 m/s) and the height of the jet rise (more than 600 m) is within an deviation of 30% of the experimental characteristics, and in 30% of the calculations it is completely consistent with the experiment. The results of numerical simulation allow evaluating the possibility of using the high-speed jet method to stimulate artificial updrafts and to create precipitation. The calculations were carried out using FlowVision CFD software on SUSU Tornado supercomputer.

The paper gives a detailed description of the developed methods for simulating the turbulent flow around a helicopter rotor and calculating its aerodynamic characteristics. The system of Reynolds-averaged Navier – Stokes equations for a viscous compressible gas closed by the Spalart –Allmaras turbulence model is used as the basic mathematical model. The model is formulated in a non-inertial rotating coordinate system associated with a rotor. To set the boundary conditions on the surface of the rotor, wall functions are used.

The numerical solution of the resulting system of differential equations is carried out on mixed-element unstructured grids including prismatic layers near the surface of a streamlined body.The numerical method is based on the original vertex-centered finite-volume EBR schemes. A feature of these schemes is their higher accuracy which is achieved through the use of edge-based reconstruction of variables on extended quasi-onedimensional stencils, and a moderate computational cost which allows for serial computations. The methods of Roe and Lax – Friedrichs are used as approximate Riemann solvers. The Roe method is corrected in the case of low Mach flows. When dealing with discontinuities or solutions with large gradients, a quasi-one-dimensional WENO scheme or local switching to a quasi-one-dimensional TVD-type reconstruction is used. The time integration is carried out according to the implicit three-layer second-order scheme with Newton linearization of the system of difference equations. To solve the system of linear equations, the stabilized conjugate gradient method is used.

The numerical methods are implemented as a part of the in-house code NOISEtte according to the two-level MPI–OpenMP parallel model, which allows high-performance computations on meshes consisting of hundreds of millions of nodes, while involving hundreds of thousands of CPU cores of modern supercomputers.

Based on the results of numerical simulation, the aerodynamic characteristics of the helicopter rotor are calculated, namely, trust, torque and their dimensionless coefficients.

Validation of the developed technique is carried out by simulating the turbulent flow around the Caradonna – Tung two-blade rotor and the KNRTU-KAI four-blade model rotor in hover mode mode, tail rotor in duct, and rigid main rotor in oblique flow. The numerical results are compared with the available experimental data.

A large number of methods have been developed to solve the Navier – Stokes equations in the case of incompressible flows, the most popular of which are methods with velocity correction by the SIMPLE algorithm and its analogue — the method of splitting by physical variables. These methods, developed more than 40 years ago, were used to solve rather simple problems — simulating both stationary flows and non-stationary flows, in which the boundaries of the calculation domain were stationary. At present, the problems of computational fluid dynamics have become significantly more complicated. CFD problems are involving the motion of bodies in the computational domain, the motion of contact boundaries, cavitation and tasks with dynamic local adaptation of the computational mesh. In this case the computational mesh changes resulting in violation of the velocity divergence condition on it. Since divergent velocities are used not only for Navier – Stokes equations, but also for all other equations of the mathematical model of fluid motion — turbulence, mass transfer and energy conservation models, violation of this condition leads to numerical errors and, often, to undivergence of the computational algorithm.

This article presents an implicit method of splitting by physical variables that uses divergent velocities from a given time step to solve the incompressible Navier – Stokes equations. The method is developed to simulate flows in the case of movable and contact boundaries treated in the Euler paradigm. The method allows to perform computations with the integration step exceeding the explicit time step by orders of magnitude (Courant – Friedrichs – Levy number $CFL\gg1$). This article presents a variant of the method for incompressible flows. A variant of the method that allows to calculate the motion of liquid and gas at any Mach numbers will be published shortly. The method for fully compressible flows is implemented in the software package FlowVision.

Numerical simulating classical fluid flow around circular cylinder at low Reynolds numbers ($50 < Re < 140$), when laminar flow is unsteady and the Karman vortex street is formed, are presented in the article. Good agreement of calculations with the experimental data published in the classical works of Van Dyke and Taneda is demonstrated.

The paper presents the results of applying a scheme of very high accuracy and resolution to obtain numerical solutions of the Navier – Stokes equations of a compressible gas describing the occurrence and development of instability of a two-dimensional laminar boundary layer on a flat plate. The peculiarity of the conducted studies is the absence of commonly used artificial exciters of instability in the implementation of direct numerical modeling. The multioperator scheme used made it possible to observe the subtle effects of the birth of unstable modes and the complex nature of their development caused presumably by its small approximation errors. A brief description of the scheme design and its main properties is given. The formulation of the problem and the method of obtaining initial data are described, which makes it possible to observe the established non-stationary regime fairly quickly. A technique is given that allows detecting flow fluctuations with amplitudes many orders of magnitude smaller than its average values. A time-dependent picture of the appearance of packets of Tollmien – Schlichting waves with varying intensity in the vicinity of the leading edge of the plate and their downstream propagation is presented. The presented amplitude spectra with expanding peak values in the downstream regions indicate the excitation of new unstable modes other than those occurring in the vicinity of the leading edge. The analysis of the evolution of instability waves in time and space showed agreement with the main conclusions of the linear theory. The numerical solutions obtained seem to describe for the first time the complete scenario of the possible development of Tollmien – Schlichting instability, which often plays an essential role at the initial stage of the laminar-turbulent transition. They open up the possibilities of full-scale numerical modeling of this process, which is extremely important for practice, with a similar study of the spatial boundary layer.