SOLVERS FOR SYSTEMS OF LARGE SPARSE LINEAR AND NONLINEAR EQUATIONS BASED ON MULTI-GPUS

Numerical treatment of engineering application problems often eventually results in a solution of systems of linear or nonlinear equations.The solution process using digital computational devices usually takes tremendous time due to the extremely large size encountered in most real-world engineering applications.So,practical solvers for systems of linear and nonlinear equations based on multi graphic process units（GPUs）are proposed in order to accelerate the solving process.In the linear and nonlinear solvers,the preconditioned bi-conjugate gradient stable（PBi-CGstab）method and the Inexact Newton method are used to achieve the fast and stable convergence behavior.Multi-GPUs are utilized to obtain more data storage that large size problems need.

Exploiting Parallelism in the Simulation of General Purpose Graphics Processing Unit Program

The simulation is an important means of performance evaluation of the computer architecture. Nowadays, the serial simulation of general purpose graphics processing unit(GPGPU) architecture is the main bottleneck for the simulation speed. To address this issue, we propose the intra-kernel parallelization on a multicore processor and the inter-kernel parallelization on a multiple-machine platform. We apply these two methods to the GPGPU-sim simulator. The intra-kernel parallelization method firstly parallelizes the serial simulation of multiple compute units in one cycle. Then it parallelizes the timing and functional simulation to reduce the performance loss caused by the synchronization between different compute units. The inter-kernel parallelization method divides multiple kernels of a CUDA program into several groups and distributes these groups across multiple simulation hosts to perform the simulation. Experimental results show that the intra-kernel parallelization method achieves a speed-up of up to 12 with a maximum error rate of 0.009 4% on a 32-core machine, and the inter-kernel parallelization method can accelerate the simulation by a factor of up to 3.9 with a maximum error rate of 0.11% on four simulation hosts. The orthogonality between these two methods allows us to combine them together on multiple multi-core hosts to get further performance improvements.

Single-particle 3D reconstruction on specialized stream architecture and comparison with GPGPUs

The wide acceptance and data deluge in medical imaging processing require faster and more efficient systems to be built.Due to the advances in heterogeneous architectures recently,there has been a resurgence in the first research aimed at FPGA-based as well as GPGPU-based accelerator design.This paper quantitatively analyzes the workload,computational intensity and memory performance of a single-particle 3D reconstruction application,called EMAN,and parallelizes it on CUDA GPGPU architectures and decouples the memory operations from the computing flow and orchestrates the thread-data mapping to reduce the overhead of off-chip memory operations.Then it exploits the trend towards FPGA-based accelerator design,which is achieved by offloading computingintensive kernels to dedicated hardware modules.Furthermore,a customized memory subsystem is also designed to facilitate the decoupling and optimization of computing dominated data access patterns.This paper evaluates the proposed accelerator design strategies by comparing it with a parallelized program on a 4-cores CPU.The CUDA version on a GTX480 shows a speedup of about 6 times.The performance of the stream architecture implemented on a Xilinx Virtex LX330 FPGA is justified by the reported speedup of 2.54 times.Meanwhile,measured in terms of power efficiency,the FPGA-based accelerator outperforms a 4-cores CPU and a GTX480 by 7.3 times and 3.4 times,respectively.

A multi-scale architecture for multi-scale simulation and its application to gas-solid flows

A multi-scale hardware and software architecture implementing the EMMS （energy-minimization multi-scale） paradigm is proven to be effective in the simulation of a two-dimensional gas-solid suspension. General purpose CPUs are employed for macro-scale control and optimization, and many integrated cores （MlCs） operating in multiple-instruction multiple-data mode are used for a molecular dynamics simulation of the solid particles at the meso-scale. Many cores operating in single-instruction multiple- data mode, such as general purpose graphics processing units （GPGPUs）, are employed for direct numerical simulation of the fluid flow at the micro-scale using the lattice Boltzmann method. This architecture is also expected to be efficient for the multi-scale simulation of other comolex systems.

Optimizing non-coalesced memory access for irregular applications with GPU computing

General purpose graphics processing units(GPGPUs)can be used to improve computing performance considerably for regular applications.However,irregular memory access exists in many applications,and the benefits of graphics processing units(GPUs)are less substantial for irregular applications.In recent years,several studies have presented some solutions to remove static irregular memory access.However,eliminating dynamic irregular memory access with software remains a serious challenge.A pure software solution without hardware extensions or offline profiling is proposed to eliminate dynamic irregular memory access,especially for indirect memory access.Data reordering and index redirection are suggested to reduce the number of memory transactions,thereby improving the performance of GPU kernels.To improve the efficiency of data reordering,an operation to reorder data is offloaded to a GPU to reduce overhead and thus transfer data.Through concurrently executing the compute unified device architecture(CUDA)streams of data reordering and the data processing kernel,the overhead of data reordering can be reduced.After these optimizations,the volume of memory transactions can be reduced by 16.7%-50%compared with CUSPARSE-based benchmarks,and the performance of irregular kernels can be improved by 9.64%-34.9%using an NVIDIA Tesla P4 GPU.

GPGPU Accelerated Fast Convolution Back-Projection for Radar Image Reconstruction

This paper describes a parallel fast convolution back-projection algorithm design for radar image reconstruction. State-of-the-art general purpose graphic processing units （GPGPU） were utilized to accelerate the processing. The implementation achieves much better performance than conventional processing systems, with a speedup of more than 890 times on NVIDIA Tesla C1060 supercomputing cards compared to an Intel P4 2.4 GHz CPU. 256×256 pixel images could be reconstructed within 6.3 s, which makes real-time imaging possible. Six platforms were tested and compared. The results show that the GPGPU super-computing system has great potential for radar image processing.

Real-time 3D Microtubule Gliding Simulation Accelerated by GPU Computing

A microtubule gliding assay is a biological experiment observing the dynamics of microtubules driven by motor proteins fixed on a glass surface. When appropriate microtubule interactions are set up on gliding assay experiments, microtubules often organize and create higher-level dynamics such as ring and bundle structures. In order to reproduce such higher-level dynamics on computers, we have been focusing on making a real-time 3D microtubule simulation. This real-time 3D microtubule simulation enables us to gain more knowledge on microtubule dynamics and their swarm movements by means of adjusting simulation paranleters in a real-time fashion. One of the technical challenges when creating a real-time 3D simulation is balancing the 3D rendering and the computing performance. Graphics processor unit （GPU） programming plays an essential role in balancing the millions of tasks, and makes this real-time 3D simulation possible. By the use of general-purpose computing on graphics processing units （GPGPU） programming we are able to run the simulation in a massively parallel fashion, even when dealing with more complex interactions between microtubules such as overriding and snuggling. Due to performance being an important factor, a performance n, odel has also been constructed from the analysis of the microtubule simulation and it is consistent with the performance measurements on different GPGPU architectures with regards to the number of cores and clock cycles.