加速比

PVM环境下提高并行计算加速比的数值实验研究

Numerical research of raising speedup of parallel calculation on PVM

MC程序并行设计及提高加速比措施

The parallel design of Monte Carlo code and measures of enhance speedup

在12个cpu的PC机群上，该算法得到了线性加速比。

Experiments on 12 CPUs cluster that has testify that the algorithm can achieve linear speedups .

三维电磁散射的网络并行FDTD计算和加速比分析

Parallel FDTD computing for 3D EM scattering problem and its speedup factor analysis

应用层并行可以在多核或CPU的计算机上达到接近CPU内核数的加速比。

The speedup of parallel can close to the number of CPU cores in the multi-core or multi-CPU computer .

经实验验证，本文GPUWM算法较原算法能够取得较好的加速比。

By the experimental result , GPU_WM algorithm gains a better speedup than the original algorithm .

然后提出的并行QR分解算法，具有通信开销小，加速比大的特点。

Then we present the parallel QR decomposition , which has less communication speeding and more speedup ratio .

在IBMSP2机上，当书点数为3时，测得的加速比为2.94，效率达到0.98。

When the node of IBM SP2 computer is 3 , the speed-up ratio will be 2.94 , and the parallel efficiency 0.98 .

在Linux集群服务器上进行了两种不同规模三对角方程组的实现，并研究了不同节点数目与计算规模下的计算时间和加速比。

Two triangular equation systems with different matrix size were realized on the Linux clusters , then the calculation time and acceleration ratio were studied under different nodes and matrix scales .

最后，通过计算三维复杂目标金属机翼目标的RCS，给出并行加速比和并行效率的测试结果。

The radar cross sections ( RCS ) of complex objects of metallic wing and NASA almond and practical object of missile are presented .

实验结果表明，SVM具有较好的扩展性，在具有不同系统配置的多核处理器上达到了接近线性的加速比，并保持了较高的负载均衡率和较低的通信开销。

Experimental results demonstrate that the SVM is scalable with processor cores and nearly reaches linear speedup on processors while keeping high load balance efficiency and low communication overhead .

利用优化加速比分析了NAS基准测试程序MG和FT在IBMSP2（66mhz/wn）上的性能。

It provides a fair performance measurement . Using optimized-speedup , we analyze the parallel performance of the NAS parallel benchmark MG and FT on IBM SP2 .

本文讨论超立方体结构上的并行FFT算法，着重分析算法的通信复杂性，并在此基础上导出算法的加速比。

This paper discusses the parallel FFT algorithm on the hypercube architecture . Based on the analysis of the communication complexity , the speedup of the algorithm is derived .

论文基于算法复杂度和加速比等评价标准对所提算法进行评价，并通过实验与传统K-Means算法做了比较。

Then some evaluation criteria like algorithm complexity and speed up has been given . Some experiments compared with the traditional K-Means algorithm have been done .

在PC集群上的计算结果表明，该算法有比较高的并行加速比和计算效率，在无任何加速措施的情况下，对于较大规模的系统已达到了在线实时计算速度。

The result of the practical sample on the PC-Cluster system demonstrates that the new algorithm possesses high calculation efficiency and parallel speedup ratio , and can achieve on-line calculation for large-scale systems without any speedup measure .

在模拟环境下的测试结果表明，应用以上确定调度模型的关键参数的方法，FIR滤波器的并行算法的加速比和效率较高，与理论分析的结果吻合。

The result of simulation experiment shows that the speedup ratio and efficiency of the parallel algorithm for the FIR filter are high and they are in accord with the theoretical analysis .

实验结果表明：该系统在16节点（32个CPU）集群服务器上的加速比最高可达30。

Experimental results show that in this scheme the AVS video encoding time can be significantly reduced . On a 16 nodes cluster with 32 CPUs , the implementation speedup ratio can reach 30 .

本文针对ALPHA中的结构特点，利用反馈式编译优化技术，提高了ALPHA中的转移预测命中率，实验结果表明，加速比效果较为明显。

This passage adopts this technique according to the features of ALPHA structure and improves the correction of branch prediction schemes . The experiment result shows that the effect of this optimization is evident .

在一个MPP并行机的500个CPU上，模拟包含21×108个粒子的三维金属微喷射模型，该算法获得了420倍的加速比。

Parallel numerical experiments for simulations of three-dimensional metal-eject physical models with 210 million particles using 500 processors of one massively parallel processors have shown that this algorithm has achieved speedup about 420 .

我们在四个处理器的晶片计算机（Transputer系统）上实现了这两种算法，取得了相对于处理器个数几乎线性的加速比。

Them has been implemented on a parallel Transputer system with four processors and a nearly linear speedup ratio has been . achieved .

设计、实现了基于MPI的三维叠前合成震源记录深度偏移的并行算法，算法并行效率和加速比理想。

Prestack depth migration method of synthetic source records retains accuracy of shot profile migration and overcomes high computational cost . The realization of parallel programming with MPI speeds up the computations .

通过测试和分析CCG（ComponentbasedComputationalGrid）系统，表明CCG系统具有较好的加速比。

At last , it analyzes the good performance and speedup of CCG ( Component based Computational Grid ) system .

实验结果表明，实现的并行算法执行性能得到了显著提升，取得了良好的并行加速比。4.研究并实现了遥感图像GPU高性能处理算法程序面向存储级的优化策略。

The experiment results show that the implementation of parallel algorithms had significantly improved performance , and achieved good parallel speedup . 4 . The optimization strategies of GPU memory level are studied .

本文针对MIMD并行机对一般的Job－Shop调度提出实时调度的并行算法，通过分析复杂性和加速比以及实例，说明并行算法对求大批工件多台机器加工的最优调度的优越性。

N this paper , we point out a parallel algorithm for real-time scheduling of multi-machine Job-shop in MIMD parallel computers . We show the advantage of this parallel algorithm by analyzing the complexity , the speedup and an example .

通过对对象存储系统的加速比进行分析得出结论：Cache对存储系统的性能改善不仅与Cache命中率有关，还与数据对象的设备访问时间有关。

Some conclusions can be reached by analyzing the speedup ratio of the storage system : the factors to influence the performance of storage system include not only Cache hit ratio but also the time of data object to access storage devices .

结果表明：在数值预报中采用并行计算技术后，其CPU加速比达到3.21，并行效率达80%，模式并行计算版本运行基本稳定，计算结果与串行计算结果基本一致。

From June 1996 , parallel arithmetic weather model is used and the result shows that the CPU accelerating ratio is 3.21 , the parallel efficiency is 80 % , the circulate is stable and the outcome is consistent with serial arithmetic essentially .

对二维倒台阶流并行计算进行了性能分析，表明基于EBE策略的迭代解法能获得较高的加速比。

By analyzing the parallel compute performance of 2D upend-step flow calculation , it is proved that EBE iterative methods can achieve upper speedup rations .

对二维高超声速MHD流场进行了并行计算，得到了较好的加速比与并行计算效率，这也表明了在非结构网格上发展多维MHD方程高效并行算法是切实可行的。

A 2-D hypersonic MHD flow was numerically simulated using the parallel algorithm , and high parallel efficiency was achieved , which shows the feasibility of developing high efficient parallel algorithms for multidimensional MHD equations .

实验结果表明最终的优化程序能在保证图像质量的前提下，获得对CPU一个的数量级的加速比，较成功的在GPU上实现了高速的小波变换。

Experimental results showed that under the premise of the quality of image can be guaranteed , the optimization process can obtain speedup of an order of magnitude to CPU . The high-speed wavelet transform is achieved successfully on GPU .

测试和评价了Fluent软件并行性能，结果表明Fluent在高性能计算机集群的硬件条件下具有更好的并行性能，其16个CPU的加速比达到了11。

The speedup performance of Fluent in the parallel computing is tested and evaluated , and the results show that Fluent have the better parallel performance in the High Performance Parallel Cluster and its speedup rate arrives at 10 under 16 CPUs .