Description: CUDA matrix multiplication
3 different optimization schemes are included
(loop unrolling, data prefetching, thread granularity) Platform: |
Size: 15360 |
Author:mghs |
Hits:
Description: MPI matrix multiplication
with two versions
(one is broad cast method
and another one is systolic method) Platform: |
Size: 3072 |
Author:mghs |
Hits:
Description: A matrix is a rectangular table of elements,which can be any number or abstract quantities that can be added and multiplied. Matrix multiplication is the operation of multiplying a matrix with a scalar or another matrix, which is simple fuction widely used to solve mathematical question in EECS field. Platform: |
Size: 6388736 |
Author:chen-che,wemg |
Hits:
Description: 基于MPI得并行矩阵乘法 Cannon算法实现 该程序基于 Boost 和 MPI 实现了Cannon 算法实现对矩阵得并行求积。
-MPI-based parallel matrix multiplication may Cannon algorithm implementation of the program and the MPI-based Boost algorithm achieved Cannon was parallel to the matrix of quadrature. Platform: |
Size: 1980416 |
Author:李劲松 |
Hits:
Description: 我大一时候的作业，实现了以下功能:
用16位8086汇编语言编写下列程序之一:
a) 两矩阵相加
b) 求Fibonacci数列
c) 输入一个正整数n, 输出1-n之间的所有素数
d) 矩阵相乘
e) 把一个十进制小数转换成IEEE754格式并输出的程序,
要求从键盘输入一个十进制小数,最后在在屏幕上输出IEEE754格式二进制码的序列.-When my freshman year of operations, realization of the following features:
8086 with 16-bit programs written in assembly language one of the following:
a) sum of two matrices
b) seek Fibonacci series
c) Enter a positive integer n, output 1-n for all prime numbers between
d) matrix multiplication
e) to convert a decimal IEEE754 format and output program
Requests from the keyboard to enter a decimal, the final output in IEEE754 format on the screen binary code sequence. Platform: |
Size: 12288 |
Author:czxu |
Hits:
Description: The main purpose of the core is a color transform tasks such as CIE XYZ<->RGB, different RGB<->RGB and RGB<->YCbCr operations. The main part of color conversions from one to another color system concludes in 3x3 matrix multiplication with vector addition. Therefore the present core performs true 3x3 matrix multiplication and vectors addition of the input data stream.
In the package ccfactors_pkg.vhd presents matrixes for all of the most popular conversions. You can use these defines or type own factors.
Input DATA_WIDTH is changeable and defines by the GENERIC param of the module. Precision of the matrix factors also is changeable. Default precision – 10-E6. Better than enough for color conversion tasks.
Platform: |
Size: 338944 |
Author:Ning |
Hits:
Description: 文章介绍了YUV向RGB颜色空间转换的硬件电路实现算法.在高基乘法算法基础上,建立了参数化高基乘法算法模型,并给出了Verilog HDL描述 小数乘法的整数乘法近似和近似误差给予了详细的讨论.采用乘法单元复用的设计结果将在两个时钟周期内完成YUV向RGB的颜色空间转换.-This paper introduces the YUV to RGB color space conversion hardware algorithm. Matrix multiplication algorithm in high-was established based on a parametric model of the high base multiplication algorithm, and gives the Verilog HDL description decimal multiplication and integer multiplication approximation error of approximation give a detailed discussion. using multiplication unit design reuse results will be completed in two clock cycles YUV to RGB color space conversion. Platform: |
Size: 3997696 |
Author:jjj |
Hits:
Description: this a matrix multiplication program using MPI and c-this is a matrix multiplication program using MPI and c++ Platform: |
Size: 337920 |
Author:mohammad |
Hits:
Description: 矩阵乘法，北航c2的某一次作业题，经测试，能通过所有测试点-Matrix multiplication, an operation of a Northern question c2, tested, pass all the test points Platform: |
Size: 1024 |
Author:火乐月半 |
Hits:
Description: 基于MPI并行程序的任意矩阵矩阵相乘，其中由主程序负责矩阵的输入输出，在本程序中便于测试默认一个矩阵。-MPI parallel programs based on an arbitrary matrix matrix multiplication, which is responsible for the main input and output matrices, in this process easy to test the default of a matrix. Platform: |
Size: 2048 |
Author:Aleny |
Hits:
Description: 主要是用c++语言对存放在文件中的两个矩阵进行乘法运算-It is mainly used c++ language stored in the file of the two matrix multiplication Platform: |
Size: 204800 |
Author:wanghuiqin |
Hits:
Description: 给定n个矩阵{A1,A2,…,An}，其中Ai与Ai+1是可乘的，i=1,2,…,n-1。要算出这n个矩阵的连乘积A1A2…An。由于矩阵乘法满足结合律，故计算矩阵的连乘积可以有许多不同的计算次序。这种计算次序可以用加括号的方式来确定。若一个矩阵连乘积的计算次序完全确定，也就是说该连乘积已完全加括号，则可以依此次序反复调用2个矩阵相乘的标准算法计算出矩阵连乘积。本文的功能是采用动态规划算法，给出矩阵的一种最优的加括号方式，是计算量最小。-Given n matrices {A1, A2, ..., An}, where Ai and Ai+1 is a mere of, i = 1,2, ..., n-1. To calculate the n matrix with the product A1A2 ... An. Since matrix multiplication to meet with law, so the calculation of the matrix with the product can have many different calculation order. This calculation can be bracketed in order to determine ways. Even if the product of a matrix calculation to determine the complete sequence, that the product is fully bracketed with, so you can call the order of 2 repeated the standard matrix multiplication algorithm to calculate the matrix with the product. This feature is the use of dynamic programming algorithm, a matrix is given in parentheses best way is to calculate the minimum amount. Platform: |
Size: 13312 |
Author:阿四 |
Hits:
Description: 该文件内的源代码，实现矩阵乘法的高速算法，效率应该不错的-The file' s source code, to achieve high-speed matrix multiplication algorithm, efficiency should be good Platform: |
Size: 4096 |
Author:张蓓蓓 |
Hits:
Search - matrix multiplication