Annotation of OpenXM/doc/calc2000p/efficient.tex, Revision 1.2
1.2 ! noro 1: % $OpenXM: OpenXM/doc/calc2000p/efficient.tex,v 1.1 2000/07/21 08:27:11 noro Exp $
1.1 noro 2: \documentclass{slides}
3: \usepackage{color}
1.2 ! noro 4: \usepackage{rgb}
1.1 noro 5: \usepackage{graphicx}
6: \usepackage{epsfig}
7: \begin{document}
8: \parskip 3pt
9: \fbox{{\huge \color{blue} We Are Efficient}}
10:
1.2 ! noro 11: {\color {red} Real speedup by parallelism}
1.1 noro 12:
1.2 ! noro 13: Each process can communicate with any other process on {\color{green} MPI}
1.1 noro 14:
1.2 ! noro 15: $\Rightarrow$ {\color{green} Collective operations} similar to \\
! 16: {\color{SeaGreen} MPI\_Bcast} and {\color{SeaGreen} MPI\_Reduce}
1.1 noro 17:
1.2 ! noro 18: \vskip 8pt
1.1 noro 19:
20: {\color{red} An example} :
1.2 ! noro 21: the product of dense univariate polynomials with 3000bit coefficients
1.1 noro 22:
1.2 ! noro 23: \vskip 8pt
! 24: {\color{green} Algorithm}(Shoup) : FFT+Chinese Remainder
! 25: \vskip 5pt
! 26: {\color{green} Parallel computing time} : $O((n\log n)/L+n{\color{red}\log L})$\\
! 27: No collective operations : $O((n\log n)/L+n{\color{red}L})$\\
! 28: ($L$: number of processes, $n$: degree)
1.1 noro 29:
1.2 ! noro 30: \vskip 8pt
1.1 noro 31:
32: \epsfxsize=17cm
33: \epsffile{3k.ps}
1.2 ! noro 34: \rightline{ {\color{red} {\tt http://www.openxm.org} }}
1.1 noro 35: \end{document}
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>