From mgu at math.berkeley.edu Thu Sep 1 17:16:35 2011 From: mgu at math.berkeley.edu (Ming Gu) Date: Thu Sep 1 17:18:16 2011 Subject: [BANANA] LAPACK seminar on Sept. 7, 2011 Message-ID: <201109020016.p820GZFW019231@panda.math.berkeley.edu> Math 290, Section 25, CS 298, Section 6, Fall 2011 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 11:10 - noon in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Sept. 7, 2011 Speaker: Gautam Wilkins, UC San Diego Title: Finding zeros of single-variable real-valued functions Date: Sept. 14, 2011 Speaker: Jiawang Nie, UC San Diego Title: Jacobian SDP Relaxation for Polynomial Optimization From mgu at math.berkeley.edu Mon Sep 5 19:30:26 2011 From: mgu at math.berkeley.edu (Ming Gu) Date: Mon Sep 5 19:32:04 2011 Subject: [BANANA] Reminder: LAPACK seminar on Sept 7, 2011 Message-ID: <201109060230.p862UQJU004740@panda.math.berkeley.edu> Math 290, Section 25, CS 298, Section 6, Fall 2011 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS noon -- 1:00PM in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Sept. 07 Speaker: Gautam Wilkins, UC San Diego Title: Finding zeros of single-variable real-valued functions Abstract: Finding zeros of single-variable, real-valued functions is a common and basic task in scientific computing. Given an interval [a,b], and a single-variable, real-valued function, f, that is continuous on [a,b], if f(a)f(b)<0, then it is guaranteed that f(x)=0 for at least one x \in [a,b]. The task of finding a zero of a function is in general twofold. First, we find an interval, [a,b], across which the function f reverses sign (also called a straddle). Second, we must reduce the size of the interval [a,b] until it is sufficiently small to bracket a zero of f. Given a straddle, we can use Brent's Method, which is one of the most popular methods for finding zeros of functions. This method usually converges very quickly to a zero; for the occasional difficult functions encountered in practice, it typically takes O(n) iterations to converge, where n is the number of steps required for the bisection method to find the zero to approximately the same accuracy. While it has long been known that in theory Brent's Method could require as many as O(n^2) iterations to find a zero, such behavior had never been observed in practice. This presentation will first show that Brent's Method can indeed take O(n^2) iterations to converge, by explicitly constructing such worst case functions, and also show how a simple modification can reduce the worst case complexity to O(n). It will then show how this approach can be generalized to ensure that any zero-finding method that is locally super-linearly convergent can use a straddle to converge in at worst O(n) time. This presentation will also discuss the computational task of finding a straddle for a function, focusing on the method used by Matlab's \textsf{fzero} function, and introduce a new method that is more robust. Date: Sept. 14, 2011 Speaker: Jiawang Nie, UC San Diego Title: Jacobian SDP Relaxation for Polynomial Optimization From mgu at math.berkeley.edu Sat Sep 10 23:37:54 2011 From: mgu at math.berkeley.edu (Ming Gu) Date: Sat Sep 10 23:40:12 2011 Subject: [BANANA] LAPACK seminar on Sept. 14, 2011 Message-ID: <201109110637.p8B6bsFQ002736@panda.math.berkeley.edu> Math 290, Section 25, CS 298, Section 6 Fall 2011 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 12:10 - 1:00PM in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For schedule and other details about the seminar, please check out math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Sept. 14, 2011 Speaker: Jiawang Nie, UC San Diego Title: Jacobian SDP Relaxation for Polynomial Optimization Abstract: Consider the global optimization problem of minimizing a polynomial function subject to polynomial equalities and/or inequalities. Jacobian SDP Relaxation is the first method that can solve this problem globally and exactly by using semidefinite programming. Its basic idea is to use the minors of Jacobian matrix of the given polynomials, add new redundant polynomial equations about the minors to the constraints, and then apply the hierarchy of Lasserre's semidefinite programming relaxations. The main result is that this new semidefinite programming relaxation will be exact for a sufficiently high (but finite) order, that is, the global minimum of the polynomial optimization can be computed by solving a semidefinite programming problem. Date: Sept. 21, 2011 Speaker: Lin Lin, LBNL Title: Numerical algorithms for the electronic structure analysis From mgu at math.berkeley.edu Mon Sep 12 20:24:49 2011 From: mgu at math.berkeley.edu (Ming Gu) Date: Mon Sep 12 20:25:41 2011 Subject: [BANANA] Reminder: LAPACK seminar on Sept. 14, 2011 In-Reply-To: <201109110637.p8B6bsFQ002736@panda.math.berkeley.edu> References: <201109110637.p8B6bsFQ002736@panda.math.berkeley.edu> Message-ID: Math 290, Section 25, CS 298, Section 6 Fall 2011 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 12:10 - 1:00PM in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For schedule and other details about the seminar, please check out math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Sept. 14, 2011 Speaker: Jiawang Nie, UC San Diego Title: Jacobian SDP Relaxation for Polynomial Optimization Date: Sept. 21, 2011 Speaker: Lin Lin, LBNL Title: Numerical algorithms for the electronic structure analysis From mgu at math.berkeley.edu Wed Sep 14 21:29:30 2011 From: mgu at math.berkeley.edu (Ming Gu) Date: Wed Sep 14 21:30:26 2011 Subject: [BANANA] LAPACK seminar on Sept. 21, 2011 Message-ID: <201109150429.p8F4TUrE030305@phoenix.math.berkeley.edu> Math 290, Section 25, CS 298, Section 6 Fall 2011 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 12:10 - 1:00PM in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Sept. 21, 2011 Speaker: Lin Lin, LBNL Title: Numerical algorithms for the electronic structure analysis Abstract: Kohn-Sham density functional theory (KSDFT) is by far the most widely used electronic structure theory for condensed matter systems. The computational time of KSDFT increases as O(N^3) with respect to the number of electrons (N), which hinders its practical application to systems of large size. We have developed an accurate and efficient algorithm to solve KSDFT which is uniformly applicable to insulating and metallic systems. Our method directly uses the property that the electron density and the electron energy are fully characterized by the diagonal elements and the nearest off diagonal elements of the single particle density matrix. This property is not reflected in the current O(N^3) scaling methods. Our new method achieves O(N) scaling for quasi 1D systems, O(N^1.5) scaling for quasi 2D systems and O(N^2) scaling for 3D bulk systems. Date: Sept. 28, 2011 Speaker: Jim Demmel, UCB From mgu at math.berkeley.edu Mon Sep 19 22:38:25 2011 From: mgu at math.berkeley.edu (Ming Gu) Date: Mon Sep 19 22:39:30 2011 Subject: Reminder: [BANANA] LAPACK seminar on Sept. 21, 2011 In-Reply-To: <21992_1316061009_p8F4U523022686_201109150429.p8F4TUrE030305@phoenix.math.berkeley.edu> References: <21992_1316061009_p8F4U523022686_201109150429.p8F4TUrE030305@phoenix.math.berkeley.edu> Message-ID: Math 290, Section 25, CS 298, Section 6 Fall 2011 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 12:10 - 1:00PM in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Sept. 21, 2011 Speaker: Lin Lin, LBNL Title: Numerical algorithms for the electronic structure analysis Date: Sept. 28, 2011 Speaker: Jim Demmel, UCB From mgu at math.berkeley.edu Thu Sep 22 10:05:05 2011 From: mgu at math.berkeley.edu (Ming Gu) Date: Thu Sep 22 10:06:42 2011 Subject: [BANANA] ICME-MATLAB Summit at Stanford (fwd) Message-ID: --- ICME MATLAB Summit invitation --- Hello, The Institute for Computational and Mathematical Engineering at Stanford University is working with MathWorks to host the first ever ICME MATLAB Summit on Tuesday, October 18th, 2011 at Stanford. The goal of the summit is to bring MATLAB users from the scientific community together with MathWorks developers. The summit will feature talks and discussions to form an open dialog about how MATLAB is used in scientific computing and teaching.? We are currently inviting participants from academia, industry, and national labs. It will be a single day event, with talks from both MathWorks and academia. There will be plenty of time between presentations for discussion.? Lunch and dinner will be provided. If you are interested in participating please RSVP at the website: ? http://icme.stanford.edu/matlabsummit/index.html If you have any questions/concerns email us at icme-matlab@lists.stanford.edu Thank you for your time, and we hope to see you there. The ICME MATLAB Summit Team, AJ Friend Arvind Saibaba Nick Henderson Margot Gerritsen icme-matlab@lists.stanford.edu From mgu at math.berkeley.edu Thu Sep 22 10:07:42 2011 From: mgu at math.berkeley.edu (Ming Gu) Date: Thu Sep 22 10:09:23 2011 Subject: [BANANA] LAPACK seminar on Sept. 28, 2011 Message-ID: <201109221707.p8MH7gks028977@panda.math.berkeley.edu> Math 290, Section 25, CS 298, Section 6 Fall 2011 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 12:10 - 1:00PM in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Sept. 28, 2011 Speaker: Jim Demmel Title: More Recent Progress on Communication-Avoiding Algorithms Date: Oct. 5, 2011 Speaker: Ming Gu Title: Low-Rank Matrix Approximations and Randomized Sampling From mgu at math.berkeley.edu Sat Sep 24 20:34:15 2011 From: mgu at math.berkeley.edu (Ming Gu) Date: Sat Sep 24 20:35:51 2011 Subject: [BANANA] LAPACK seminar on Sept. 28. Pls note change of talks Message-ID: <201109250334.p8P3YFBd001616@panda.math.berkeley.edu> Math 290, Section 25, CS 298, Section 6 Fall 2011 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 12:10 - 1:00PM in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. Please note that it has become necessary to make a late change of schedule. Ming Gu will be speaker for Sept. 28, and Jim Demmel speaker for Oct. 5. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Sept. 28, 2011 Speaker: Ming Gu Title: Low-Rank Matrix Approximations and Randomized Sampling Abstract: A classical problem in matrix computations is the efficient and reliable approximation of a given matrix by a lower ranked one, with applications throughout wide areas of computational sciences and engineering. The truncated singular value decomposition (SVD) is known to provide the best such approximation for any given fixed rank. However, the SVD is also known to be very costly to compute. Recently, a number of randomized algorithms for low-rank matrix approximations have attracted researchers' attention due to their surprising reliability and computational efficiency in different application areas. In this talk, we present a novel analysis based on a connection between randomized algorithms and the traditional subspace iteration methods, allowing us to estabilish new and better error bounds for both. We present various numerical experimental results that are in support of our analysis and show that different domain applications lead to different accuracy requirements, and therefore different oversampling sizes. Date: Oct. 5, 2011 Speaker: Jim Demmel Title: More Recent Progress on Communication-Avoiding Algorithms From mgu at math.berkeley.edu Mon Sep 26 18:18:05 2011 From: mgu at math.berkeley.edu (Ming Gu) Date: Mon Sep 26 18:18:59 2011 Subject: [BANANA] Reminder: LAPACK seminar on Sept. 28. In-Reply-To: <201109250334.p8P3YFBd001616@panda.math.berkeley.edu> References: <201109250334.p8P3YFBd001616@panda.math.berkeley.edu> Message-ID: Math 290, Section 25, CS 298, Section 6 Fall 2011 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 12:10 - 1:00PM in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Sept. 28, 2011 Speaker: Ming Gu Title: Low-Rank Matrix Approximations and Randomized Sampling Date: Oct. 5, 2011 Speaker: Jim Demmel Title: More Recent Progress on Communication-Avoiding Algorithms From mgu at math.berkeley.edu Thu Sep 29 23:14:03 2011 From: mgu at math.berkeley.edu (Ming Gu) Date: Thu Sep 29 23:15:46 2011 Subject: [BANANA] LAPACK seminar on Oct. 5, 2011 Message-ID: <201109300614.p8U6E3No013461@panda.math.berkeley.edu> Math 290, Section 25, CS 298, Section 6 Fall 2011 (Matrix Computations and Scientific Computing) We meet WEDNESDAYS 12:10 - 1:00PM in Room 380 Soda Hall, Berkeley campus. The coordinators are Profs. J. Demmel (demmel@cs.berkeley.edu) and M. Gu (mgu@math.berkeley.edu). The program will be a mixture of research talks and tutorials. The tutorials will provide a partial sequel to Math 221. For the schedule and other details about the seminar, please see math.berkeley.edu/~mgu/LAPACKSeminar.htm Date: Oct. 5, 2011 Speaker: Jim Demmel, UC Berkeley Title: More Recent Progress on Communication-Avoiding Algorithms Date: Oct. 12, 2011 Speaker: Michael Mahoney, Stanford U. Title: Linear Algebra and Machine Learning of Large Informatics Graphs