You are invited to attend a lecture

by

Dor Gabay

Ph.D. candidate of

Professor Amir Boag and Dr. Amir Natan of Electrical Engineering, Physical Electronics Department

The solution of the Poisson equation describing the electrostatic potential plays an important role in first principles methods such as Density Functional Theory (DFT), time dependent DFT (TDDFT), Hartree-Fock (HF), GW, and others. In recent years this is becoming even more important as many DFT methods utilize hybrid functionals that involve the inclusion of Fock exchange or screened exchange. The Poisson equation for the electrostatic potential, , is given by , where  is the electronic density. An alternative approach to solving the Poisson Equation is to write it in its equivalent integral form, i.e. , where  is the volumetric domain of integration and  is the electrostatic Green’s function. The volumetric integral can be efficiently evaluated using the Multi-Dimensional Fast Fourier Transform (FFT) which leads to  scaling for  grid points. The FFT-based method was therefore implemented in PARSEC, a real-space first principles Density Functional Theory (DFT) package.

In developing a Poisson solver using FFT, we analyzed both existing and novel Green’s function kernels. The proposed kernels took both analytical and numerical form. The purely analytical kernels were computationally efficient, but tended to produce insufficiently accurate results. Various optimization methods were used to fine-tune the singular region of the analytically derived kernels. The numerical kernels, although more computationally expensive, could be adjusted in accuracy by considering finer integration schemes or higher orders of approximations, making them more accurate than the non-corrected analytically derived kernels.

Finally, we demonstrate how the errors associated with the solution of the electrostatic potential propagate into errors in the eigenvalues of the Density Functional Theory (DFT) calculations.

Monday, July 4th, 2016, at 17:00

Room 011, Kitot Building

~~Speaker: Gadi Oxman
Ph.D. student under the supervision of Prof. Shlomo Weiss

Wednesday, July 6th, 2016 at 15:00
Room 011, Kitot Bldg., Faculty of Engineering

Efficient Communication in Manycore Processors using Deflection Routing

Abstract

Scaling technology limitations shifted the focus in the computer architecture field from complex single threaded cores with ever increasing operating frequencies, to integration of many parallel simpler cores on the same silicon die. A key component that facilitates communication among the cores is the network-on-chip (NoC). The NoC embeds many interconnected routers with the cores in the same chip. Since the router is replicated many times, it is important to improve its performance and power consumption.

A low cost NoC routing scheme that we focus on in this work is deflection routing (also called hot potato routing), originally proposed for off-chip networks. Deflection routing can be bufferless, thus reducing power consumption by eliminating the router's buffers altogether. However, this comes at a network performance cost, especially under heavy load. We propose the use of buffers in deflection routing to improve performance and power efficiency, and show that substantial improvement can be achieved even with a small number of buffers.

Deflection routing is inherently adaptive, and there are multiple routes by which a packet can reach its destination. We propose prioritization schemes that improve performance and reduce network congestion. We also improve deflection routing performance on diagonal, irregular, and hierarchical mesh topologies.

We describe a new cycle based NoC simulation tool that was developed during our research. The simulator can simulate the performance, power and energy of NoCs using deflection routing. It supports both stand-alone simulation using synthetic traffic, as well as co-simulation along with a system simulator, facilitating closed-loop simulation of NoC traffic from application benchmarks.

~~(The talk will be given in English)

Speaker:   Dr. Rami Pugatch
                       Department of Condensed Physics, Tel Aviv University

Monday, June 27th, 2016
15:00 - 16:00
Room 011, Kitot Bldg., Faculty of Engineering

Temporal organization of cellular self-replication

Abstract
What determines the growth rate of a cell? Why single cells grow exponentially?  In this talk I will offer a novel systems analysis perspective on these biological questions. I will explain two generic methods to accelerate a complex assembly process beyond its critical path duration - the duration of the longest serial process that is bound to occur. I will then apply these ideas to the process of cellular self-replication and derive a novel type of growth law relating the doubling rate of a cell with the number of progenies concurrently under production, and the critical path duration.

(No background in biology is required).

Bio:
BA in mathematics and physics - Technion
MsC in physics (Ultracold atoms) - TAU
PhD in Physics (Atomic physics) - Weizmann Institute
5 years in the operation research division - Israeli ministry of Defense as a system analyst
(IS ops-res representative in the US-IS Tactical high energy laser project)
Member and research associate - Institute for advanced study, Princeton NJ USA.
notable prizes:
Israel's president award for excellence (IMOD finalist)
Fulbright fellow 2012.