A System-C based Micro architectural Exploration Framework for Latency, Power and Performance Trade-offs of On-Chip Interconnection

We describe a System-C based framework we are developing, to explore the impact of various architectural and microarchitectural level parameters of the on-chip interconnection network elements on its power and performance. The framework enables one to choose from a variety of architectural options like topology, routing policy, etc., as well as allows experimentation with various microarchitectural options for the individual links like length, wire width, pitch, pipelining, supply voltage and frequency. The framework also supports a flexible traffic generation and communication model. We provide preliminary results of using this framework to study the power, latency and throughput of a 4x4 multi-core processing array using mesh, torus and folded torus, for two different communication patterns of dense and sparse linear algebra. The traffic consists of both Request-Response messages (mimicing cache accesses)and One-Way messages. We find that the average latency can be reduced by increasing the pipeline depth, as it enables higher link frequencies. We also find that there exists an optimum degree of pipelining which minimizes energy-delay product.

Best Node Selection Through Distributed Fast Variable Power Multiple Access

In many wireless applications, it is highly desirable to have a fast mechanism to resolve or select the packet from the user with the highest priority. Furthermore, individual priorities are often known only locally at the users. In this paper we introduce an extremely fast, local-information-based multiple access algorithm that selects the best node in 1.8 to 2.1 slots,which is much lower than the 2.43 slot average achieved by the best algorithm known to date. The algorithm, which we call Variable Power Multiple Access Selection (VP-MAS), uses the local channel state information from the accessing nodes to the receiver, and maps the priorities into the receive power.It is inherently distributed and scales well with the number of users. We show that mapping onto a discrete set of receive power levels is optimal, and provide a complete characterization for it. The power levels are chosen to exploit packet capture that inherently occurs in a wireless physical layer. The VP-MAS algorithm adjusts the expected number of users that contend in each step and their respective transmission powers, depending on whether previous transmission attempts resulted in capture,idle channel, or collision.

An Adaptive Feature Based Low Power Motion Estimation Algorithm

Motion Estimation is one of the most power hungry operations in video coding. While optimal search (eg. full search)methods give best quality, non optimal methods are often used in order to reduce cost and power. Various algorithms have been used in practice that trade off quality vs. complexity. Global elimination is an algorithm based on pixel averaging to reduce complexity of motion search while keeping performance close to that of full search. We propose an adaptive version of the global elimination algorithm that extracts individual macro-block features using Hadamard transform to optimize the search. Performance achieved is close to the full search method and global elimination. Operational complexity and hence power is reduced by 30% to 45% compared to global elimination method.

A low power, process invariant keeper design for high speed dynamic logic circuits

A low power keeper circuit using the concept of rate sensing has been proposed. The proposed technique reduces the amount of short circuit power dissipation in the domino gate by 70% compared to the conventional keeper technique. Also the total power-delay product is 26% lower compared to the previously reported techniques. The process tracking capability of the design enables the domino gate to achieve uniform delay across different process corners. This reduces the amount of short circuit power dissipation that occurs in the cascaded domino gates by 90%. The use of the proposed technique in the read path of a register file reduces the energy requirement by 26% as compared to the other keeper techniques. The proposed technique has been prototyped in 130nm CMOS technology.

Unified Vdd Vth Optimization Based DVFM Controller for a Logic Block

In this paper analytical expressions for optimal Vdd and Vth to minimize energy for a given speed constraint are derived. These expressions are based on the EKV model for transistors and are valid in both strong inversion and sub threshold regions. The effect of gate leakage on the optimal Vdd and Vth is analyzed. A new gradient based algorithm for controlling Vdd and Vth based on delay and power monitoring results is proposed. A Vdd-Vth controller which uses the algorithm to dynamically control the supply and threshold voltage of a representative logic block (sum of absolute difference computation of an MPEG decoder) is designed. Simulation results using 65 nm predictive technology models are given.

Realizing a flexible constraint length Viterbi decoder for software radio on a de Bruijn interconnection network

Building flexible constraint length Viterbi decoders requires us to be able to realize de Bruijn networks of various sizes on the physically provided interconnection network. This paper considers the case when the physical network is itself a de Bruijn network and presents a scalable technique for realizing any n-node de Bruijn network on an N-node de Bruijn network, where n < N. The technique ensures that the length of the longest path realized on the network is minimized and that each physical connection is utilized to send only one data item, both of which are desirable in order to reduce the hardware complexity of the network and to obtain the best possible performance.

Nonlinear Model Predictive Spread Acceleration Guidance with Impact Angle Constraint for Stationary Targets

A new technique named as model predictive spread acceleration guidance (MPSAG) is proposed in this paper. It combines nonlinear model predictive control and spread acceleration guidance philosophies. This technique is then used to design a nonlinear suboptimal guidance law for a constant speed missile against stationary target with impact angle constraint. MPSAG technique can be applied to a class of nonlinear problems, which leads to a closed form solution of the lateral acceleration (latax) history update. Guidance command assumed is the lateral acceleration (latax), applied normal to the velocity vector. The new guidance law is validated by considering the nonlinear kinematics with both lag-free as well as first order autopilot delay. The simulation results show that the proposed technique is quite promising to come up with a nonlinear guidance law that leads to both very small miss distance as well as the desired impact angle.

Nonlinear Model Predictive Spread Acceleration Guidance with Impact Angle Constraint for Slow-moving Targets

A new technique named as model predictive spread acceleration guidance (MPSAG) is proposed in this paper. It combines nonlinear model predictive control and spread acceleration guidance philosophies. This technique is then used to design a nonlinear suboptimal guidance law for a constant speed missile against stationary target with impact angle constraint. MPSAG technique can be applied to a class of nonlinear problems, which leads to a closed form solution of the lateral acceleration (latax) history update. Guidance command assumed is the lateral acceleration (latax), applied normal to the velocity vector. The new guidance law is validated by considering the nonlinear kinematics with both lag-free as well as first order autopilot delay. The simulation results show that the proposed technique is quite promising to come up with a nonlinear guidance law that leads to both very small miss distance as well as the desired impact angle.

Power Efficient Redundant Execution for Chip Multiprocessor

This paper describes the design of a power efficient microarchitecture for transient fault detection in chip multiprocessors (CMPs) We introduce a new per-core dynamic voltage and frequency scaling (DVFS) algorithm for our architecture that significantly reduces power dissipation for redundant execution with a minimal performance overhead. Using cycle accurate simulation combined with a simple first order power model, we estimate that our architecture reduces dynamic power dissipation in the redundant core by an mean value of 79% and a maximum of 85% with an associated mean performance overhead of only 1:2%