Balancing design options with Sherpa

Abstract

Application specific processors offer the potential of rapidly designed logic specifically constructed to meet the performance and area demands of the task at hand. Recently, there have been several major projects that attempt to automate the process of transforming a predetermined processor configuration into a low level description for fabrication. These projects either leave the specification of the processor to the designer, which can be a significant engineering burden, or handle it in a fully automated fashion, which completely removes the designer from the loop.In this paper we introduce a technique for guiding the design and optimization of application specific processors. The goal of the Sherpa design framework is to automate certain design tasks and provide early feedback to help the designer navigate their way through the architecture design space. Our approach is to decompose the overall problem of choosing an optimal architecture into a set of sub-problems that are, to the first order, independent. For each sub-problem, we create a model that relates performance to area. From this, we build a constraint system that can be solved using integer-linear programming techniques, and arrive at an ideal parameter selection for all architectural components. Our approach only takes a few minutes to explore the design space allowing the designer or compiler to see the potential benefits of optimizations rapidly. We show that the expected performance using our model correlates strongly to detailed pipeline simulations, and present results showing design tradeoffs for several different benchmarks.

Selection of instruction set extensions for an FPGA embedded processor core

“…In architectural exploration, the design engineer uses tools that guide the selection of parameters such as cache size, branch prediction strategy, and number and type of functional units. Some examples of tools that perform this type of processor customization are Sherpa [2] and BUILDABONG [3]. Such approaches allow the engineer to customize the processor to a target application by modifying the micro-architecture of the processor and thereby improve its performance for a given set of assumptions and constraints associated with the application.…”

“…A model for the design space is developed using a data driven analytical model or a simulation. Finally, the design parameters for each architectural feature of the processor are tuned using integer-linear programming in order to optimize the entire processor design [2].…”

“…Motivated by Linear Programming models used to optimize the parameters of ASIPs using architectural exploration in [2], we propose a formal optimization model for the selection of a hybrid instruction set Recall from Figure 2 that the full ISA version of the application can be profiled. The optimization technique proposed in the present section requires that the number cycles required to execute each instruction j in hardware in the full ISA version of the application (j FISA) is known.…”

Automatic application-specific microarchitecture reconfiguration

“…[2] performs analytical (hierarchical) searching of parameters in their own dimensions, with some full parameter exploration to avoid local minimal, for tuning multi-level cache for low-energy embedded systems. [23] explores design options of instruction and data caches, branch predictor, and multiplier, by dividing the search space into piece-wise linear models and solving their results using integer linear programming.…”