A Single (Unified) Shader GPU Microarchitecture for Embedded Systems
Research work supported by the Department of Universities, Research and Society of the Generalitat de Catalunya and the European Social Fund
High Performance Embedded Architectures and Compilers, Lecture Notes in Computer Science, 2005, Volume 3793/2005, 286-301
@article{moya2005single,
title={A Single (Unified) Shader GPU Microarchitecture for Embedded Systems},
author={Moya, V. and Gonzalez, C. and Roca, J. and Fern{‘a}ndez, A. and Espasa, R.},
journal={High Performance Embedded Architectures and Compilers},
pages={286–301},
year={2005},
publisher={Springer}
}
We present and evaluate the TILA-rin GPU microarchitecture for embedded systems using the ATTILA GPU simulation framework. We use a trace from an execution of the Unreal Tournament 2004 PC game to eval uate and compare the performance of the proposed embedded GPU against a baseline GPU architecture for the PC. We evaluate the different elements that have been removed from the baseline GPU architecture to accommodate the architecture to the restricted power, bandwidth and area budgets of em bedded systems. The unified shader architecture we present processes verti ces, triangles and fragments in a single processing unit saving space and re ducing hardware complexity. The proposed embedded GPU architecture sustains 20 frames per second on the selected UT 2004 trace.
December 18, 2010 by hgpu
