This paper describes the modeling and performance analysis of MS accelerators integrating SRAM and F-2T2R compute cells, carried out with a Matlab analytical framework. The accelerator models include the fundamental circuit non-idealities for accurate analysis. The framework allows exploration of the design space and optimization of MS accelerators for the best system-level performance. The comparison made in a 22-nm technology shows that SRAM-based accelerators can obtain significantly higher computational efficiency, while the high density of the RRAM devices allows for better area efficiency in accelerators embedding the F-2T2R cells.
Modeling and Analysis of MS Accelerators Embedding SRAM and RRAM Compute Cells / Caselli, Michele; Boni, Andrea. - (2025). ( 2025 21st International Conference on Synthesis, Modeling, Analysis and Simulation Methods, and Applications to Circuits Design (SMACD)) [10.1109/SMACD65553.2025.11092257].
Modeling and Analysis of MS Accelerators Embedding SRAM and RRAM Compute Cells
Michele Caselli
;Andrea Boni
2025-01-01
Abstract
This paper describes the modeling and performance analysis of MS accelerators integrating SRAM and F-2T2R compute cells, carried out with a Matlab analytical framework. The accelerator models include the fundamental circuit non-idealities for accurate analysis. The framework allows exploration of the design space and optimization of MS accelerators for the best system-level performance. The comparison made in a 22-nm technology shows that SRAM-based accelerators can obtain significantly higher computational efficiency, while the high density of the RRAM devices allows for better area efficiency in accelerators embedding the F-2T2R cells.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


