CIM-Tuner: Balancing the Compute and Storage Capacity of SRAM-CIM Accelerator via Hardware-mapping Co-exploration

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Chen, Jinwu, Shi, Yuhui, Wang, He, Jiang, Zhe, Yang, Jun, Si, Xin, Zhu, Zhenhua
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915755454365696
author Chen, Jinwu
Shi, Yuhui
Wang, He
Jiang, Zhe
Yang, Jun
Si, Xin
Zhu, Zhenhua
author_facet Chen, Jinwu
Shi, Yuhui
Wang, He
Jiang, Zhe
Yang, Jun
Si, Xin
Zhu, Zhenhua
contents As an emerging type of AI computing accelerator, SRAM Computing-In-Memory (CIM) accelerators feature high energy efficiency and throughput. However, various CIM designs and under-explored mapping strategies impede the full exploration of compute and storage balancing in SRAM-CIM accelerator, potentially leading to significant performance degradation. To address this issue, we propose CIM-Tuner, an automatic tool for hardware balancing and optimal mapping strategy under area constraint via hardware-mapping co-exploration. It ensures universality across various CIM designs through a matrix abstraction of CIM macros and a generalized accelerator template. For efficient mapping with different hardware configurations, it employs fine-grained two-level strategies comprising accelerator-level scheduling and macro-level tiling. Compared to prior CIM mapping, CIM-Tuner's extended strategy space achieves 1.58$\times$ higher energy efficiency and 2.11$\times$ higher throughput. Applied to SOTA CIM accelerators with identical area budget, CIM-Tuner also delivers comparable improvements. The simulation accuracy is silicon-verified and CIM-Tuner tool is open-sourced at https://github.com/champloo2878/CIM-Tuner.git.
format Preprint
id arxiv_https___arxiv_org_abs_2601_18070
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle CIM-Tuner: Balancing the Compute and Storage Capacity of SRAM-CIM Accelerator via Hardware-mapping Co-exploration
Chen, Jinwu
Shi, Yuhui
Wang, He
Jiang, Zhe
Yang, Jun
Si, Xin
Zhu, Zhenhua
Hardware Architecture
As an emerging type of AI computing accelerator, SRAM Computing-In-Memory (CIM) accelerators feature high energy efficiency and throughput. However, various CIM designs and under-explored mapping strategies impede the full exploration of compute and storage balancing in SRAM-CIM accelerator, potentially leading to significant performance degradation. To address this issue, we propose CIM-Tuner, an automatic tool for hardware balancing and optimal mapping strategy under area constraint via hardware-mapping co-exploration. It ensures universality across various CIM designs through a matrix abstraction of CIM macros and a generalized accelerator template. For efficient mapping with different hardware configurations, it employs fine-grained two-level strategies comprising accelerator-level scheduling and macro-level tiling. Compared to prior CIM mapping, CIM-Tuner's extended strategy space achieves 1.58$\times$ higher energy efficiency and 2.11$\times$ higher throughput. Applied to SOTA CIM accelerators with identical area budget, CIM-Tuner also delivers comparable improvements. The simulation accuracy is silicon-verified and CIM-Tuner tool is open-sourced at https://github.com/champloo2878/CIM-Tuner.git.
title CIM-Tuner: Balancing the Compute and Storage Capacity of SRAM-CIM Accelerator via Hardware-mapping Co-exploration
topic Hardware Architecture
url https://arxiv.org/abs/2601.18070