Abstract
With Moore’s Law slowing down, the development of SoCs has encountered a bottleneck. Integrating more functional units and larger on-chip storage leads to a dramatic increase in chip area, resulting in lower chip yields and higher costs. Most researches and industry products began to seek to use advanced connection and packaging technologies to decompose the raw chip into multiple smaller, higher yield, and more cost-effective chiplets, and then packet them. Interposer-based 2.5D integration, as an emerging packaging technology, is widely used in chiplet-based systems. However, even if both the interposer and chiplets are deadlock-free, deadlock dependency cycles across them may still occur after integration. To address these problems, this paper proposes a deadlock resolution called Absorb for 2.5D integrated chiplet systems, which is different from deadlock avoidance and deadlock recovery. By regularly absorbing inter-chiplet packets, global deadlock freedom is achieved, and no extra VCs are for deadlock resolution. Our proposed Absorb maintains the modularity of each chiplet and imposes no restrictions on the routing algorithm. Our evaluations show that compared with the previously proposed deadlock-free designs in 2.5D-chiplet systems, Absorb provides an average performance improvement of about 7.5%, and the area overhead is less than 6%.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Agarwal, N., Krishna, T., Peh, L.S., et al.: GARNET: a detailed on-chip network model inside a full-system simulator. In: 2009 IEEE International Symposium on Performance Analysis of Systems and Software, pp. 33–42. IEEE (2009)
Binkert, N., Beckmann, B., Black, G., et al.: The gem5 simulator. ACM SIGARCH Comput. Archit. News 39(2), 1–7 (2011)
Dally, W.J., Seitz, C.L.: Deadlock-free message routing in multiprocessor interconnection networks. IEEE Trans. Comput. C-36(5), 547–553 (1988)
Duato, J.: A new theory of deadlock-free adaptive routing in wormhole networks. IEEE Trans. Parallel Distrib. Syst. 4(12), 1320–1331 (1993)
Duato, J., Pinkston, T.M.: A general theory for deadlock-free adaptive routing using a mixed set of resources. IEEE Trans. Parallel Distrib. Syst. 12(12), 1219–1235 (2001)
Fallin, C., Craik, C., Mutlu, O.: CHIPPER: a low-complexity bufferless deflection router. In: 2011 IEEE 17th International Symposium on High Performance Computer Architecture, pp. 144–155. IEEE (2011)
Ramrakhyani, A., Krishna, T.: Static bubble: a framework for deadlock-free irregular on-chip topologies. In: 2017 IEEE International Symposium on High Performance Computer Architecture (HPCA), pp. 253–264. IEEE (2017)
Chen, L., Wang, R., Pinkston, T.M.: Critical bubble scheme: an efficient implementation of globally aware network flow control. In: 2011 IEEE International Parallel & Distributed Processing Symposium, pp. 592–603. IEEE (2011)
Ramrakhyani, A., Gratz, P.V., Krishna, T.: Synchronized progress in interconnection networks (SPIN): a new theory for deadlock freedom. In: 2018 ACM/IEEE 45th Annual International Symposium on Computer Architecture (ISCA), pp. 699–711. IEEE (2018)
Parasar, M., Jerger, N.E., Gratz, P.V., et al.: SWAP: synchronized weaving of adjacent packets for network deadlock resolution. In: Proceedings of the 52nd Annual IEEE/ACM International Symposium on Microarchitecture, pp. 873–885 (2019)
Parasar, M., Farrokhbakht, H., Jerger, N.E., et al.: DRAIN: deadlock removal for arbitrary irregular networks. In: 2020 IEEE International Symposium on High Performance Computer Architecture (HPCA), pp. 447–460. IEEE (2020)
Yin, J., Lin, Z., Kayiran, O., et al.: Modular routing design for chiplet-based systems. In: 2018 ACM/IEEE 45th Annual International Symposium on Computer Architecture (ISCA), pp. 726–738. IEEE (2018)
Majumder, P., Kim, S., Huang, J., et al.: Remote control: a simple deadlock avoidance scheme for modular systems-on-chip. IEEE Trans. Comput. 70(11), 1928–1941 (2020)
Taheri, E., Pasricha, S., Nikdast, M.: DeFT: a deadlock-free and fault-tolerant routing algorithm for 2.5 D Chiplet Networks. In: 2022 Design, Automation & Test in Europe Conference & Exhibition (DATE), pp. 1047–1052. IEEE (2022)
Wu, Y., Wang, L., Wang, X., et al.: Upward packet popup for deadlock freedom in modular chiplet-based systems. In: 2022 IEEE International Symposium on High-Performance Computer Architecture (HPCA), pp. 986–1000. IEEE (2022)
Zhan, X., Bao, Y., Bienia, C., et al.: PARSEC3. 0: a multicore benchmark suite with network stacks and SPLASH-2X. ACM SIGARCH Comput. Archit. News 44(5), 1–16 (2017)
Author information
Authors and Affiliations
Corresponding authors
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2024 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Yang, Y., Li, T., Dai, Y., Wang, B., Ma, S., Sun, Y. (2024). Absorb: Deadlock Resolution for 2.5D Modular Chiplet Based Systems. In: Tari, Z., Li, K., Wu, H. (eds) Algorithms and Architectures for Parallel Processing. ICA3PP 2023. Lecture Notes in Computer Science, vol 14487. Springer, Singapore. https://doi.org/10.1007/978-981-97-0834-5_27
Download citation
DOI: https://doi.org/10.1007/978-981-97-0834-5_27
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-97-0833-8
Online ISBN: 978-981-97-0834-5
eBook Packages: Computer ScienceComputer Science (R0)