Logic-Compatible Embedded DRAM Architecture for Multifunctional Digital Storage and Compute-in-Memory
Source abstract
Abstract
The compute-in-memory (CIM) which embeds computation inside memory is an attractive scheme to circumvent von Neumann bottlenecks. This study proposes a logic-compatible embedded DRAM architecture that supports data storage as well as versatile digital computations. The proposed configurable memory unit operates in three modes: (1) memory mode in which it works as a normal dynamic memory, (2) logic–arithmetic mode where it performs bit-wise Boolean logic and full adder operations on two words stored within the memory array, and (3) convolution mode in which it executes digitally XNOR-and-accumulate (XAC) operation for binarized neural networks. A 1.0-V 4096-word × 8-bit computational DRAM implemented in a 45-nanometer CMOS technology performs memory, logic and arithmetic operations at 241, 229, and 224 MHz while consuming the energy of 7.92, 8.09, and 8.19 pJ/cycle. Compared with conventional digital computing, it saves energy and latency of the arithmetic operation by at least 47% and 46%, respectively. For VDD = 1.0 V, the proposed CIM unit performs two 128-input XAC operations at 292 MHz with an energy consumption of 20.8 pJ/cycle, achieving 24.6 TOPS/W. This marks at least 11.9× better energy efficiency and 38.8× better delay, thereby achieving at least 461× better energy-delay product than traditional 8-bit wide computing hardware.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Advanced Memory and Neural Computing · Ferroelectric and Negative Capacitance Devices · Parallel Computing and Optimization Techniques
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
School of Electronic and Electrical Engineering, Kyungpook National University, Daegu 41566, Republic of Korea
Funding
Korea government (MSIT) · Ministry of Education of Korea
References (62)
Show source-provided references
1. In-memory computing advances and prospects10.1109/MSSC.2019.2922889 ↗
2. Processing-in-memory technology for machine learning: From basic to ASIC10.1109/TCSII.2022.3168404 ↗
3. A 28 nm configurable memory (TCAM/BCAM/SRAM) using push-rule 6T bit cell enabling logic-in-memory10.1109/JSSC.2016.2515510 ↗
4. A 4 + 2T SRAM for searching and in-memory computing with 0.3-V VDDmin10.1109/JSSC.2017.2776309 ↗
5. A 28-nm compute SRAM with bit-serial logic/arithmetic operations for programmable in-memory vector computing10.1109/JSSC.2019.2939682 ↗
6. Chih, Y.-D., Lee, P.-H., Fujiwara, H., Shih, Y.-C., Lee, C.-F., Naous, R., Chen, Y.-L., Lo, C.-P., Lu, C.-H., and Mori, H. (2021, January 13–22). An 89TOPS/W and 16.3TOPS/mm2 all-digital SRAM-based full-precision compute-in memory macro in 22nm for machine-learning edge applications. Proceedings of the IEEE International Solid-State Circuits Conference, San Francisco, CA, USA.10.1109/ISSCC42613.2021.9365766 ↗
7. Z-PIM: A sparsity-aware processing-in-memory architecture with fully variable weight bit-precision for energy-efficient deep neural networks10.1109/JSSC.2020.3039206 ↗
8. Two-direction in-memory computing based on 10T SRAM with horizontal and vertical decoupled read ports10.1109/JSSC.2021.3061260 ↗
9. A 40-nm CMOS multifunctional computing-in-memory (CIM) using single-ended disturb-free 7T 1-Kb SRAM10.1109/TVLSI.2021.3115970 ↗
10. A 1-16b reconfigurable 80Kb 7T SRAM-based digital near-memory computing macro for processing neural networks10.1109/TCSI.2022.3232648 ↗
11. In situ storing 8T SRAM-CIM macro for full-array Boolean logic and copy operations10.1109/JSSC.2022.3206318 ↗
12. In-memory computation of a machine-learning classifier in a standard 6T SRAM array10.1109/JSSC.2016.2642198 ↗
13. A multi-functional in-memory inference processor using a standard 6T SRAM array10.1109/JSSC.2017.2782087 ↗
14. CONV-SRAM: An energy-efficient SRAM with in-memory dot-product computation for low-power convolutional neural networks10.1109/JSSC.2018.2880918 ↗
15. Yang, J., Kong, Y., Wang, Z., Liu, Y., Wang, B., Yin, S., and Shi, L. (2019, January 17–21). Sandwich-RAM: An energy-efficient in-memory BWN architecture with pulse-width modulation. Proceedings of the IEEE International Solid-State Circuits Conference, San Francisco, CA, USA.10.1109/ISSCC.2019.8662435 ↗
16. A dual-split 6T SRAM-based computing-in-memory unit-macro with fully parallel product-sum operation for binarized DNN edge processors10.1109/TCSI.2019.2928043 ↗
17. A twin-8T SRAM computation-in-memory unit-macro for multibit CNN-based AI edge processors10.1109/JSSC.2019.2952773 ↗
18. XNOR-SRAM: In-memory computing SRAM macro for binary/ternary deep neural networks
19. C3SRAM: An in-memory-computing SRAM macro based on robust capacitive coupling computing mechanism10.1109/JSSC.2020.2992886 ↗
20. A 7-nm compute-in-memory SRAM macro supporting multi-bit input, weight and output and achieving 351 TOPS/W and 372.4 GOPS10.1109/JSSC.2020.3031290 ↗
21. CAP-RAM: A charge-domain in-memory computing 6T-SRAM for accurate and precision-programmable CNN inference10.1109/JSSC.2021.3056447 ↗
22. ±CIM SRAM for signed in-memory broad-purpose computing from DSP to neural processing10.1109/JSSC.2021.3092759 ↗
23. Two-way transpose multibit 6T SRAM computing-in-memory macro for inference-training AI edge chips10.1109/JSSC.2021.3108344 ↗
24. An in-memory computing SRAM macro for memory-augmented neural network10.1109/TCSII.2021.3132063 ↗
25. A 65 nm 73 kb SRAM-based computing-in-memory macro with dynamic-sparsity controlling10.1109/TCSII.2022.3162017 ↗
26. A 65-nm 8T SRAM compute-in-memory macro with column ADCs for processing neural networks10.1109/JSSC.2022.3162602 ↗
27. A charge domain SRAM compute-in-memory macro with C-2C ladder-based 8-bit MAC unit in 22-nm FinFET process for edge inference10.1109/JSSC.2022.3232601 ↗
28. A 2941-TOPS/W charge-domain 10T SRAM compute-in-memory for ternary neural network10.1109/TCSI.2023.3241385 ↗
29. A 28nm 32Kb SRAM computing-in-memory macro with hierarchical capacity attenuator and input sparsity-optimized ADC for 4b Mac operation10.1109/TCSII.2023.3234620 ↗
30. A bit-serial, compute-in-SRAM design featuring hybrid-integrating ADCs and input dependent binary scaled precharge eliminating DACs for energy-efficient DNN inference10.1109/JSSC.2023.3235210 ↗
31. A floating-point 6T SRAM in-memory-compute macro using hybrid-domain structure for advanced AI edge chips10.1109/JSSC.2023.3309966 ↗
32. eDRAM-CIM: Reconfigurable charge domain compute-in-memory design with embedded dynamic random access memory array realizing adaptive data converters10.1109/JSSC.2023.3326094 ↗
33. Chen, W.-H., Li, K.-X., Lin, W.-Y., Hsu, K.-H., Li, P.-Y., Yang, C.-H., Xue, C.X., Yang, E.Y., Chen, Y.K., and Chang, Y.S. (2018, January 11–15). A 65nm 1Mb nonvolatile computing-in-memory ReRAM macro with sub-16ns multiply-and-accumulate for binary DNN AI edge processors. Proceedings of the IEEE International Solid-State Circuits Conference, San Francisco, CA, USA.10.1109/ISSCC.2018.8310400 ↗
34. Bocquet, M., Hirztlin, T., Klein, J.-O., Nowak, E., Vianello, E., Portal, J.-M., and Querlioz, D. (2018, January 1–5). In-memory and error-immune differential RRAM implementation of binarized deep neural networks. Proceedings of the IEEE International Electron Devices Meeting, San Francisco, CA, USA.10.1109/IEDM.2018.8614639 ↗
35. Embedded 1-Mb ReRAM-based computing-in-memory macro with multibit input and weight for CNN-based AI edge processors10.1109/JSSC.2019.2951363 ↗
36. Liquid silicon: A nonvolatile fully programmable processing-in-memory processor with monolithically integrated ReRAM10.1109/JSSC.2019.2963005 ↗
37. Xue, C.-X., Hung, J.-M., Kao, H.-Y., Huang, Y.-H., Huang, S.-P., Chang, F.-C., Chen, P., Liu, T.W., Jhang, C.J., and Su, C.I. (2021, January 13–22). A 22nm 4Mb 8b-precision ReRAM computing-in-memory macro with 11.91 to 195.7TOPS/W for tiny AI edge devices. Proceedings of the IEEE International Solid-State Circuits Conference, San Francisco, CA, USA.10.1109/ISSCC42613.2021.9365769 ↗
38. A 40-nm 118.44-TOPS/W voltage-sensing compute-in-memory RRAM macro with write verification and multi-bit encoding10.1109/JSSC.2022.3141370 ↗
39. A 40-nm MLC-RRAM compute-in-memory macro with sparsity control, on-chip write-verify, and temperature-independent ADC references10.1109/JSSC.2022.3163197 ↗
40. A 28-nm RRAM computing-in-memory macro using weighted hybrid 2T1R cell array and reference subtracting sense amplifier for AI edge inference10.1109/JSSC.2023.3280357 ↗
41. HERMES-core—A 1.59-TOPS/mm2 PCM on 14-nm CMOS in-memory compute core using 300-ps/LSB linearized CCO-based ADCs10.1109/JSSC.2022.3140414 ↗
42. An 8b-precision 8-Mb STT-MRAM near-memory-compute macro using weight-feature and input-sparsity aware schemes for energy-efficient edge AI devices10.1109/JSSC.2023.3324335 ↗
43. Programmable in-memory computing circuit for solving combinatorial matrix operation in one step10.1109/TCSI.2023.3263217 ↗
44. Nail, C., Molas, G., Blaise, P., Piccolboni, G., Sklenard, B., Cagli, C., Bernard, M., Roule, A., Azzaz, M., and Vianello, E. (2016, January 3–7). Understanding RRAM endurance, retention and window margin trade-off using experimental results and simulations. Proceedings of the IEEE International Electron Devices Meeting, San Francisco, CA, USA.
45. A survey of architectural approaches for managing embedded DRAM and non-volatile on-chip caches10.1109/TPDS.2014.2324563 ↗
46. Dou, C.-M., Chen, W.-H., Xue, C.-X., Lin, W.-Y., Lin, W.-E., Li, J.-Y., Lin, H.T., and Chang, M.F. (2018, January 18–22). Nonvolatile circuits-devices interaction for memory, logic and artificial intelligence. Proceedings of the IEEE Symposium on VLSI Technology, Honolulu, HI, USA.10.1109/VLSIT.2018.8510627 ↗
47. Recent advances in in-memory computing: Exploring memristor and memtransistor arrays with 2D materials10.1007/s40820-024-01335-2 ↗
48. A 3T gain cell embedded DRAM utilizing preferential boosting for high density and low power on-die caches10.1109/JSSC.2011.2128150 ↗
49. A 667 MHz logic-compatible embedded DRAM featuring an asymmetric 2T gain cell for high speed on-die caches10.1109/JSSC.2011.2168729 ↗
50. A 2T1C embedded DRAM macro with no boosted supplies featuring a 7T SRAM based repair and a cell storage monitor10.1109/JSSC.2012.2206685 ↗
51. A write-back-free 2T1D embedded DRAM with local voltage sensing and a dual-row-access low power mode10.1109/TCSI.2013.2252652 ↗
52. A logic-compatible embedded DRAM utilizing common-body toggled capacitive cross-talk10.5573/JSTS.2016.16.6.781 ↗
53. P-channel logic 2 T eDRAM macro with high retention bit architecture10.1002/cta.2496 ↗
54. Chen, Z., Chen, X., and Gu, J. (2021, January 13–22). A 65nm 3T dynamic analog RAM-based computing-in-memory macro and CNN accelerator with retention enhancement, adaptive analog sparsity and 44TOPS/W system energy efficiency. Proceedings of the IEEE International Solid-State Circuits Conference, San Francisco, CA, USA.10.1109/ISSCC42613.2021.9366045 ↗
55. A logic-compatible eDRAM compute-in-memory with embedded ADCs for processing neural networks10.1109/TCSI.2020.3036209 ↗
56. A 36.2 dB high SNR and PVT/leakage-robust eDRAM computing-in-memory macro with segmented BL and reference cell array10.1109/TCSII.2022.3159808 ↗
57. Xie, S., Ni, C., Jain, P., Hamzaoglu, F., and Kulkarni, J.P. (2022, January 13–17). Gain-cell CIM: Leakage and bitline swing aware 2T1C gain-cell eDRAM compute in memory design with bitline precharge DACs and compact Schmitt trigger ADCs. Proceedings of the IEEE Symposium on VLSI Technology and Circuits, Honolulu, HI, USA.10.1109/VLSITechnologyandCir46769.2022.9830338 ↗
58. DynaPlasia: An eDRAM in-memory computing-based reconfigurable spatial accelerator with triple-mode cell10.1109/JSSC.2023.3319962 ↗
59. Scaling-CIM: eDRAM in-memory-computing accelerator with dynamic-scaling ADC and adaptive analog operation10.1109/JSSC.2024.3362699 ↗
60. An eDRAM-based computing-in-memory macro with full-valid-storage and channel-wise-parallelism for depthwise neural network10.1109/TCSII.2024.3375319 ↗
61. He, Y., Fan, S., Li, X., Lei, L., Jia, W., Tang, C., Li, Y., Huang, Z., Du, Z., and Yue, J. (2024, January 18–22). A 28nm 2.4Mb/mm2 6.9-16.3TOPS/mm2 eDRAM-LUT-based digital-computing-in-memory macro with in-memory encoding and refreshing. Proceedings of the IEEE International Solid-State Circuits Conference, San Francisco, CA, USA.10.1109/ISSCC49657.2024.10454323 ↗
62. Simons, T., and Lee, D.-J. (2019). A review of binarized neural networks. Electronics, 8.10.3390/electronics8060661 ↗