KiF: Accelerating Low-Batch LLM Inference Using In-Flash KV Cache.

Jae Yong Lee 0004 Jaeyong Lee 0004 Seoul National University, Seoul, Korea https://orcid.org/0000-0002-2724-8888 Jae Yong Lee Jaeyong Lee Jae-Yong Lee Jae Yong Lee 0001 Jae-Yong Lee 0001 Jaeyong Lee 0001 Chungnam National University, Daejeon, Korea https://orcid.org/0000-0002-4967-911X https://www.wikidata.org/entity/Q80634610 Jae Yong Lee 0002 Korea Institute for Advanced Study, Seoul, Korea https://orcid.org/0000-0003-0193-545X Jae Yong Lee 0003 Jaeyong Lee 0003 Seoul National University, Seoul, Korea https://jylee749.wordpress.com/ https://orcid.org/0000-0003-3630-5933 Jae Yong Lee 0006 University of Illinois at Urbana-Champaign, Department of Computer Science, IL, USA Jae Yong Lee Jaeyong Lee Jae-Yong Lee Jae Yong Lee 0001 Jae-Yong Lee 0001 Jaeyong Lee 0001 Chungnam National University, Daejeon, Korea https://orcid.org/0000-0002-4967-911X https://www.wikidata.org/entity/Q80634610 Jae Yong Lee 0003 Jaeyong Lee 0003 Seoul National University, Seoul, Korea https://jylee749.wordpress.com/ https://orcid.org/0000-0003-3630-5933 Jaeyong Lee 0005 Soongsil University, Department of Electronic Engineering, Seoul, Korea https://orcid.org/0000-0002-0073-2733

Haeeun Jung Jaeyong Lee 0004 Sanggu Lee Huiwon Yun Dongsuk Jeon Nam Sung Kim Jihong Kim 0001 KiF: Accelerating Low-Batch LLM Inference Using In-Flash KV Cache. 2026 January - June 174-177 25 IEEE Comput. Archit. Lett. 1 https://doi.org/10.1109/LCA.2026.3689575 db/journals/cal/cal25.html#JungLLYJKK26 streams/journals/cal

Hyeunjoo Kim Sanghun Oh Jaeyong Lee 0004 Jihong Kim 0001 Lazy Discharge: A High- Speed Energy-Efficient Read Technique for NAND Flash Memory. 49-57 2026 February 43 IEEE Des. Test 1 https://doi.org/10.1109/MDAT.2025.3591413 db/journals/dt/dt43.html#KimOLK26 streams/journals/dt

Myoungjun Chun Jaeyong Lee 0004 Inhyuk Choi Jisung Park 0001 Myungsuk Kim Jihong Kim 0001 STRAW: Stress-Aware WL-Based Read Disturbance Management for High-Density NAND Flash Memory. 1897-1911 2026 ASPLOS (2) https://doi.org/10.1145/3779212.3790228 conf/asplos/2026-2 db/conf/asplos/asplos2026-2.html#ChunLC0K026

Myoungjun Chun Jae Yong Lee 0004 Inhyuk Choi Jisung Park 0001 Myungsuk Kim Jihong Kim 0001 Straw: A Stress-Aware WL-Based Read Reclaim Technique for High-Density NAND Flash-Based SSDs. 5-8 2025 January - June 24 IEEE Comput. Archit. Lett. 1 https://doi.org/10.1109/LCA.2024.3516205 db/journals/cal/cal24.html#ChunLCPKK25 streams/journals/cal

Jeongho Lee Sangjun Kim Jae Yong Lee 0004 Jaeyoung Kang 0004 Sungjin Lee 0001 Nam Sung Kim Jihong Kim 0001 srNAND: A Novel NAND Flash Organization for Enhanced Small Read Throughput in SSDs. 197-200 2025 July - December 24 IEEE Comput. Archit. Lett. 2 https://doi.org/10.1109/LCA.2025.3571321 db/journals/cal/cal24.html#LeeKLKLKK25 streams/journals/cal

Jaeyong Lee 0004 Hyeunjoo Kim Sanghun Oh Myoungjun Chun Myungsuk Kim Jihong Kim 0001 AiF: Accelerating On-Device LLM Inference Using In-Flash Processing. 529-543 2025 ISCA https://doi.org/10.1145/3695053.3731073 conf/isca/2025 db/conf/isca/isca2025.html#LeeKOCK025 Jaeyong Lee 0004 Beomjun Kim Myoungjun Chun Myungsuk Kim Jihong Kim 0001 DEAR: Improving Performance and Lifetime of SSDs Using Dynamic Error-Aware Refresh. 1207-1220 2025 MICRO https://doi.org/10.1145/3725843.3756134 conf/micro/2025 db/conf/micro/micro2025.html#LeeKCK025

Myoungjun Chun Jaeyong Lee 0004 Inhyuk Choi Jisung Park 0001 Myungsuk Kim Jihong Kim 0001 STRAW: A Stress-Aware WL-Based Read Reclaim Technique for High-Density NAND Flash-Based SSDs. 2025 January abs/2501.02517 CoRR https://doi.org/10.48550/arXiv.2501.02517 db/journals/corr/corr2501.html#abs-2501-02517 streams/journals/corr

Myoungjun Chun Jaeyong Lee 0004 Myungsuk Kim Jisung Park 0001 Jihong Kim 0001 RiF: Improving Read Performance of Modern SSDs Using an On-Die Early-Retry Engine. 643-656 2024 HPCA https://doi.org/10.1109/HPCA57654.2024.00056 conf/hpca/2024 db/conf/hpca/hpca2024.html#ChunLKPK24 Jaeyong Lee 0004 Myungsuk Kim Wonil Choi Sanggu Lee Jihong Kim 0001 TailCut: improving performance and lifetime of SSDs using pattern-aware state encoding. 409-414 2022 DAC https://doi.org/10.1145/3489517.3530471 https://www.wikidata.org/entity/Q130955660 conf/dac/2022 db/conf/dac/dac2022.html#LeeKCL022 Myoungjun Chun Jaeyong Lee 0004 Sanggu Lee Myungsuk Kim Jihong Kim 0001 PiF: in-flash acceleration for data-intensive applications. 106-112 2022 HotStorage https://doi.org/10.1145/3538643.3539742 https://www.wikidata.org/entity/Q130962078 conf/hotstorage/2022 db/conf/hotstorage/hotstorage2022.html#ChunLLK022 Inhyuk Choi Wonil Choi Myoungjun Chun Dongsuk Jeon Haeeun Jung Jaeyoung Kang 0004 Beomjun Kim Hyeunjoo Kim Jihong Kim 0001 Myungsuk Kim Nam Sung Kim Sangjun Kim Jeongho Lee Sanggu Lee Sungjin Lee 0001 Sanghun Oh Jisung Park 0001 Huiwon Yun