속성:Context
하드웨어, 부하, 버전, 규모처럼 결과에 영향을 줄 수 있는 조건을 적습니다.
s
Venue: OSDI. Year: 2022.
Meta needed one distribution service for millions of client processes and heterogeneous workloads.
Verification: official USENIX OSDI 2022 paper page/abstract. +
Venue: EuroSys. Year: 2024.
Shared XArray operations and serialized writeback become bottlenecks before a multi-NVMe array is saturated.
Verification: official_abstract; confidence=high. +
Venue: FAST. Year: 2026.
Linux swap centralizes queues/LRU metadata and underutilizes parallel all-flash arrays.
Verification: official USENIX page and abstract; confidence=high. +
Venue: OSDI. Year: 2025.
Centralized target processing and inconsistent cache states limit throughput in DPU/JBOF deployments.
Verification: official USENIX page and abstract; confidence=high. +
Publication scope: international. Lab publication metadata:
1. Hwanju Kim, Jinkyu Jeong, Jaeho Hwang, Joonwon Lee, Seungryoul Maeng, "Scheduler Support for Video-oriented Multimedia on Client-side Virtualization," in Proceedings of ACM Multimedia Systems (MMSys 2012), Chapel Hill, North Carolina, USA, February, 2012
Verification level: full_text. Sources checked:
https://doi.org/10.1145/2155555.2155566
https://yonsei.elsevierpure.com/en/publications/scheduler-support-for-video-oriented-multimedia-on-client-side-vi/
https://www.researchgate.net/publication/221636635_Scheduler_support_for_video-oriented_multimedia_on_client-side_virtualization +
Publication scope: international. Lab publication metadata:
1. Minwoo Ahn, Donghyun Kim, Taekeun Nam, Jinkyu Jeong, "SCOZ: A System‐wide Causal Profiler for Multicore Systems," Software: Practice and Experience, vol. 51, issue 5, pp. 1043-1058, May 2021, doi:10.1002/spe.2930
Verification level: full_text. Sources checked:
https://onlinelibrary.wiley.com/doi/full/10.1002/spe.2930
https://doi.org/10.1002/spe.2930
https://yonsei.elsevierpure.com/en/publications/scoz-a-system-wide-causal-profiler-for-multicore-systems/ +
Publication scope: international. Lab publication metadata:
1. Sung-hun Kim, Jinkyu Jeong, and Joonwon Lee, "Selective Memory Deduplication for Cost Efficiency in Mobile Smart Devices," IEEE Transactions on Consumer Electronics, vol. 60, no. 2, pp. 276-284, May 2014
Verification level: official_abstract. Sources checked:
https://doi.org/10.1109/TCE.2014.6852004
https://yonsei.elsevierpure.com/en/publications/selective-memory-deduplication-for-cost-efficiency-in-mobile-smar/ +
Sequential Qwen and Mistral target-generation migration candidates on one isolated NVIDIA B200 GPU. a03 used commit dd81c5ff98ffce72cd907a416fd42efbb41f920b and failed only at a buggy suite-end source comparison; a04 used fixed commit 99ccfb7b7f55199876191ff443fe7398db7a8e1a and passed. +
serverless in the wild characterizing and optimizing the serverless workload at a large cloud pr f2fa1121 +
Venue: USENIX ATC. Year: 2020.
공개 실측이 부족해 서버리스 자원 정책이 단순 가정에 의존했다.
Verification: abstract_only; confidence=high. +
Venue: OSDI. Year: 2024.
Loading model checkpoints from remote storage dominates cold starts, while keeping every model resident is too costly.
Verification: full_text; confidence=high. +
Shape-adaptive attention dispatch on one NVIDIA L40S, with five fixed Triton tile configurations, causal GQA prefill/decode cases, eager and CUDA Graph execution, and a production-motivated hypothesis that per-shape specialization plus RoPE-attention fusion would materially outperform one global kernel. +
Venue: FAST. Year: 2025.
Clients repeatedly fail atomic acquisition of a held lock, inflating tail latency and consuming lock-server inbound IOPS.
Verification: official_abstract; confidence=high. +
Venue: NSDI. Year: 2024.
최신 정책은 복잡하고 생산 캐시 라이브러리에 넣기 어렵다.
Verification: abstract_only; confidence=high. +
Venue: USENIX ATC. Year: 2019.
Background compaction competes with user I/O and can create long stalls despite high average throughput.
Verification: official USENIX page and abstract; confidence=high. +
Venue: SOSP. Year: 2023.
Rowhammer 완화는 큰 성능·용량 비용이 들거나 하드웨어 지원이 필요하다.
Verification: abstract_only; confidence=high. +
Publication scope: international. Lab publication metadata:
1. Sang-Hoon Kim, Jinkyu Jeong, Jin-Soo Kim, and Seungryoul Maeng, "SmartLMK: A Memory Reclamation Scheme for Improving User-perceived App Launch Time," ACM Transactions on Embedded Computing Systems, vol. 15, no. 3, article 47, May 2016
Verification level: official_abstract. Sources checked:
https://doi.org/10.1145/2894755
https://yonsei.elsevierpure.com/en/publications/smartlmk-a-memory-reclamation-scheme-for-improving-user-perceived/ +
specinfer accelerating generative large language model serving with tree based speculative infer e609aa25 +
Venue: ASPLOS. Year: 2024.
A single small draft model can propose poor tokens, limiting acceptance and leaving target-model parallelism unused.
Verification: full_text; confidence=high. +
Publication scope: international. Lab publication metadata:
1. Jongwon Park, Jinkyu Jeong, "Speculative Multi-Level Access in LSM Tree-Based KV Store," Computer Architecture Letters, Vol. 21, Issue 2, pp. 145-148, Nov. 2022, 10.1109/LCA.2022.3219808
Verification level: official_abstract. Sources checked:
https://yonsei.elsevierpure.com/en/publications/speculative-multi-level-access-in-lsm-tree-based-kv-store/
https://doi.org/10.1109/LCA.2022.3219808 +
speed is all you need on device acceleration of large diffusion models via gpu aware optimizatio c53959fd +
Venue: CVPR Workshops. Year: 2023.
Stable Diffusion의 반복 UNet과 attention/convolution kernel 및 메모리 경로가 모바일 GPU 지연을 키운다.
Verification: full_text; confidence=high. +
Publication scope: international. Lab publication metadata:
1. Jonghyun Bae, Hakbeom Jang, Jeonghun Gong, Wenjing Jin, Shine Kim, Jaeyoung Jang, Tae Jun Ham, Jinkyu Jeong, Jae W. Lee, "SSDStreamer: Specializing I/O Stack for Large-Scale Machine Learning," IEEE Micro, Vol. 39, issue. 5, pp. 73-81, Sep/Oct, 2019.
Verification level: official_abstract. Sources checked:
https://doi.org/10.1109/MM.2019.2930497
https://yonsei.elsevierpure.com/en/publications/ssdstreamer-specializing-io-stack-for-large-scale-machine-learnin/ +