]> 2026-09-08T15:37:37+09:00 Lesson:identifying on off cpu bottlenecks together with blocked samples 1f04444a 3000 ko 스토리지·락·스케줄링 대기가 섞인 Linux 서비스와 DB 성능 분석에 높음. off-CPU 구간에도 주기적 blocked sample(IP, callchain, weight, 원인 유형)을 생성한다. bperf가 on/off-CPU 샘플을 통합하고, BCOZ는 causal virtual speedup을 I/O·동기화·스케줄링의 세부 blocked class로 확장한다. S3ResearchAgent medium Publication scope: international. Lab publication metadata: 1. Minwoo Ahn, Jeongmin Han, Youngjin Kwon, Jinkyu Jeong, "Identifying On-/Off-CPU Bottlenecks Together with Blocked Samples," in Proceedings of the 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI '24), July 2024 Verification level: full_text. Sources checked: https://www.usenix.org/conference/osdi24/presentation/ahn https://www.usenix.org/system/files/osdi24-ahn.pdf 2026-07-16T15:02:39Z 2461238.1268403 Publication record 1: Minwoo Ahn, Jeongmin Han, Youngjin Kwon, Jinkyu Jeong, "Identifying On-/Off-CPU Bottlenecks Together with Blocked Samples," in Proceedings of the 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI '24), July 2024 BCOZ의 27.6% 평균/64.7% 최대 오버헤드는 상시 운영 관측도구로는 크다. 가상 speedup은 실제 최적화 효과를 과대·과소평가할 수 있어 정확한 수치 예측보다 우선순위 도구로 써야 한다. blocked sample 주기와 지연 보정은 진단 해상도와 오버헤드의 trade-off를 갖는다. Technical interpretation: wall-clock 병목 분석은 실행 샘플만으로 부족하며, 대기 구간도 시간 가중 샘플과 원인 분류를 가져야 동일한 통계·인과 프레임으로 처리할 수 있다. Identifying On-/Off-CPU Bottlenecks Together with Blocked Samples bperf의 평균 성능 저하는 1.6%로 perf sampling 0.9%, tracing 3.6%와 비교해 perf보다 0.7%p 추가 비용. BCOZ end-to-end overhead는 평균 27.6%, 최대 64.7%. RocksDB 혼합 워크로드에서 병목을 식별했고 실제 최적화의 방향/순위가 가상 speedup 예측과 대체로 일치. CPU를 쓰는 구간과 I/O·동기화·스케줄링으로 막힌 구간을 한 프로파일에서 코드 라인 수준으로 연결하고, off-CPU 코드 최적화의 잠재효과까지 예측할 수 있는가? lab wall-clock 병목 분석은 실행 샘플만으로 부족하며, 대기 구간도 시간 가중 샘플과 원인 분류를 가져야 동일한 통계·인과 프레임으로 처리할 수 있다. Draft 2026-07-18T14:58:36Z 2461240.1240278 2026-07-18T14:58:36Z 2461240.1240278 identifying on off cpu bottlenecks together with blocked samples 1f04444a 3000 ev_a21c50fcb7804018 2026-07-16T15:02:40Z 2461238.1268519 S3ResearchAgent Minwoo Ahn, Jeongmin Han, Youngjin Kwon, Jinkyu Jeong, "Identifying On-/Off-CPU Bottlenecks Together with Blocked Samples," in Proceedings of the 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI '24), July 2024 paper Source: Yonsei University Computer Systems Laboratory publication list supplied by the user. Manifestation 1 of 1. not_recorded identifying on off cpu bottlenecks together with blocked samples 1f04444a# 23c81ec42b302d10fe2b6dd718d564fb 3000 canonical-paper-v2-1f04444a 2026-07-18T05:24:38Z 2461239.7254398 S3ResearchAgent Minwoo Ahn, Jeongmin Han, Youngjin Kwon, Jinkyu Jeong, "Identifying On-/Off-CPU Bottlenecks Together with Blocked Samples" in Proceedings of the 18th USENIX Symposium on Operating Systems Design and Implementation (OSDI '24), July 2024 paper 정본 Lesson 보강 근거. 검토 원본: Lesson:technical_review_identifying_on_off_cpu_bottlenecks_together_with_blocked_samples_2f1283b1. 확인 범위: full_text. 확인한 자료: https://www.usenix.org/conference/osdi24/presentation/ahn ; https://www.usenix.org/system/files/osdi24-ahn.pdf. 질문, 방법, 평가, 해석, 재사용 교훈, 적용 범위와 한계를 같은 Lesson 본문에 통합했습니다. full_text identifying on off cpu bottlenecks together with blocked samples 1f04444a# 176c28a38fa32fc0641bfc7c0e381d44 3000 2026-07-18T14:58:35Z 2461240.1240162 S3ResearchAgent verify_cb2b35afd0f77dce4dc5 partial_source context 보존 원문 객체를 확보했으나 이 일괄 검증에서는 claim-bearing 범위를 재판정하지 않아 결론을 보류함. inconclusive 076d0d62f9f8ab9bbd595206b8d6f46b4d27c6479d9579ad6eb82c8ca3a3022d R2-RESTIC:7f893ca5afd2cfb6fe320e9b61063ccc70e75a7a96589420038c8cf338b273be; archive-manifest-sha256=e28171fb69e141ce306d92dfe4b10e6cdc6e81d4fa910c30a846204dbcf8edf8; sha256=076d0d62f9f8ab9bbd595206b8d6f46b4d27c6479d9579ad6eb82c8ca3a3022d 보존 파일 objects/sha256/07/076d0d62f9f8ab9bbd595206b8d6f46b4d27c6479d9579ad6eb82c8ca3a3022d ev_a21c50fcb7804018 93fbace8b3851198a5271fc8e6cdd1c8f9297e048a09690f870f10f4ccfe60da identifying on off cpu bottlenecks together with blocked samples 1f04444a# 7c537649f99d2d7783ea6113ae9715fd 3000 2026-07-18T14:58:36Z 2461240.1240278 S3ResearchAgent verify_182a3204c23e2589c1e4 partial_source context 보존 원문 객체를 확보했으나 이 일괄 검증에서는 claim-bearing 범위를 재판정하지 않아 결론을 보류함. inconclusive 076d0d62f9f8ab9bbd595206b8d6f46b4d27c6479d9579ad6eb82c8ca3a3022d R2-RESTIC:7f893ca5afd2cfb6fe320e9b61063ccc70e75a7a96589420038c8cf338b273be; archive-manifest-sha256=e28171fb69e141ce306d92dfe4b10e6cdc6e81d4fa910c30a846204dbcf8edf8; sha256=076d0d62f9f8ab9bbd595206b8d6f46b4d27c6479d9579ad6eb82c8ca3a3022d 보존 파일 objects/sha256/07/076d0d62f9f8ab9bbd595206b8d6f46b4d27c6479d9579ad6eb82c8ca3a3022d canonical-paper-v2-1f04444a ebf0291250f7cb2e7002a5fbcfec57a2c5f74b37e17c52cc87ced64892a7e2a5 identifying on off cpu bottlenecks together with blocked samples 1f04444a# bfee74f0221e6c16036bd49fbd51ecd1 3000 2026-07-18T14:58:36Z 2461240.1240278 S3ResearchAgent verify_490088029e076049dd6f official_abstract observation,interpretation,reusable_lesson claim-bearing O/I/R coverage was not established; confidence forced to low inconclusive 076d0d62f9f8ab9bbd595206b8d6f46b4d27c6479d9579ad6eb82c8ca3a3022d R2-RESTIC:7f893ca5afd2cfb6fe320e9b61063ccc70e75a7a96589420038c8cf338b273be; archive-manifest-sha256=e28171fb69e141ce306d92dfe4b10e6cdc6e81d4fa910c30a846204dbcf8edf8; sha256=076d0d62f9f8ab9bbd595206b8d6f46b4d27c6479d9579ad6eb82c8ca3a3022d; high-confidence adjudication ledger Saved official abstract inspected in full; it supports the paper identity/headline only, not every detailed metric, limitation, and derived O/I/R statement. canonical-paper-v2-1f04444a ebf0291250f7cb2e7002a5fbcfec57a2c5f74b37e17c52cc87ced64892a7e2a5 identifying on off cpu bottlenecks together with blocked samples 1f04444a# 515926cf0c76a6fb70b743ff6419e856 Lesson:technical review identifying on off cpu bottlenecks together with blocked samples 2f1283b1 3000 ko technical review identifying on off cpu bottlenecks together with blocked samples 2f1283b1 Supersedes 102 ko Supersedes