속성:Evidence note
Optional note explaining what an evidence item establishes or limits.
a
정본 Lesson 보강 근거. 검토 원본: Lesson:technical_review_a_case_for_hardware_based_demand_paging_163a61b2. 확인 범위: official_abstract. 확인한 자료: https://yonsei.elsevierpure.com/en/publications/a-case-for-hardware-based-demand-paging/ ; https://doi.org/10.1109/ISCA45697.2020.00093. 질문, 방법, 평가, 해석, 재사용 교훈, 적용 범위와 한계를 같은 Lesson 본문에 통합했습니다. +
Source: Yonsei University Computer Systems Laboratory publication list supplied by the user. Manifestation 1 of 1. +
Bibliographic paper record. +
Verification: official USENIX page and abstract; confidence=high.
Canonical title: A Comprehensive Analysis of Superpage Management Mechanisms and Policies
Question: Which superpage lifecycle mechanisms cause latency spikes or memory bloat, and how should policies combine them?
Context: Transparent huge-page systems mix allocation, promotion, compaction, and reclamation decisions whose interactions are poorly understood.
Method: The study builds a lifecycle design-space framework, identifies root causes, and implements the Quicksilver policy.
Evaluation: workloads=superpage-management workloads; baselines=existing superpage policies; metrics=latency spikes, performance, memory bloat; results=qualitative improvement; exact values unavailable in official abstract
Interpretation: Superpage policy should be evaluated as an end-to-end lifecycle, not isolated promotion heuristics.
Reusable lesson: Map mechanisms to lifecycle stages and measure their cross-stage side effects.
Applicability: OS huge-page/superpage management on server workloads.
Limits: Quantitative results and hardware/workload scope remain abstract-only. +
Verification: official USENIX page and abstract; confidence=medium.
Canonical title: A Comprehensive Analysis of Superpage Management Mechanisms and Policies
Question: Which superpage lifecycle mechanisms cause latency spikes or memory bloat, and how should policies combine them?
Context: Transparent huge-page systems mix allocation, promotion, compaction, and reclamation decisions whose interactions are poorly understood.
Method: The study builds a lifecycle design-space framework, identifies root causes, and implements the Quicksilver policy.
Evaluation: workloads=superpage-management workloads; baselines=existing superpage policies; metrics=latency spikes, performance, memory bloat; results=qualitative improvement; exact values unavailable in official abstract
Interpretation: Superpage policy should be evaluated as an end-to-end lifecycle, not isolated promotion heuristics.
Reusable lesson: Map mechanisms to lifecycle stages and measure their cross-stage side effects.
Applicability: OS huge-page/superpage management on server workloads.
Limits: Quantitative results and hardware/workload scope remain abstract-only. +
정본 Lesson 보강 근거. 검토 원본: Lesson:technical_review_a_fully_associative_tagless_dram_cache_8377cc32. 확인 범위: official_abstract. 확인한 자료: https://doi.org/10.1145/2749469.2750383 ; https://yonsei.elsevierpure.com/en/publications/a-fully-associative-tagless-dram-cache/. 질문, 방법, 평가, 해석, 재사용 교훈, 적용 범위와 한계를 같은 Lesson 본문에 통합했습니다. +
Source: Yonsei University Computer Systems Laboratory publication list supplied by the user. Manifestation 1 of 1. +
Source: Yonsei University Computer Systems Laboratory publication list supplied by the user. Manifestation 1 of 1. +
정본 Lesson 보강 근거. 검토 원본: Lesson:technical_review_a_hybrid_web_browser_architecture_for_mobile_devices_19197da7. 확인 범위: full_text. 확인한 자료: https://doi.org/10.4316/AECE.2014.03001 ; https://yonsei.elsevierpure.com/en/publications/a-hybrid-web-browser-architecture-for-mobile-devices/ ; https://www.researchgate.net/publication/277676543_A_Hybrid_Web_Browser_Architecture_for_Mobile_Devices. 질문, 방법, 평가, 해석, 재사용 교훈, 적용 범위와 한계를 같은 Lesson 본문에 통합했습니다. +
Source: Yonsei University Computer Systems Laboratory publication list supplied by the user. Manifestation 1 of 1. +
정본 Lesson 보강 근거. 검토 원본: Lesson:technical_review_a_neural_network_accelerator_for_mobile_application_processors_d4383cb5. 확인 범위: official_abstract. 확인한 자료: https://doi.org/10.1109/TCE.2015.7389812 ; https://yonsei.elsevierpure.com/en/publications/a-neural-network-accelerator-for-mobile-application-processors/. 질문, 방법, 평가, 해석, 재사용 교훈, 적용 범위와 한계를 같은 Lesson 본문에 통합했습니다. +
Source: Yonsei University Computer Systems Laboratory publication list supplied by the user. Manifestation 1 of 1. +
정본 Lesson 보강 근거. 검토 원본: Lesson:technical_review_a_performance_stable_numa_management_scheme_for_linux_based_hpc_systems_e846c351. 확인 범위: full_text. 확인한 자료: https://yonsei.elsevierpure.com/en/publications/a-performance-stable-numa-management-scheme-for-linux-based-hpc-s/ ; https://doi.org/10.1109/ACCESS.2021.3069991. 질문, 방법, 평가, 해석, 재사용 교훈, 적용 범위와 한계를 같은 Lesson 본문에 통합했습니다. +
정본 Lesson 보강 근거. 검토 원본: Lesson:technical_review_a_scalable_and_overflow_tolerant_mechanism_for_minimum_virtual_time_tracking_7a68da61. 확인 범위: official_abstract. 확인한 자료: https://yonsei.elsevierpure.com/en/publications/a-scalable-and-overflow-tolerant-mechanism-for-minimum-virtual-ti/ ; https://doi.org/10.1109/ICCD65941.2025.00027 ; https://www.iccd-conf.com/agenda.html. 질문, 방법, 평가, 해석, 재사용 교훈, 적용 범위와 한계를 같은 Lesson 본문에 통합했습니다. +
Source: Yonsei University Computer Systems Laboratory publication list supplied by the user. Manifestation 1 of 1. +
Source: Yonsei University Computer Systems Laboratory publication list supplied by the user. Manifestation 1 of 1. +
정본 Lesson 보강 근거. 검토 원본: Lesson:technical_review_a_secure_fast_and_resource_efficient_serverless_platform_with_function_rewind_588a4b15. 확인 범위: full_text. 확인한 자료: https://www.usenix.org/conference/atc24/presentation/song ; https://www.usenix.org/system/files/atc24-song.pdf ; https://github.com/s3yonsei/rewind_serverless. 질문, 방법, 평가, 해석, 재사용 교훈, 적용 범위와 한계를 같은 Lesson 본문에 통합했습니다. +
Bibliographic paper record. +
Verification: abstract_only; confidence=high.
Canonical title: ACCL+: an FPGA-Based Collective Engine for Distributed Applications
Question: 분산 애플리케이션의 collective 통신을 FPGA에서 범용·확장 가능하게 가속할 수 있는가?
Context: 고정 하드웨어 collective는 재합성이 필요하고 CPU MPI는 통신·CPU 비용이 크다.
Method: ACCL+는 UDP/TCP/RDMA, FPGA 직접 사용과 CPU offload, 재합성 없는 확장성을 제공한다.
Evaluation: workloads=100 Gb/s FPGA cluster; CPU vector-matrix multiply; FPGA DLR inference; baselines=software MPI; software RDMA collectives; metrics=collective performance; application performance; results=Significant/competitive gains reported; exact aggregate not abstract-verified.
Interpretation: collective를 재사용 가능한 FPGA 네트워크 엔진으로 분리하면 여러 실행 모델을 지원할 수 있다.
Reusable lesson: 가속기 통신은 특정 앱 회로보다 프로토콜·collective 계층의 재사용성을 설계하라.
Applicability: FPGA 클러스터 HPC·분산 추론.
Limits: FPGA/AMD toolchain과 평가 클러스터 규모에 제약되며 초록에 정확한 수치가 없다. +
Verification: abstract_only; confidence=medium.
Canonical title: ACCL+: an FPGA-Based Collective Engine for Distributed Applications
Question: 분산 애플리케이션의 collective 통신을 FPGA에서 범용·확장 가능하게 가속할 수 있는가?
Context: 고정 하드웨어 collective는 재합성이 필요하고 CPU MPI는 통신·CPU 비용이 크다.
Method: ACCL+는 UDP/TCP/RDMA, FPGA 직접 사용과 CPU offload, 재합성 없는 확장성을 제공한다.
Evaluation: workloads=100 Gb/s FPGA cluster; CPU vector-matrix multiply; FPGA DLR inference; baselines=software MPI; software RDMA collectives; metrics=collective performance; application performance; results=Significant/competitive gains reported; exact aggregate not abstract-verified.
Interpretation: collective를 재사용 가능한 FPGA 네트워크 엔진으로 분리하면 여러 실행 모델을 지원할 수 있다.
Reusable lesson: 가속기 통신은 특정 앱 회로보다 프로토콜·collective 계층의 재사용성을 설계하라.
Applicability: FPGA 클러스터 HPC·분산 추론.
Limits: FPGA/AMD toolchain과 평가 클러스터 규모에 제약되며 초록에 정확한 수치가 없다. +