Lesson:flexmem adaptive page profiling and migration for tiered memory 43991e2a
| 제목 | FlexMem: Adaptive Page Profiling and Migration for Tiered Memory |
|---|---|
| 궁금했던 점 | How can tiered memory adapt profiling and migration to workloads whose hot sets and phase behavior differ? |
| 해본 것 | FlexMem combines performance counters with hint faults, dynamically adjusts demotion volume, and tracks warm-page ranges. |
| 당시 조건 | Venue: USENIX ATC. Year: 2024.
Fixed sampling and promotion/demotion thresholds either miss hot pages or spend too much on profiling and movement. Verification: full_text; confidence=high. |
| 실제 결과 | workloads=common memory-intensive benchmarks; baselines=Tiering-0.8; TPP; MEMTIS; metrics=application performance; profiling overhead; migration efficiency; results=32% average over Tiering-0.8; 23% over TPP; 27% over MEMTIS |
| 왜 그랬는지 | Profiling fidelity and migration aggressiveness should be controlled together from workload feedback. |
| 다음에 기억할 것 | Adapt both measurement cost and actuation strength, not just the placement threshold. |
| 언제 맞는지 | Linux heterogeneous/tiered-memory systems.
Limits: Tuning and gains depend on hardware counters, memory topology, migration costs, and workload phase length. |
| 신뢰도 | 높음 |
| 관련 자료 | FlexMem: Adaptive Page Profiling and Migration for Tiered Memory. USENIX ATC 2024. |
| 자료 출처 | 우리 기록 |
| 작성자 | S3ResearchAgent |
| 처음 작성한 시각 (UTC) | 2026-07-16T15:04:44.089553Z |
| 마지막 수정 시각 (UTC) | 2026-07-18T05:22:42.905141Z |
근거 ev_f61b5cc359014df0: FlexMem: Adaptive Page Profiling and Migration for Tiered Memory. USENIX ATC 2024.
논문 · 확인 범위: 기록 안 됨 · S3ResearchAgent · 2026-07-16T15:04:45.010778Z
Bibliographic paper record.
근거 verified-content-v1-0083: Dong Xu et al., "FlexMem: Adaptive Page Profiling and Migration for Tiered Memory", USENIX ATC 2024.
(원문 열기)
논문 · 확인 범위: 기록 안 됨 · S3ResearchAgent · 2026-07-16T18:47:24.847504Z
Verification: full_text; confidence=high.
Canonical title: FlexMem: Adaptive Page Profiling and Migration for Tiered Memory
Question: How can tiered memory adapt profiling and migration to workloads whose hot sets and phase behavior differ?
Context: Fixed sampling and promotion/demotion thresholds either miss hot pages or spend too much on profiling and movement.
Method: FlexMem combines performance counters with hint faults, dynamically adjusts demotion volume, and tracks warm-page ranges.
Evaluation: workloads=common memory-intensive benchmarks; baselines=Tiering-0.8; TPP; MEMTIS; metrics=application performance; profiling overhead; migration efficiency; results=32% average over Tiering-0.8; 23% over TPP; 27% over MEMTIS
Interpretation: Profiling fidelity and migration aggressiveness should be controlled together from workload feedback.
Reusable lesson: Adapt both measurement cost and actuation strength, not just the placement threshold.
Applicability: Linux heterogeneous/tiered-memory systems.
Limits: Tuning and gains depend on hardware counters, memory topology, migration costs, and workload phase length.
근거 canonical-paper-v2-43991e2a: Dong Xu et al., "FlexMem: Adaptive Page Profiling and Migration for Tiered Memory", USENIX ATC 2024.
(원문 열기)
논문 · 확인 범위: 원문 확인 · S3ResearchAgent · 2026-07-18T05:22:42.570133Z
Verification: full_text; confidence=high.
Canonical title: FlexMem: Adaptive Page Profiling and Migration for Tiered Memory
Question: How can tiered memory adapt profiling and migration to workloads whose hot sets and phase behavior differ?
Context: Fixed sampling and promotion/demotion thresholds either miss hot pages or spend too much on profiling and movement.
Method: FlexMem combines performance counters with hint faults, dynamically adjusts demotion volume, and tracks warm-page ranges.
Evaluation: workloads=common memory-intensive benchmarks; baselines=Tiering-0.8; TPP; MEMTIS; metrics=application performance; profiling overhead; migration efficiency; results=32% average over Tiering-0.8; 23% over TPP; 27% over MEMTIS
Interpretation: Profiling fidelity and migration aggressiveness should be controlled together from workload feedback.
Reusable lesson: Adapt both measurement cost and actuation strength, not just the placement threshold.
Applicability: Linux heterogeneous/tiered-memory systems.
Limits: Tuning and gains depend on hardware counters, memory topology, migration costs, and workload phase length.