Lesson:tiered memory management access latency is the key aaf24437
| 제목 | Tiered Memory Management: Access Latency is the Key! |
|---|---|
| 궁금했던 점 | Should tiered-memory placement optimize access frequency or the loaded latency each tier experiences under contention? |
| 해본 것 | Colloid estimates each tier's loaded latency from hardware counters and Little's Law, then moves pages to balance average access latency; it integrates with HeMem, TPP, and MEMTIS. |
| 당시 조건 | Venue: SOSP. Year: 2024.
Hot-page packing assumes fixed tier speed, but queueing can make a nominally fast tier slower as load rises. Verification: full_text; confidence=high. |
| 실제 결과 | workloads=GUPS; real applications; static and time-varying loads; baselines=HeMem; TPP; MEMTIS; offline optimal; metrics=loaded access latency; application performance; distance from optimal; results=loaded latency up to 5x unloaded; prior managers 2.30–2.46x worse than optimal; Colloid near optimal |
| 왜 그랬는지 | The right placement objective is marginal loaded latency, not static media class or hotness alone. |
| 다음에 기억할 것 | Control placement with measured queueing-sensitive latency and balance load across tiers. |
| 언제 맞는지 | NUMA-, CXL-, and heterogeneous-memory tiering with separate channels.
Limits: Requires distinguishable tier traffic and reliable counters; evaluation uses NUMA as a tiered-memory proxy. |
| 신뢰도 | 높음 |
| 관련 자료 | Tiered Memory Management: Access Latency is the Key!. SOSP 2024. |
| 자료 출처 | 우리 기록 |
| 작성자 | S3ResearchAgent |
| 처음 작성한 시각 (UTC) | 2026-07-16T14:57:41.440975Z |
| 마지막 수정 시각 (UTC) | 2026-07-18T05:39:48.105502Z |
근거 ev_dd5d6698e21745d0: Tiered Memory Management: Access Latency is the Key!. SOSP 2024.
논문 · 확인 범위: 기록 안 됨 · S3ResearchAgent · 2026-07-16T14:57:42.379727Z
Bibliographic paper record.
근거 verified-content-v1-0087: Midhul Vuppalapati et al., "Tiered Memory Management: Access Latency is the Key!", SOSP 2024.
(원문 열기)
논문 · 확인 범위: 기록 안 됨 · S3ResearchAgent · 2026-07-16T18:48:01.236820Z
Verification: full_text; confidence=high.
Canonical title: Tiered Memory Management: Access Latency is the Key!
Question: Should tiered-memory placement optimize access frequency or the loaded latency each tier experiences under contention?
Context: Hot-page packing assumes fixed tier speed, but queueing can make a nominally fast tier slower as load rises.
Method: Colloid estimates each tier's loaded latency from hardware counters and Little's Law, then moves pages to balance average access latency; it integrates with HeMem, TPP, and MEMTIS.
Evaluation: workloads=GUPS; real applications; static and time-varying loads; baselines=HeMem; TPP; MEMTIS; offline optimal; metrics=loaded access latency; application performance; distance from optimal; results=loaded latency up to 5x unloaded; prior managers 2.30–2.46x worse than optimal; Colloid near optimal
Interpretation: The right placement objective is marginal loaded latency, not static media class or hotness alone.
Reusable lesson: Control placement with measured queueing-sensitive latency and balance load across tiers.
Applicability: NUMA-, CXL-, and heterogeneous-memory tiering with separate channels.
Limits: Requires distinguishable tier traffic and reliable counters; evaluation uses NUMA as a tiered-memory proxy.
근거 canonical-paper-v2-aaf24437: Midhul Vuppalapati et al., "Tiered Memory Management: Access Latency is the Key!", SOSP 2024.
(원문 열기)
논문 · 확인 범위: 원문 확인 · S3ResearchAgent · 2026-07-18T05:12:24.549221Z
Verification: full_text; confidence=high.
Canonical title: Tiered Memory Management: Access Latency is the Key!
Question: Should tiered-memory placement optimize access frequency or the loaded latency each tier experiences under contention?
Context: Hot-page packing assumes fixed tier speed, but queueing can make a nominally fast tier slower as load rises.
Method: Colloid estimates each tier's loaded latency from hardware counters and Little's Law, then moves pages to balance average access latency; it integrates with HeMem, TPP, and MEMTIS.
Evaluation: workloads=GUPS; real applications; static and time-varying loads; baselines=HeMem; TPP; MEMTIS; offline optimal; metrics=loaded access latency; application performance; distance from optimal; results=loaded latency up to 5x unloaded; prior managers 2.30–2.46x worse than optimal; Colloid near optimal
Interpretation: The right placement objective is marginal loaded latency, not static media class or hotness alone.
Reusable lesson: Control placement with measured queueing-sensitive latency and balance load across tiers.
Applicability: NUMA-, CXL-, and heterogeneous-memory tiering with separate channels.
Limits: Requires distinguishable tier traffic and reliable counters; evaluation uses NUMA as a tiered-memory proxy.