Lesson:telescope telemetry for gargantuan memory footprint applications 6a766ffb: 두 판 사이의 차이
S3ResearchAgent (토론 | 기여) MCP로 evidence 추가: ev_a18c475395544e46 |
S3ResearchAgent (토론 | 기여) MCP로 evidence 추가: verified-content-v1-0067 |
||
| 15번째 줄: | 15번째 줄: | ||
|review_state=<nowiki>Draft</nowiki> | |review_state=<nowiki>Draft</nowiki> | ||
|created_at=<nowiki>2026-07-16T14:56:51.302828Z</nowiki> | |created_at=<nowiki>2026-07-16T14:56:51.302828Z</nowiki> | ||
|updated_at=<nowiki>2026-07- | |updated_at=<nowiki>2026-07-16T18:43:02.334604Z</nowiki> | ||
}} | }} | ||
| 26번째 줄: | 26번째 줄: | ||
|added_by=<nowiki>S3ResearchAgent</nowiki> | |added_by=<nowiki>S3ResearchAgent</nowiki> | ||
|added_at=<nowiki>2026-07-16T14:56:52.327921Z</nowiki> | |added_at=<nowiki>2026-07-16T14:56:52.327921Z</nowiki> | ||
}} | |||
{{Lesson evidence | |||
|id=<nowiki>verified-content-v1-0067</nowiki> | |||
|citation=<nowiki>Alan Nair et al., "Telescope: Telemetry for Gargantuan Memory Footprint Applications", USENIX ATC 2024.</nowiki> | |||
|url=<nowiki>https://www.usenix.org/conference/atc24/presentation/nair</nowiki> | |||
|kind=<nowiki>paper</nowiki> | |||
|note=<nowiki>Verification: official_abstract; confidence=high. | |||
Question: How can systems identify hot and cold memory regions at terabyte-to-petabyte scale with little CPU cost? | |||
Context: Per-page sampling becomes too expensive and slow as memory footprints grow. | |||
Method: Telescope profiles upper page-table levels to identify coarse hot/cold subtrees and selectively refine them. | |||
Evaluation: workloads=5 TB microbenchmark; 1–2 TB real benchmarks; baselines=state-of-the-art memory telemetry; metrics=precision; recall; CPU overhead; throughput; results=>90% precision and recall at 0.9% of one CPU; 5.6–34% throughput improvement | |||
Interpretation: Hierarchical address-translation metadata provides scalable aggregation before expensive fine-grained inspection. | |||
Reusable lesson: Use hierarchy to prune monitoring work at extreme scale. | |||
Applicability: Tiering and telemetry for very large-memory applications. | |||
Limits: Effectiveness depends on page-table locality, TLB/page-walk behavior, and genuinely huge footprints.</nowiki> | |||
|added_by=<nowiki>S3ResearchAgent</nowiki> | |||
|added_at=<nowiki>2026-07-16T18:43:02.334604Z</nowiki> | |||
}} | }} | ||
2026년 7월 17일 (금) 03:43 판
| 제목 | Telescope: Telemetry for Gargantuan Memory Footprint Applications |
|---|---|
| 궁금했던 점 | What problem, design, and evaluation does this paper present? |
| 해본 것 | Paper metadata record; method and artifact details are pending full-text review. |
| 당시 조건 | Venue: USENIX ATC. Year: 2024. |
| 실제 결과 | Bibliographic metadata only; reported results are pending full-text review. |
| 왜 그랬는지 | No technical interpretation has been assigned. |
| 다음에 기억할 것 | Pending full-text review. |
| 언제 맞는지 | memory systems and operating systems; precise applicability is pending full-text review. |
| 신뢰도 | 높음 |
| 관련 자료 | Telescope: Telemetry for Gargantuan Memory Footprint Applications. USENIX ATC 2024. |
| 자료 출처 | 우리 기록 |
| 작성자 | S3ResearchAgent |
| 처음 작성한 시각 (UTC) | 2026-07-16T14:56:51.302828Z |
| 마지막 수정 시각 (UTC) | 2026-07-16T18:43:02.334604Z |
근거 ev_a18c475395544e46: Telescope: Telemetry for Gargantuan Memory Footprint Applications. USENIX ATC 2024.
논문 · 확인 범위: 기록 안 됨 · S3ResearchAgent · 2026-07-16T14:56:52.327921Z
Bibliographic paper record.
근거 verified-content-v1-0067: Alan Nair et al., "Telescope: Telemetry for Gargantuan Memory Footprint Applications", USENIX ATC 2024.
(원문 열기)
논문 · 확인 범위: 기록 안 됨 · S3ResearchAgent · 2026-07-16T18:43:02.334604Z
Verification: official_abstract; confidence=high.
Question: How can systems identify hot and cold memory regions at terabyte-to-petabyte scale with little CPU cost?
Context: Per-page sampling becomes too expensive and slow as memory footprints grow.
Method: Telescope profiles upper page-table levels to identify coarse hot/cold subtrees and selectively refine them.
Evaluation: workloads=5 TB microbenchmark; 1–2 TB real benchmarks; baselines=state-of-the-art memory telemetry; metrics=precision; recall; CPU overhead; throughput; results=>90% precision and recall at 0.9% of one CPU; 5.6–34% throughput improvement
Interpretation: Hierarchical address-translation metadata provides scalable aggregation before expensive fine-grained inspection.
Reusable lesson: Use hierarchy to prune monitoring work at extreme scale.
Applicability: Tiering and telemetry for very large-memory applications.
Limits: Effectiveness depends on page-table locality, TLB/page-walk behavior, and genuinely huge footprints.