]> 2026-09-08T17:23:26+09:00 Lesson:pact a criticality first design for tiered memory 20db3445 3000 ko DRAM plus NUMA/persistent/CXL memory where standard performance counters are available. Limits: Depends on Intel-style queue/performance counters and phase stability; when not best, average/max gap is 4.1%/11.8%. PACT defines per-page access criticality from four hardware counters and per-tier MLP, then uses eager demotion and adaptive promotion online. S3ResearchAgent low Venue: ASPLOS. Year: 2026. High-MLP hot pages can hide slow-tier latency, while lower-frequency pointer-chasing pages may be performance-critical. Verification: publication metadata is preserved, but the current fetched object is a landing/cookie page; claim-bearing full text must be reattached before raising confidence; confidence=low. 2026-07-16T15:03:42Z 2461238.1275694 PACT: A Criticality-First Design for Tiered Memory. ASPLOS 2026. Online page placement should optimize attributed CPU stall time rather than access count. PACT: A Criticality-First Design for Tiered Memory workloads=13 graph, HPC, in-memory-cache, and ML workloads; 96-workload model study; baselines=Soar, Alto, Memtis, Colloid, Nomad, TPP, Linux NBT; metrics=performance, migrations, model correlation; results=up to 61% faster; up to 50x fewer migrations; Pearson >0.98 Which pages deserve DRAM when access frequency fails to capture their actual CPU-stall impact? lab Estimate per-item criticality from exposed latency and parallelism, then design migration policies around its skew. Draft 2026-07-18T15:00:29Z 2461240.1253357 2026-07-18T15:00:29Z 2461240.1253357 pact a criticality first design for tiered memory 20db3445 3000 ev_445a5dc6a68e45fd 2026-07-16T15:09:01Z 2461238.1312616 S3ResearchAgent PACT: A Criticality-First Design for Tiered Memory. ASPLOS 2026. paper Bibliographic paper record. not_recorded pact a criticality first design for tiered memory 20db3445# 2c6d14ac1fa651911a69f96a73fe4c4f 3000 verified-content-v1-0159 2026-07-16T18:59:31Z 2461238.291331 S3ResearchAgent Hamid Hadian; Jinshu Liu; Hanchen Xu; Hansen Idden; Huaicheng Li. PACT: A Criticality-First Design for Tiered Memory. ASPLOS, 2026. paper Verification: author-hosted full paper including abstract, introduction, evaluation, and conclusion; ...; confidence=high. Canonical title: PACT: A Criticality-First Design for Tiered Memory Question: Which pages deserve DRAM when access frequency fails to capture their actual CPU-stall impact? Context: High-MLP hot pages can hide slow-tier latency, while lower-frequency pointer-chasing pages may be performance-critical. Method: PACT defines per-page access criticality from four hardware counters and per-tier MLP, then uses eager demotion and adaptive promotion online. Evaluation: workloads=13 graph, HPC, in-memory-cache, and ML workloads; 96-workload model study; baselines=Soar, Alto, Memtis, Colloid, Nomad, TPP, Linux NBT; metrics=performance, migrations, model correlation; results=up to 61% faster; up to 50x fewer migrations; Pearson >0.98 Interpretation: Online page placement should optimize attributed CPU stall time rather than access count. Reusable lesson: Estimate per-item criticality from exposed latency and parallelism, then design migration policies around its skew. Applicability: DRAM plus NUMA/persistent/CXL memory where standard performance counters are available. Limits: Depends on Intel-style queue/performance counters and phase stability; when not best, average/max gap is 4.1%/11.8%. not_recorded pact a criticality first design for tiered memory 20db3445# f37e0d3e9ff3541d9b928d00ab954e4a 3000 canonical-paper-v2-20db3445 2026-07-18T05:32:57Z 2461239.7312153 S3ResearchAgent Hamid Hadian; Jinshu Liu; Hanchen Xu; Hansen Idden; Huaicheng Li. PACT: A Criticality-First Design for Tiered Memory. ASPLOS, 2026. paper Verification: author-hosted full paper including abstract, introduction, evaluation, and conclusion; .; confidence=high. Canonical title: PACT: A Criticality-First Design for Tiered Memory Question: Which pages deserve DRAM when access frequency fails to capture their actual CPU-stall impact? Context: High-MLP hot pages can hide slow-tier latency, while lower-frequency pointer-chasing pages may be performance-critical. Method: PACT defines per-page access criticality from four hardware counters and per-tier MLP, then uses eager demotion and adaptive promotion online. Evaluation: workloads=13 graph, HPC, in-memory-cache, and ML workloads; 96-workload model study; baselines=Soar, Alto, Memtis, Colloid, Nomad, TPP, Linux NBT; metrics=performance, migrations, model correlation; results=up to 61% faster; up to 50x fewer migrations; Pearson >0.98 Interpretation: Online page placement should optimize attributed CPU stall time rather than access count. Reusable lesson: Estimate per-item criticality from exposed latency and parallelism, then design migration policies around its skew. Applicability: DRAM plus NUMA/persistent/CXL memory where standard performance counters are available. Limits: Depends on Intel-style queue/performance counters and phase stability; when not best, average/max gap is 4.1%/11.8%. full_text pact a criticality first design for tiered memory 20db3445# 33c20aba40e28e02efcd94eeac23dcbd 3000 2026-07-18T14:58:49Z 2461240.1241782 S3ResearchAgent verify_18c92b8971442dad3a2e metadata_only context 보존 객체는 cookie/landing page이므로 서지 위치만 확인했고 본문 주장을 검증하지 못함. inconclusive ef5ce107b3077738d2351e4a37f6ee434e2d7e8c1fbcfd07da607633fd349337 R2-RESTIC:7f893ca5afd2cfb6fe320e9b61063ccc70e75a7a96589420038c8cf338b273be; archive-manifest-sha256=e28171fb69e141ce306d92dfe4b10e6cdc6e81d4fa910c30a846204dbcf8edf8; sha256=ef5ce107b3077738d2351e4a37f6ee434e2d7e8c1fbcfd07da607633fd349337 보존 파일 objects/sha256/ef/ef5ce107b3077738d2351e4a37f6ee434e2d7e8c1fbcfd07da607633fd349337 ev_445a5dc6a68e45fd b4e4523d3da75ac24edffcd594529f93bd24bf90d95c2918775c95e8b345af4d pact a criticality first design for tiered memory 20db3445# 268d363feb23ec727409dfe064aaf48a 3000 2026-07-18T14:58:50Z 2461240.1241898 S3ResearchAgent verify_9ded199ede15752fa2ca metadata_only context 보존 객체는 cookie/landing page이므로 서지 위치만 확인했고 본문 주장을 검증하지 못함. inconclusive ef5ce107b3077738d2351e4a37f6ee434e2d7e8c1fbcfd07da607633fd349337 R2-RESTIC:7f893ca5afd2cfb6fe320e9b61063ccc70e75a7a96589420038c8cf338b273be; archive-manifest-sha256=e28171fb69e141ce306d92dfe4b10e6cdc6e81d4fa910c30a846204dbcf8edf8; sha256=ef5ce107b3077738d2351e4a37f6ee434e2d7e8c1fbcfd07da607633fd349337 보존 파일 objects/sha256/ef/ef5ce107b3077738d2351e4a37f6ee434e2d7e8c1fbcfd07da607633fd349337 verified-content-v1-0159 711f71d76bab91cde0bdc6b67bf3ada3708e2d3f45bd4316323c4e15707a82b0 pact a criticality first design for tiered memory 20db3445# 8cb1d7d21e5fa12f10649053d48a9f78 3000 2026-07-18T14:58:50Z 2461240.1241898 S3ResearchAgent verify_2fccf30452746a8f7c4a metadata_only context 보존 객체는 cookie/landing page이므로 서지 위치만 확인했고 본문 주장을 검증하지 못함. inconclusive ef5ce107b3077738d2351e4a37f6ee434e2d7e8c1fbcfd07da607633fd349337 R2-RESTIC:7f893ca5afd2cfb6fe320e9b61063ccc70e75a7a96589420038c8cf338b273be; archive-manifest-sha256=e28171fb69e141ce306d92dfe4b10e6cdc6e81d4fa910c30a846204dbcf8edf8; sha256=ef5ce107b3077738d2351e4a37f6ee434e2d7e8c1fbcfd07da607633fd349337 보존 파일 objects/sha256/ef/ef5ce107b3077738d2351e4a37f6ee434e2d7e8c1fbcfd07da607633fd349337 canonical-paper-v2-20db3445 98e90911e93239628a1350df54b726ae75d6fdc709c27c4a86fc71f03e86b4f0 pact a criticality first design for tiered memory 20db3445# 1b1f0465d8c1664ea5893e629631f73c 3000 2026-07-18T15:00:28Z 2461240.1253241 S3ResearchAgent verify_0762ca9191c1fdaf3014 metadata_only observation,interpretation,reusable_lesson claim-bearing O/I/R coverage was not established; confidence forced to low inconclusive ef5ce107b3077738d2351e4a37f6ee434e2d7e8c1fbcfd07da607633fd349337 R2-RESTIC:7f893ca5afd2cfb6fe320e9b61063ccc70e75a7a96589420038c8cf338b273be; archive-manifest-sha256=e28171fb69e141ce306d92dfe4b10e6cdc6e81d4fa910c30a846204dbcf8edf8; sha256=ef5ce107b3077738d2351e4a37f6ee434e2d7e8c1fbcfd07da607633fd349337; high-confidence adjudication ledger Saved HTML inspected in full; it is a cookie/landing/generic metadata page and contains no claim-bearing text covering O/I/R. canonical-paper-v2-20db3445 98e90911e93239628a1350df54b726ae75d6fdc709c27c4a86fc71f03e86b4f0 pact a criticality first design for tiered memory 20db3445# 795e83fc25abca253dcbbeaa1bc919a3 3000 2026-07-18T15:00:29Z 2461240.1253357 S3ResearchAgent verify_aac4ed2d53c8d1fe78ab metadata_only context CD-009 curated source-bound correction; primary locator recorded in review. inconclusive a140f3572e510459b9eed5bf42d4e2096c4684fd4f1ee746960d6675661d1077 R2-RESTIC:7f893ca5afd2cfb6fe320e9b61063ccc70e75a7a96589420038c8cf338b273be; archive-manifest-sha256=e28171fb69e141ce306d92dfe4b10e6cdc6e81d4fa910c30a846204dbcf8edf8; sha256=a140f3572e510459b9eed5bf42d4e2096c4684fd4f1ee746960d6675661d1077; adjudicated correction content-debt-review CD-009 canonical-paper-v2-20db3445 98e90911e93239628a1350df54b726ae75d6fdc709c27c4a86fc71f03e86b4f0 pact a criticality first design for tiered memory 20db3445# cad12002289d246fafb284e02081d5cb