속성으로 검색

이 문서는 속성과 이름이 지정된 값으로 설명된 개체를 찾기 위한 간단한 탐색 인터페이스를 제공합니다. 그 밖에 사용 가능한 검색 인터페이스에는 문서 속성 검색 및 Ask 쿼리 빌더가 있습니다.

결과 목록

• Lesson:smartlmk a memory reclamation scheme for improving user perceived app launch time d60522c6  + (Cached-process 기반 모바일 lifecycle 관리에 직접 해당한다.)
• Lesson:technical review transparently exploiting device reserved memory for application performance in d2f8d46b  + (Camera, codec, GPU용 contiguous reserved memory가 큰 mobile 또는 embedded system에 해당한다.)
• Lesson:transparently exploiting device reserved memory for application performance in mobile systems 43e55454  + (Camera, codec, GPU용 contiguous reserved memory가 큰 mobile 또는 embedded system에 해당한다.)
• Lesson:technical review enlightening the i o path a holistic approach for application performance dabf0f9c  + (Checkpoint, journaling, compaction이 foreground request와 경합하는 database 및 key-value store에 높다.)
• Lesson:enlightening the i o path a holistic approach for application performance a3c2b020  + (Checkpoint, journaling, compaction이 foreground request와 경합하는 database 및 key-value store에 높다.)
• Lesson:technical review controlling physical memory fragmentation in mobile systems 863971bc  + (Codec과 camera 등 큰 contiguous DMA buffer가 필요한 mobile 또는 embedded kernel에 해당한다.)
• Lesson:controlling physical memory fragmentation in mobile systems 0365acff  + (Codec과 camera 등 큰 contiguous DMA buffer가 필요한 mobile 또는 embedded kernel에 해당한다.)
• Lesson:technical review dram architecture for efficient data lifetime management 0f2dfdbe  + (Cold/warm boot 또는 deallocated-page disclosure를 막아야 하는 DRAM 설계에 해당한다.)
• Lesson:dram architecture for efficient data lifetime management f27c8820  + (Cold/warm boot 또는 deallocated-page disclosure를 막아야 하는 DRAM 설계에 해당한다.)
• Lesson:how to copy memory coordinated asynchronous copy as a first class os service 8e43b859  + (Copy-heavy file, network, IPC, and mobile OS paths. Limits: Applications/subsystems must express asynchronous copy dependencies; hardware availability affects benefit.)
• Lesson:kvcache cache in the wild characterizing and optimizing kvcache cache at a large cloud provider 0b39b53c  + (Cross-request prefix/KV reuse가 있는 shared LLM serving cluster에 적용합니다. Limits: 한 cloud provider의 두 production 기간을 분석했고 논문은 각 trace의 대표 하루 결과를 주로 제시합니다. Reasoning workload와 prefix 이외의 KV reuse, global scheduling, cache fairness는 평가 범위 밖입니다.)
• Lesson:moat securely mitigating rowhammer with per row activation counters 6e5ed0e1  + (DDR5 PRAC/ABO RowHammer mitigation designs. Limits: Security bound depends on timing and threshold assumptions; evaluation is simulation/workload based and must track JEDEC/device implementation details.)
• Lesson:predicting future system reliability with a component level dram fault model 1f248bb6  + (DRAM architecture, ECC, repair, and future-system reliability planning. Limits: Predictions extrapolate public production logs through assumed component mappings and future-device parameters.)
• Lesson:pact a criticality first design for tiered memory 20db3445  + (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%.)
• Lesson:tiered memory management beyond hotness 6ad152df  + (DRAM plus slower NUMA/CXL/persistent-memory tiers. Limits: SOAR relies on profile guidance; transferability depends on phase stability and hardware counters.)
• Lesson:breakhammer enhancing rowhammer mitigations by carefully throttling suspect threads 19eb947b  + (DRAM systems deploying preventive RowHammer mitigations. Limits: Relies on correct culprit identification and simulated mitigation models; benign high-activity threads can be throttled.)
• Lesson:understanding and mitigating covert and side channel vulnerabilities introduced by rowhammer def 0e272653  + (DRAM systems deploying stateful or throttling-based RowHammer defenses. Limits: Channel feasibility and mitigation overhead depend on the specific defense and platform; arXiv versions report different older rates.)
• Lesson:optimizing file systems on heterogeneous memory by integrating dram cache with virtual memory ma aaa7aa17  + (DRAM+PM 계층형 로컬 파일시스템. Limits: 특정 PM/DRAM 구성과 library FS 통합을 요구한다.)
• Lesson:fully harnessing the performance potential of dram less mobile flash storage ff995292  + (DRAM-less UFS/eMMC 설계와 host-managed flash metadata 가속에 높음; 상용 적용에는 firmware 변경 필요.)
• Lesson:technical review fully harnessing the performance potential of dram less mobile flash storage d4449ca9  + (DRAM-less UFS/eMMC 설계와 host-managed flash metadata 가속에 높음; 상용 적용에는 firmware 변경 필요.)
• Lesson:nomad non exclusive memory tiering via transactional page migration 3990fa86  + (DRAM/CXL or DRAM/PMEM tiered-memory systems under pressure. Limits: Shadow pages consume slow-tier capacity and require transactional coherence; gains vary with pressure and reuse.)
• Lesson:technical review ssdstreamer specializing i o stack for large scale machine learning 5d1f4a63  + (DRAM보다 큰 반복 ML working set과 Spark류 데이터 분석 프레임워크에 해당한다.)
• Lesson:ssdstreamer specializing i o stack for large scale machine learning 6b597ff1  + (DRAM보다 큰 반복 ML working set과 Spark류 데이터 분석 프레임워크에 해당한다.)
• Lesson:extmem enabling application aware virtual memory management for data intensive applications 8a6dc0e4  + (Data-intensive Linux applications needing custom allocation, migration, or reclamation. Limits: Applications must supply useful policy semantics; correctness and benefit remain policy- and backend-dependent.)
• Lesson:crossprefetch accelerating i o prefetching for modern storage 59296dd1  + (Data-intensive runtimes over local or remote file systems. Limits: Requires OS/runtime integration and sufficiently predictable access patterns; aggressive prefetch can waste bandwidth.)
• Lesson:disaggregated raid storage in modern datacenters 4eb679fe  + (Datacenter disaggregated storage and object stores. Limits: Benefits depend on topology, network balance, failure model, reconstruction traffic, and object-store integration.)

• Lesson:a fully associative tagless dram cache f6383680  + (Die-stacked 또는 in-package DRAM cache architecture에 해당한다.)
• Lesson:technical review a fully associative tagless dram cache 8377cc32  + (Die-stacked 또는 in-package DRAM cache architecture에 해당한다.)

• Lesson:calcspar a contract aware lsm store for cloud storage with low latency spikes accc7c9d  + (EBS 위 LSM KV/DB. Limits: EBS와 선정 워크로드 중심이며 캐시·IOPS 정책의 이식성은 미검증이다.)
• Lesson:exoflow a universal workflow system for exactly once dags d9d62c90  + (ETL, ML 파이프라인, stateful serverless DAG. Limits: annotation 정확성과 외부 부작용의 멱등성·커밋 가정에 의존한다.)
• Lesson:accl an fpga based collective engine for distributed applications 896d30fd  + (FPGA 클러스터 HPC·분산 추론. Limits: FPGA/AMD toolchain과 평가 클러스터 규모에 제약되며 초록에 정확한 수치가 없다.)
• Lesson:serverless in the wild characterizing and optimizing the serverless workload at a large cloud pr f2fa1121  + (FaaS 용량 계획과 콜드 스타트 완화. Limits: 단일 공급자·14일 관측이며 초록에서 정확한 개선 수치는 확인되지 않는다.)
• Lesson:technical review request aware cooperative i o scheduling for scale out database applications fe4a3e09  + (Fan-out 또는 sharded read 요청을 수행하는 scale-out database에 해당한다.)
• Lesson:request aware cooperative i o scheduling for scale out database applications 25327d2c  + (Fan-out 또는 sharded read 요청을 수행하는 scale-out database에 해당한다.)
• Lesson:avoiding read stalls on flash storage 34c6809a  + (Flash-backed databases and buffer caches under dirty-page pressure. Limits: Requires extra DRAM and careful ordering/recovery; value falls when dirty-victim stalls are rare.)
• Lesson:ugache a unified gpu cache for embedding based deep learning 227e7c8f  + (GNN 학습과 추천 임베딩 추론. Limits: 읽기 전용·편향·예측 가능한 임베딩 접근과 단일 노드 다중 GPU를 전제한다.)
• Lesson:pod attention unlocking full prefill decode overlap for faster llm inference caae5b3e  + (GPU LLM serving that mixes ongoing decode with new prompt prefill. Limits: Requires custom attention kernels and careful resource partitioning; gains depend on prefill/decode mix.)
• Lesson:nanoflow towards optimal large language model serving throughput f0633332  + (GPU LLM serving with predictable operator profiles. Limits: May duplicate work and requires model/hardware-specific profiling and scheduling.)
• Lesson:asynchrony and gpus bridging this dichotomy for i o with agio 6fab41bf  + (GPU applications with direct SSD access and overlapable computation. Limits: Requires GPU/SSD integration and enough independent work; proactive access is easier for predictable patterns.)
• Lesson:technical review managing gpu buffers for caching more apps in mobile systems 46ece2ab  + (GPU buffer가 system memory를 공유하고 background app caching이 중요한 mobile SoC에 해당한다.)
• Lesson:managing gpu buffers for caching more apps in mobile systems afba2ee0  + (GPU buffer가 system memory를 공유하고 background app caching이 중요한 mobile SoC에 해당한다.)
• Lesson:shape adaptive attention negative 20260722  + (GPU serving and compiler projects that cho … GPU serving and compiler projects that choose among shape-specialized kernels, especially attention, RoPE, and other launch-sensitive operations under CUDA Graphs. Quantitative bounds are limited to one L40S, five Triton configurations, one controlled GQA compatibility partition, and a PyTorch CUDA Graph proxy; a materially different native candidate set or real trace could reopen the question only if it first demonstrates substantially larger global-to-oracle headroom.antially larger global-to-oracle headroom.)
• Lesson:dwkv kv saturation unproven  + (GPU serving profiling, KV-cache scheduling, simulator validation, bottleneck attribution)
• Lesson:technical review gpu e130421c  + (GPU 인과관계 profiling과 최적화 후보 예측 검토에 한정합니다. 현재 서지 자료는 성능 향상 주장을 지지하지 않습니다.)
• Lesson:scalaafa constructing user space all flash array engine with holistic designs 396837bd  + (High-performance user-space all-flash arrays. Limits: The design assumes SSD offload capabilities and its evaluated hardware/software stack.)
• Lesson:ic cache efficient large language model serving via in context caching 4f05d8b7  + (High-volume LLM services with recurrent intents or tasks. Limits: Requires representative history and safe example selection; novel or distribution-shifted queries may not benefit.)
• Lesson:understanding silent data corruptions in a large production cpu population 56bebc4c  + (Hyperscale CPU qualification, fleet monitoring, and SDC mitigation. Limits: Production data and CPU population are provider-specific; sensitive rates and all sampling biases are not exposed in the abstract.)
• Lesson:aegonkv a high bandwidth low tail latency and low storage cost kv separated lsm store with smart 5a081f42  + (KV-separated LSM stores on computational/SmartSSD platforms. Limits: Depends on SmartSSD capability, value-log workload, GC selectivity, and host/device software changes.)
• Lesson:impress an importance informed multi tier prefix kv storage system for large language model infe 51b47015  + (LLM applications with recurring long contexts and disk-backed prefix caches. Limits: Accuracy and speed depend on importance stability across heads/models/tasks, prefix reuse, disk bandwidth, and tier capacity.)
• Lesson:dwkv prompt seed bug  + (LLM serving benchmark, stochastic workload replay, multi-policy paired comparison, seed-repeat 분석)
• Lesson:constructing and analyzing the lsm compaction design space 1aa4ace5  + (LSM key-value stores and compaction-policy design. Limits: The explored strategies and workloads do not exhaust the design space, and the study does not provide an online adaptive selector.)
• Lesson:silk preventing latency spikes in log structured merge key value stores 8576916f  + (LSM key-value stores with latency-sensitive foreground traffic. Limits: Results target the evaluated RocksDB-derived implementation and storage/workload mixes.)