본문으로 이동

속성:Reusable lesson

S3 연구 메모리

Text

비슷한 문제를 만났을 때 다시 쓸 수 있는 내용을 적습니다.

( | ) (20 | 50 | 100 | 250 | 500) 보기
이 속성을 사용하는 문서 20개를 보여줍니다.
s
scale and flexibility in distribution of hot content a83ac58f +
Centralizing control decisions need not prevent data-plane scalability when bulk transfer remains peer-to-peer.  +
scalecache a scalable page cache for multiple solid state drives 60b5c227 +
Remove coordination bottlenecks from both the foreground index and background drain path.  +
scaleswap a scalable os swap system for all flash swap arrays 4041bd54 +
Pair per-core fast paths with opportunistic assistance instead of global work queues.  +
scalio scaling up dpu based jbof key value store with nvme of target offload 8db39764 +
Place control and data processing near the storage fabric, with an explicit consistency protocol.  +
scheduler support for video oriented multimedia on client side virtualization c8366be4 +
애플리케이션 내부 신호가 없어도 low-level device event를 QoS proxy metric으로 만들고 feedback scheduling에 사용할 수 있다.  +
scoz a systemwide causal profiler for multicore systems 7b5e72e2 +
causal profiling의 범위를 넓히려면 프로세스가 아니라 CPU core를 공통 관측·지연 단위로 삼되 idle dependency와 migration semantics를 함께 고쳐야 한다.  +
selective memory deduplication for cost efficiency in mobile smart devices 9f3139a0 +
모바일 프레임워크와 앱 메모리 배치의 구조적 규칙을 이용하면 content-based 메모리 관리의 탐색공간을 줄일 수 있다.  +
separate semantic replication hashes from provenance bound artifact hashes e7ab7a4e +
Keep two hashes with distinct contracts: bind inventories and archives to the complete artifact bytes, and use a versioned canonical semantic projection only for cross-run replication claims. Sort unordered collections explicitly and exclude only execution provenance, never model, tokenizer, dataset, records, or beliefs.  +
serverless in the wild characterizing and optimizing the serverless workload at a large cloud pr f2fa1121 +
단일 고정 keep-alive보다 워크로드별 하이브리드 정책이 낫다.  +
serverlessllm low latency serverless inference for large language models 69b8cd6d +
Co-design persistent layout, cache hierarchy, migration, and placement policy for large cold state.  +
shape adaptive attention negative 20260722 +
Before implementing an adaptive GPU kernel portfolio, first measure the global-to-oracle gap on a correctness-matched, time-ordered trace and require that it exceed compile, cache, dispatch, and switching costs. Separate policy gain from fusion gain with a policy-by-fusion factorial. Measure eager and CUDA Graph modes independently, because launch-elimination benefits are not portable across them. Calibrate roofline constants only on code that actually reaches the assumed memory regime, and model L2 residency before charging HBM traffic. If the global kernel is already within about 1% of oracle and graph replay collapses the fusion margin, stop adaptive-policy work and preserve the result as a bounded negative finding.  +
shiftlock mitigate one sided rdma lock contention via handover a5d67e63 +
Under high contention, hand work directly to a known successor instead of reopening global competition.  +
sieve is simpler than lru an efficient turn key eviction algorithm for web caches 86f86293 +
생산 캐시 정책은 코드 복잡성과 lock 비용도 1급 지표로 보라.  +
silk preventing latency spikes in log structured merge key value stores 8576916f +
Make background storage work preemptible and prioritize tasks that avert future write stalls.  +
siloz leveraging dram isolation domains to prevent inter vm rowhammer 9d15762c +
보안 격리를 물리 메모리 고장 도메인과 정렬하라.  +
smartlmk a memory reclamation scheme for improving user perceived app launch time d60522c6 +
Reclamation victim은 recency 하나가 아니라 재실행 비용, 사용 확률, 회수량, 앞으로 필요한 메모리를 함께 보아야 한다.  +
specinfer accelerating generative large language model serving with tree based speculative infer e609aa25 +
When cheap predictors are uncertain, aggregate diverse candidates and verify them in one expensive parallel pass.  +
speculative multi level access in lsm tree based kv store 72e42f29 +
빠른 병렬 SSD에서는 순차 의존 탐색보다 여러 후보를 제한적으로 추측 실행해 CPU와 I/O를 겹치는 편이 유리할 수 있다.  +
speed is all you need on device acceleration of large diffusion models via gpu aware optimizatio c53959fd +
온디바이스 diffusion은 end-to-end 지연만 보지 말고 kernel별 latency, intermediate tensor memory, initialized weight memory를 함께 측정하고 장치별 최적화 효과를 분리한다.  +
ssdstreamer specializing i o stack for large scale machine learning 6b597ff1 +
순차·반복 접근을 예측할 수 있다면 작은 DRAM을 범용 cache가 아니라 SSD를 지속적으로 공급하는 stream buffer로 사용하는 편이 유리할 수 있다.  +
원본 주소 "https://s3wiki.yonsei.ac.kr/index.php?title=속성:Reusable_lesson&oldid=149"