본문으로 이동

속성:Attempt

S3 연구 메모리

Text

직접 해본 방법이나 설정을 적습니다.

( | ) (20 | 50 | 100 | 250 | 500) 보기
이 속성을 사용하는 문서 20개를 보여줍니다.
s
scale and flexibility in distribution of hot content a83ac58f +
Owl combines ephemeral peer-to-peer distribution trees in a decentralized data plane with centralized tracker services. Trackers maintain peer/cache/download metadata, choose fetch sources and retries, control caching/eviction, and expose a configurable policy interface; peers remain simple state machines.  +
scalecache a scalable page cache for multiple solid state drives 60b5c227 +
ScaleCache introduces concurrent ccXArray operations and dflush for parallel, opportunistic direct flushing.  +
scaleswap a scalable os swap system for all flash swap arrays 4041bd54 +
ScaleSwap assigns per-core swap resources, adds opportunistic inter-core metadata assistance, and maintains page/LRU affinity.  +
scalio scaling up dpu based jbof key value store with nvme of target offload 8db39764 +
Scalio offloads target functions to DPUs and uses RDMA-based coordination to keep DPU DRAM and SSD cache states linearizable.  +
scheduler support for video oriented multimedia on client side virtualization c8366be4 +
framebuffer write 또는 GPU interrupt rate로 frame rate를 추정하고, desired frame rate와의 feedback으로 VM CPU share를 동적으로 조절한다. Xen Credit scheduler의 boost를 multimedia-aware하게 정제한다.  +
scoz a systemwide causal profiler for multicore systems 7b5e72e2 +
virtual speedup 단위를 thread에서 physical core로 바꾸고 per-core handler, perf CPU-clock sampling, user/kernel mapping table, shared delay bookkeeping을 둔다. idle-core dependency와 asynchronous delay 중 thread migration을 처리하도록 Linux scheduler도 수정한다.  +
selective memory deduplication for cost efficiency in mobile smart devices 9f3139a0 +
중복될 가능성이 낮은 페이지를 시간 영역과 공간 영역에서 제외한다. 절약한 메모리는 앱 캐싱/실행시간 개선에 사용한다.  +
separate semantic replication hashes from provenance bound artifact hashes e7ab7a4e +
The B200 a03 and a04 reruns were first compared by whole-file SHA-256, then by a canonical projection containing only dataset, model, tokenizer, records sorted by item_id, and belief_model with beliefs sorted by prompt_bin.  +
serverless in the wild characterizing and optimizing the serverless workload at a large cloud pr f2fa1121 +
Azure Functions 14일 플릿을 분석하고 앱별 keep-alive와 ARIMA 기반 예측을 혼합한다.  +
serverlessllm low latency serverless inference for large language models 69b8cd6d +
ServerlessLLM uses a loading-optimized checkpoint format, multi-tier local loading, live migration, and a scheduler aware of locality and startup cost.  +
shape adaptive attention negative 20260722 +
Built global, exact-key, bucketed, clustered, regret-aware, random, and oracle policies over a matched fused/unfused Triton attention candidate set; measured three fresh L40S processes with chronological train/test splits, lifecycle costs, correctness checks, and a separate CUDA Graph replay experiment across L2-resident and spilled working sets.  +
shiftlock mitigate one sided rdma lock contention via handover a5d67e63 +
ShiftLock uses nonblocking direct client-to-client handover inspired by MCS queues, with reader-writer semantics, starvation freedom, CPU efficiency, and fault handling.  +
sieve is simpler than lru an efficient turn key eviction algorithm for web caches 86f86293 +
FIFO 큐, hand, lazy promotion, quick demotion만으로 SIEVE를 구성한다.  +
silk preventing latency spikes in log structured merge key value stores 8576916f +
SILK schedules internal I/O opportunistically, prioritizes lower LSM levels, and makes compaction preemptible in a RocksDB-derived store.  +
siloz leveraging dram isolation domains to prevent inter vm rowhammer 9d15762c +
DRAM subarray group을 VM별 isolation domain으로 배정하고 EPT 무결성을 보호한다.  +
smartlmk a memory reclamation scheme for improving user perceived app launch time d60522c6 +
App launch time과 사용 통계로 보존 가치를 모델링하고, 과거 메모리 사용으로 footprint와 임박한 memory demand를 추정해 가치가 낮은 앱들을 한 번에 종료한다.  +
specinfer accelerating generative large language model serving with tree based speculative infer e609aa25 +
SpecInfer has multiple small models construct a token tree, then verifies tree nodes in parallel with the target LLM using distribution-preserving acceptance.  +
speculative multi level access in lsm tree based kv store 72e42f29 +
키가 있을 가능성이 있는 여러 LSM level에 speculative read를 병렬 발행해 CPU filter/index 처리와 SSD I/O를 겹친다.  +
speed is all you need on device acceleration of large diffusion models via gpu aware optimizatio c53959fd +
Stable Diffusion 1.4의 group normalization/GELU specialized kernels, attention softmax/FlashAttention 경로, Winograd convolution, GPU 실행과 memory-buffer 재사용을 모바일 장치에 맞게 최적화한다.  +
ssdstreamer specializing i o stack for large scale machine learning 6b597ff1 +
DRAM을 상위 캐시로 쓰지 않고 coarse-grained prefetch용 stream buffer로 사용하며, 큰 SSD cache는 무거운 커널 I/O 경로를 우회하는 전용 user-space I/O stack 위에서 구동한다.  +
원본 주소 "https://s3wiki.yonsei.ac.kr/index.php?title=속성:Attempt&oldid=145"