USENIX ATC '23
2023-07-10..2023-07-12
Boston, United States
as the organisers put it
2023 USENIX Annual Technical Conference will take place July 10–12, 2023, at the Sheraton Boston, Boston, MA, USA. USENIX ATC '23 will bring together leading systems researchers for cutting-edge systems research and the opportunity to gain insight into a wealth of must-know topics.
| title | author | length |
|---|---|---|
| Sky Computing | Ion Stoica | 52m |
Sky Computing proposes an intercloud broker that maps application requirements to services across clouds; prototypes Sky Pilot (ML job broker using Ray) and… | ||
| Bifrost: Analysis and Optimization of Network I/O Tax in Confidential Virtual... | Dingji Li, Dingji Li, Zeyu Mi, Chenhui Ji, Yifan Tan, Binyu Zang, Haibing Guan, Haibo Chen | 19m |
Existing confidential VMs lose CPU to a measurable CVM-IO tax that can consume >50% of vCPU cycles; this work builds Bifrost, a para-virtualized I/O prototype… | ||
| SecretFlow-SPU: A Performant and User-Friendly Framework for Privacy-Preserving... | Junming Ma, Yancheng Zheng, Jun Feng, Derun Zhao, Haoqi Wu, Wenjing Fang, Jin Tan, Chaofan Yu, Benyu Zhang, Benyu Zhang, Lei Wang | 23m |
SecretFlow‑SPU compiles XLA frontend graphs into an MPC‑specific IR and runs them on a high‑performance runtime that supports pluggable MPC protocols; compiler… | ||
| Portunus: Re-imagining Access Control in Distributed Systems | Watson Ladd, Tanya Verma, Marloes Venema, Armando Faz-Hernández, Brendan McMillion, Avani Wildani, Nick Sullivan | 22m |
Portunus applies ciphertext-policy ABE to encrypt TLS private keys under customer-chosen locality policies, issues per-server attribute secret keys from a core… | ||
| GLogS: Interactive Graph Pattern Matching Query At Large Scale | Longbin Lai, Yufan Yang, Zhibin Wang, Yuxuan Liu, Haotian Ma, Sijie Shen, Bingqing Lyu, Xiaoli Zhou, Wenyuan Yu, Zhengping Qian, Chen Tian, Sheng Zhong, Yeh-Ching Chung, Jingren Zhou | 24m |
GLogS is a distributed system that compiles declarative graph pattern queries (supports Cypher-like languages) into dataflow programs, builds a GLog catalog of… | ||
| Cyclosa: Redundancy-Free Graph Pattern Mining via Set Dataflowi | Chuangyi Gui, Xiaofei Liao, Long Zheng, Hai Jin | 19m |
Cyclosa implements a set dataflow prototype that rewrites pattern-centric mining into directed set operators (generator, combiner, reducer), runs on a single… | ||
| SOWalker: An I/O-Optimized Out-of-CoreGraph Processing System for Second-Order... | Yutong Wu, Zhan Shi, Shicai Huang, Zhipeng Tian, Pengwei Zuo, Peng Fang, Dan Feng, Fang Wang | 18m |
SOWalker is an out‑of‑core implementation that cuts block I/O to 6–41% of GraphWalker and 8–73% of GraSorw and achieves up to 10.2× speedup over GraphWalker… | ||
| Light-Dedup: A Light-weight Inline Deduplication Framework for Non-Volatile... | Jiansheng Qiu, Yanqi Pan, Wen Xia, Xiaojia Huang, Wenjun Wu, Xiangyu Zou, Shiyi Li, Shiyi Li, Yu Hua | 17m |
Light-Dedup is a lightweight inline deduplication framework for NVM file systems that combines a non-cryptographic hash with speculative prefetch–driven… | ||
| TiDedup: A New Distributed Deduplication Architecture for Ceph4 | Myoungwon Oh, Sungmin Lee, Sungmin Lee, Samuel Just, Young Jin Yu, Duck-Ho Bae, Sage Weil, Sangyeun Cho, Heon Y. Yeom | 20m |
TiDedup introduces a Ceph-integrated deduplication design that claims up to 34% data reduction, improves foreground I/O throughput by about 50% during… | ||
| LoopDelta: Embedding Locality-aware Opportunistic Delta Compression in Inline... | Yucheng Zhang, Yucheng Zhang, Hong Jiang, Dan Feng, Nan Jiang, Taorong Qiu, Wei Huang | 22m |
LoopDelta embeds delta compression into inline deduplication and implements four concrete techniques: dual-locality-based similarity tracking, locality-aware… | ||
| TC-GNN: Bridging Sparse GNN Computation and Dense Tensor Cores on GPUs | Yuke Wang, Boyuan Feng, Zheng Wang, Guyue Huang, Yufei Ding | 18m |
TC-GNN implements a GPU pipeline that turns irregular GNN sparse ops into dense tiles processed on tensor cores, combining sparse graph translation with a… | ||
| Legion: Automatically Pushing the Envelope of Multi-GPU System for Billion-Scale... | Jie Sun, Li Su, Zuocheng Shi, Wenting Shen, Zeke Wang, Lei Wang, Jie Zhang, Yong Li, Wenyuan Yu, Jingren Zhou, fei wu | 20m |
Claims an end-to-end multi-GPU system that trains billion‑node GNNs on a single machine by combining hierarchical graph partitioning to minimize cache… | ||
| Bridging the Gap between Relational OLTP and Graph-based OLAP | Sijie Shen, Zihang Yao, Lin Shi, Lei Wang, Longbin Lai, Qian Tao, Li Su, Rong Chen, Wenyuan Yu, Haibo Chen, Binyu Zang, Jingren Zhou | 20m |
GART extends HTAP into HTGAP by adding a relational-to-graph mapping layer and an in-memory dynamic graph store that claims tens of milliseconds freshness and… | ||
| Comosum: An Extensible, Reconfigurable, and Fault-Tolerant IoT Platform for... | Gloire Rubambiza, Shiang-Wan Chin, Mueed Rehman, Sachille Atapattu, José F. Martínez, Hakim Weatherspoon | 20m |
Comosum implements an extensible, reconfigurable, fault-tolerant stack called FarmBIOS and runs as a software-defined farm across edge, core, and vendor… | ||
| Oakestra: A Lightweight Hierarchical Orchestration Framework for Edge Computing | Giovanni Bartolomeo, Mehdi Yosofie, Simon Bäurle, Oliver Haluszczynski, Nitinder Mohan, Jorg Ott | 20m |
Oakestra is a three-tier hierarchical orchestration that aggregates clusters, delegates placement to cluster schedulers, and runs a tiny node engine, net… | ||
| Explore Data Placement Algorithm for Balanced Recovery Load Distribution | Yingdi Shan, Kang Chen, Kang Chen, Yongwei Wu | 10m |
Defines optimal recovery-load placement as NP-hard and gives a concrete, implementable greedy data placement algorithm that picks nodes with minimal current… | ||
| Luci: Loader-based Dynamic Software Updates for Off-the-shelf Shared Objects | Bernhard Heinloth, Peter Wägemann, Wolfgang Schröder-Preikschat | 19m |
Luci implements a linker/loader that performs dynamic software updates for off-the-shelf shared objects by exploiting PIC relocations and GOT patching, using… | ||
| MELF: Multivariant Executables for a Heterogeneous World | Dominik Töllner, Christian Dietrich, Illia Ostapyshyn, Florian Rommel, Daniel Lohmann | 19m |
Provides multiple per-function compile-time variants inside one ELF by using a modified LLVM linker plus a runtime overlay manager that swaps same-size… | ||
| APRON: Authenticated and Progressive System Image Renovation | Sangho Lee | 21m |
APRON performs authenticated, progressive repair of a corrupt or outdated system image by verifying each requested block in-kernel with a signed Merkle hash… | ||
| zpoline: a system call hook mechanism based on binary rewriting | Kenichi Yasukata, Hajime Tazaki, Pierre-Louis Aublin, Kenta Ishiguro | 14m |
zpoline implements a syscall hook by binary rewriting: it replaces two-byte syscall/sysenter with callq %rax and places a trampoline at address 0 that… | ||
| Sponge: Fast Reactive Scaling for Stream Processing with Serverless Frameworks | Won Wook Song, Taegeon Um, Sameh Elnikety, Myeongjae Jeon, Byung-Gon Chun | 19m |
Sponge is a stream-processing runtime that pre-inserts router, transient and merge operators at compile time to avoid runtime state migration, redirects bursty… | ||
| On-demand Container Loading in AWS Lambda | Marc Brooker, Mike Danilov, Chris Greenwood, Phil Piwonka | 23m |
AWS Lambda increases deployable artifacts from zip files to 10 GiB images while keeping cold starts low by chunking flattened container tar layers… | ||
| Decentralized and Stateful Serverless Computing on the Internet Computer Blockchain | Maksym Arutyunyan, Andriy Berestovskyy, Adam Bratschi-Kaye, Ulan Degenbaev, Manu Drijvers, Islam El-Ashi, Stefan Kaestle, Roman Kashitsyn, Maciej Kot, Yvonne-Anne Pignolet, Rostislav Rumenov, Dimitris Sarlis, Alin Sinpalean, Alexandru Uta, Bogdan Warinschi, Alexandra Zapuc | 21m |
The Internet Computer runs stateful canisters compiled to WebAssembly under a Wasmtime sandbox, with canisters allowed up to 64 GB memory and deployed across… | ||
| Pinolo: Detecting Logical Bugs in Database Management Systems with Approximate... | Zongyin Hao, Quanfeng Huang, Chengpeng Wang, Jianfeng Wang, Yushan Zhang, Rongxin Wu, Charles Zhang | 19m |
PINOLO synthesizes queries that are guaranteed to be over-approximations or under-approximations of a seed SQL query by applying concrete mutators (predicate… | ||
| AutoARTS: Taxonomy, Insights and Tools for Root Cause Labelling of Incidents in... | Pradeep Dogga, Chetan Bansal, Richard Costleigh, Gopinath Jayagopal, Suman Nath, Xuchao Zhang | 20m |
Empirical study of over 2,000 incidents across 450+ Microsoft Azure services yields a 4-level hierarchical Arts taxonomy with 346 distinct contributing factors… | ||
| Avoiding the Ordering Trap in Systems Performance Measurement | Dmitry Duplyakin, Nikhil Ramesh, Carina Imburgia, Hamza Fathallah Al Sheikh, Semil Jain, Prikshit Tekta, Aleksander Maricq, Gary Wong, Robert Ricci | 17m |
Performance results often depend on trial order, not just configuration: a literature survey of 56 reproducible papers finds only 7% specify reset procedures… | ||
| AWARE: Automate Workload Autoscaling with Reinforcement Learning in Production... | Haoran Qiu, Weichao Mao, Chen Wang, Hubertus Franke, Alaa Youssef, Zbigniew T. Kalbarczyk, Tamer Başar, Ravishankar K. Iyer | 22m |
AWARE is an extensible framework that deploys RL autoscalers on Kubernetes with an OpenAI Gym-like interface, combines offline replay-buffer training with… | ||
| Nodens: Enabling Resource Efficient and Fast QoS Recovery of Dynamic Microservice... | Jiuchen Shi, Hang Zhang, Zhixin Tong, Quan Chen, Kaihua Fu, Minyi Guo | 19m |
Nodens is a runtime that builds a traffic-based load monitor, an blocking-aware load updater, and a resource-efficient query drainer to cut QoS recovery from… | ||
| Lifting the veil on Meta’s microservice architecture: Analyses of topology and... | Darby Huye, Yuri Shkuro, Raja R. Sambasivan | 20m |
Analysis of Meta's microservice estate uses Canopy traces and 22 months of deployment records to show topology is heterogeneous and in constant flux: the… | ||
| Tectonic-Shift: A Composite Storage Fabric for Large-Scale ML Training | Mark Zhao, Satadru Pan, Niket Agarwal, Zhaoduo Wen, David Xu, Anand Natarajan, Pavan Kumar, Shiva Shankar P, Ritesh Tijoriwala, Karan Asher, Hao Wu, Aarti Basant, Daniel Ford, Delia David, Nezih Yigitbasi, Pratap Singh, Carole-Jean Wu | 20m |
Composite fabric combines Tectonic HDD capacity with a flash tier called Shift that sits in a consistent‑hash ring and uses Meta’s cashlib for node‑local… | ||
| Calcspar: A Contract-Aware LSM Store for Cloud Storage with Low Latency Spikes | Yuanhui Zhou, Jian Zhou, Shuning Chen, Peng Xu, Peng Wu, Yanguang Wang, Xian Liu, Ling Zhan, Jiguang Wan | 20m |
Calcspar is an LSM-store variant built to tame Amazon EBS latency by regulating I/O rate and absorbing bursts with a fluctuation-aware cache and a… | ||
| Adaptive Online Cache Capacity Optimization via Lightweight Working Set Size... | Rong Gu, Simian Li, Haipeng Dai, Hancheng Wang, Yili Luo, Bin Fan, Ran Ben Basat, Ke Wang, Zhenyu Song, Shouwei Chen, Beinan W | 20m |
Cuki is a tiny, thread-safe cuckoo-style sketch that estimates working set size and item-repetition ratio online for variable-size objects, with provable error… | ||
| SAGE: Software-based Attestation for GPU Execution | Andrei Ivanov, Benjamin Rothenberger, Arnaud Dethise, Marco Canini, Torsten Hoefler, Adrian Perrig | 23m |
Software-only attestation that runs on NVIDIA A100 (Ampere) GPUs and pairs with a CPU SGX verifier to prove code and data integrity and secrecy. Implements a… | ||
| Confidential Computing within an AI Accelerator | Kapil Vaswani, Stavros Volos, Cédric Fournet, Antonio Nino Diaz, Ken Gordon, Balaji Vembu, Sam Webster, David Chisnall, Saurabh Kulkarni, Graham Cunningham, Richard Osborne, Daniel Wilkinson | 20m |
Built an AI accelerator extension called IPU Trusted Extensions (ITX) that adds an on‑chip hardware root-of-trust and PCIe Gen4‑rate cryptographic engines… | ||
| Arbitor: A Numerically Accurate Hardware Emulation Tool for DNN Accelerators | Chenhao Jiang, Anand Jayarajan, Hao Lu, Gennady Pekhimenko | 18m |
Arbitor is an open‑source compiler‑based emulator that rewrites ML compiler IR to insert custom low‑precision, approximate and sparsity policies, running on… | ||
| oBBR: Optimize Retrmissions of BBR Flows on the Internet | Pengqiang Bi, Mengbai Xiao, Dongxiao Yu, Guanghui Zhang | 25m |
USENIX ATC '23 - oBBR: Optimize Retransmissions of BBR Flows on the Internet
Pengqiang Bi, Shandong University, Mengbai Xiao, Shandong University, Dongxiao… | ||
| Bridging the Gap between QoE and QoS in Congestion Control: A Large-scale Mobile... | Jia Zhang, Yixuan Zhang, Enhuan Dong, Quan Cheng, Yan Zhang, Shaorui Ren, Zili Meng | 17m |
Builds Floo, a shim layer between apps and CCAs that monitors app state and network metrics, runs a reinforcement learning selection policy, and performs fast… | ||
| FarReach: Write-back Caching in Programmable Switches | Siyuan Sheng, Huancheng Puyang, Qun Huang, Lu Tang, Patrick P. C. Lee | 19m |
FarReach implements an in-switch write-back cache on a Tofino switch with P4 and a RoxDB backend, claiming up to 6.6× throughput versus a NetCache baseline… | ||
| CXL-ANNS: Software-Hardware Collaborative Memory Disaggregation and Computation... | Junhyeok Jang, Hanjin Choi, Hanyeoreum Bae, Seungjun Lee, Miryeong Kwon, Myoungsoo Jung | 22m |
CXL-ANNS places billion-point graphs and embeddings in a CXL memory pool, adds a local DRAM cache of inner nodes (59% hit rate with 128 GB local memory) and… | ||
| Overcoming the Memory Wall with CXL-Enabled SSD | Shao-Peng Yang, Minjae Kim, Sanghyun Nam, Juhyung Park, Jin-yong Choi, Eyee Hyun Nam, Eunji Lee, Sungjin Lee, Bryan S. Kim | 18m |
Study builds a physical memory tracer and a trace-driven simulator to model CXL-enabled SSDs and integrates a DRAM cache, MSHR and several prefetchers to… | ||
| STYX: Exploiting SmartNIC Capability to Reduce Datacenter Memory Tax | Houxiang Ji, Mark Mansi, Yan Sun, Yifan Yuan, Jinghan Huang, Reese Kuper, Michael M. Swift, Nam Sung Kim | 18m |
STYX RDMA-copies target memory regions to a SmartNIC (tested on a BlueField‑2), runs the heavy data-path work there, and RDMA-writes results back so the host… | ||
| Change Management in Physical Network Lifecycle Automation | Mohammad Al-Fares, Virginia Jean Beauregard, Kevin Grant, Angus Griffith, Jahangir Hasan, Chen Huang, Quan Leng, Jiayao Li, Alexander Lin, Zhuotao Liu, Ahmed Mansy, Bill Martinusen, Nikil Mehta, Jeffrey C. Mogul, Andrew Narver, Anshul Nigham, Melanie Obenberger, Sean Smith, Kurt Steinkraus, Sheng Sun, Edward Thiele, Amin Vahdat | 21m |
Google describes a deployed system that automates physical network lifecycle changes using the MALT language and a suite of services: a model design service, a… | ||
| AAsclepius: Monitoring, Diagnosing, and Detouring at the Internet Peering Edge | Kaicheng Yang, Yuanpeng Li, Sheng Long, Tong Yang, Ruijie Miao | 21m |
AAsclepius automates detection, localization, direction identification, and detouring of middle faults at the peering edge for Huawei Cloud. It probes… | ||
| Deploying User-space TCP at Cloud Scale with LUNA | Lingjun Zhu, Yifan Shen, Erci Xu, Bo Shi, Ting Fu, Shu Ma, Shuguang Chen, Zhongyu Wang, Haonan Wu, Xingyu Liao, Zhendan Yang, Zhongqing Chen, Wei Lin, Yijun Hou, Rong Liu, Chao Shi, Jiaji Zhu, Jiesheng Wu | 19m |
LUNA is a user-space TCP libOS built with DPDK and SR-IOV/flow-steering that runs on a run-to-completion model (batch and inline r2c), a user-space… | ||
| RubbleDB: CPU-Efficient Replication with NVMe-oF | Haoyu Li, Sheng Jiang, Chen Chen, Ashwini Raina, Xingyu Zhu, Changxu Luo, Asaf Cidon | 23m |
RubbleDB uses NVMe-oF target offload to ship compaction outputs once from a primary to secondaries, avoiding redundant compactions and offloading disk/network… | ||
| Distributed Transactions at Scale in Amazon DynamoDB | Joseph Idziorek, Alex Keyes, Colin Lazier, Somu Perianayagam, Prithvi Ramanathan, James Christopher Sorenson III, Doug Terry, Akshat Vig | 24m |
Explains how DynamoDB adds full ACID transactions by implementing timestamp ordering with a fleet of transaction coordinators, a durable ledger checkpointing… | ||
| Prefix Siphoning: Exploiting LSM-Tree Range Filters For Information Disclosure | Adi Kaufman, Moshik Hershcovitch, Adam Morrison | 19m |
Shows a concrete timing attack called prefix siphoning that abuses LSM-tree range filters (demonstrated on SuRF and the prefix Bloom filter) to recover key… | ||
| EPF: Evil Packet Filter | Di Jin, Vaggelis Atlidakis, Vasileios P. Kemerlis | 20m |
EPF demonstrates two concrete attacks, BPF-Reuse and BPF-ROP, that hijack kernel control flow and redirect execution into the BPF interpreter so unaligned or… | ||
| Translation Pass-Through for Near-Native Paging Performance in VMs | Shai Aviram Bergman, Mark Silberstein, Takahiro Shinagawa, Peter Pietzuch, Lluís Vilanova | 20m |
Proposes Translation Pass-Through (TPT): guests build one-dimensional guest-virtual→host-physical page tables while the host enforces isolation via a per-VM… | ||
| Efficient Memory Overcommitment for I/O Passthrough Enabled VMs via Fine-grained... | Yaohui Wang, Ben Luo, Yibin Shen | 25m |
V-Probe implements fine-grained page metadata inspection by injecting tiny helper binaries (e.g., Page_free and Page_order) so the hypervisor reads guest… | ||
| LPNS: Scalable and Latency-Predictable Local Storage Virtualization for... | Bo Peng, Cheng Guo, Jianguo Yao, Haibing Guan | 23m |
LPNS is an OS-level NVMe virtualization that provides latency‑predictable QoS for hybrid local cloud storage by implementing self‑feedback control, flexible… | ||
| P2CACHE: Exploring Tiered Memory for In-Kernel File Systems Caching | Zhen Lin, Lingfeng Xiang, Jia Rao, Hui Lu | 19m |
An in-kernel cache (P2CACHE) implements a read/write-distinguishable PM+DRAM hierarchy and a lightweight operation log to persist writes immediately to Intel… | ||
| Revisiting Secondary Indexing in LSM-based Storage Systems with Persistent Memory | Jing Wang, Youyou Lu, Qing Wang, Yuhao Zhang, Jiwu Shu | 20m |
Claims a complete design called PERSEID that uses a PS3 two-layer PM index, a hybrid PM-DRAM hash validation to avoid expensive primary-table reads, and two… | ||
| Zhuque: Failure is Not an Option, it’s an Exception | George Hodgkins, Yi Xu, Steven Swanson, Joseph Izraelevitz | 14m |
Claims a new programming model Whole Process Persistence and a userspace runtime Zhuque runtime that interposes on libc to map all process memory to PMEM and… | ||
| EnvPipe: Performance-preserving DNN Training Framework for Saving Energy | Sangjin Choi, Inhoe Koo, Jeongseob Ahn, Myeongjae Jeon, Youngjin Kwon | 21m |
EnvPipe is a PyTorch library that profiles each pipeline stage, reschedules pipeline units to expose usable pipeline bubbles, and selectively lowers GPU SM… | ||
| Decentralized Application-Level Adaptive Scheduling for Multi-Instance DNNs on... | Hsin-Hsuan Sung, Jou-An Chen, Wei Niu, Jiexiong Guan, Bin Ren, Xipeng Shen | 23m |
A user‑level scheduler implements decentralized DQN agents (lightweight 2–5 fully connected layers) inside each app to pick execution configurations (CPU… | ||
| UnFaaSener: Latency and Cost Aware Offloading of Functions from Serverless Platforms | Ghazal Sadeghian, Mohamed Elsakhawy, Mohanna Shahrad, Joe Hattori, Mohammad Shahrad | 22m |
Framework harvests spare VMs and on‑prem hosts and routes individual functions off the serverless path using pub/sub glue and a tiny injected header that reads… | ||
| LLFree: Scalable and Optionally-Persistent Page-Frame Allocation | Lars Wrenger, Florian Rommel, Alexander Halbuer, Christian Dietrich, Daniel Lohmann | 19m |
A new page-frame allocator LLFree implements a lock- and log-free design using a global bitfield (1 bit per 4 KiB), 2 MiB counters for 512-page groups, and… | ||
| SingularFS: A Billion-Scale Distributed File System Using a Single Metadata Server | Hao Guo, Youyou Lu, Wenhao Lv, Xiaojian Liao, Shaoxun Zeng, Jiwu Shu | 17m |
SingularFS builds a billion-scale distributed file system that uses a single metadata server and persistent memory with RDMA, replacing journaling with a KV… | ||
| The Hitchhiker's Guide to Operating Systems | Yanyan Jiang | 18m |
This work claims that treating programs and OS components as state transition systems improves OS teaching. It implements a minimal OS model with nine system… | ||
| Accelerating Distributed MoE Training and Inference | Lina | 20m |
Lina is a systems layer for distributed MoE that opportunistically prioritizes expensive all-to-all token exchanges over concurrent allreduce by splitting… | ||
| SmartMoE: Efficiently Training Sparsely-Activated Models through Combiningand... | Mingshu Zhai, Jiaao He, Zixuan Ma, Zan Zong, Runqing Zhang, Jidong Zhai | 12m |
SmartMoE is an automatic system that combines two-stage parallelization (offline pool construction + online selection) with a data-sensitive performance model… | ||
| MSRL: Distributed Reinforcement Learning with Dataflow Fragments | Huanzhou Zhu, Bo Zhao, Gang Chen, Weifeng Chen, Yijie Chen, Liang Shi, Yaodong Yang, Peter Pietzuch, Lei Chen | 26m |
Claims a new fragmented dataflow graph abstraction that splits an RL training loop into independent fragments mapped to CPU, CUDA kernels or DNN operator… | ||
| Beware of Fragmentation: Scheduling GPU-Sharing Workloads with Fragmentation... | Qizhen Weng, Lingyun Yang, Yinghao Yu, Wei Wang, Xiaochuan Tang, Guodong Yang, Liping Zhang | 22m |
Paper defines a statistical fragmentation metric that predicts expected unallocatable GPU capacity given task distributions, implements a scheduler called… | ||
| Towards Iterative Relational Algebra on the GPU | Ahmedur Rahman Shovon, Thomas Gilray, Kristopher Micinski, Sidharth Kumar | 17m |
Implements a CUDA GPU backend that runs iterative relational algebra for Datalog via a custom open-addressing hash table and fused join+projection operators… | ||
| VectorVisor: A Binary Translation Scheme for Throughput-Oriented GPU Acceleration | Samuel Ginzburg, Mohammad Shahrad, Michael J. Freedman | 25m |
Describes VectorVisor, a vectorized binary translator that compiles WebAssembly IR to OpenCL C, runs many copies of unmodified single‑threaded server programs… | ||
| nothing matches | ||