J
Javier Picorel
Researcher at École Polytechnique Fédérale de Lausanne
Publications - 21
Citations - 667
Javier Picorel is an academic researcher from École Polytechnique Fédérale de Lausanne. The author has contributed to research in topics: Dram & Server. The author has an hindex of 7, co-authored 17 publications receiving 549 citations. Previous affiliations of Javier Picorel include University of Edinburgh & Huawei.
Papers
More filters
Proceedings ArticleDOI
Meet the walkers: accelerating index traversals for in-memory databases
Onur Kocberber,Boris Grot,Javier Picorel,Babak Falsafi,Kevin T. Lim,Parthasarathy Ranganathan +5 more
TL;DR: Widx is introduced, an on-chip accelerator for database hash index lookups, which achieves both high performance and flexibility by decoupling key hashing from the list traversal, and processing multiple keys in parallel on a set of programmable walker units.
Journal ArticleDOI
Scale-out processors
Pejman Lotfi-Kamran,Boris Grot,Michael Ferdman,Stavros Volos,Onur Kocberber,Javier Picorel,Almutaz Adileh,Djordje Jevdjic,Sachin Satish Idgunji,Emre Ozer,Babak Falsafi +10 more
TL;DR: This work introduces a methodology for designing scalable and efficient scale-out server processors based on a metric of performance-density, and facilitates the design of optimal multi-core configurations, called pods.
Proceedings ArticleDOI
The Mondrian Data Engine
Mario Drumond,Alexandros Daglis,Nooshin Mirzadeh,Dmitrii Ustiugov,Javier Picorel,Babak Falsafi,Boris Grot,Dionisios Pnevmatikatos +7 more
TL;DR: This thesis is that efficient NMP calls for an algorithm-hardware co-design that favors algorithms with sequential accesses to enable simple hardware that accesses memory in streams, and introduces an instance of such a co-designed NMP architecture for data analytics, the Mondrian Data Engine.
Proceedings ArticleDOI
BuMP: Bulk Memory Access Prediction and Streaming
TL;DR: Bulk Memory Access Prediction and Streaming employs a low-cost predictor to identify high-density pages and triggers bulk transfer operations upon the first read or write to the page, thereby reducing DRAM energy per access by 23%, and improves server throughput by 11% across a wide range of server applications.
Proceedings ArticleDOI
Near-Memory Address Translation
TL;DR: This paper proposes the Distributed Inverted Page Table (DIPTA), a near-memory structure in which the smallest memory partition keeps the translation information for its data share, ensuring that the translation completes together with the data fetch.