Seminaris de Recerca Severo Ochoa
Dimarts, 11 Juliol, 2017
The most common software update on the AppStore *by far* is "Bug fixes and performance enhancements." Now that Moore's Law Free Lunch has ended, programmers have to work hard to get high performance for their applications. But why is performance so hard to deliver?
Dimecres, 14 Juny, 2017
Dilluns, 12 Juny, 2017
This talk will discuss recent technologies that Bruce Jacob and his group have helped to develop in high-performance systems, including Micron’s Hybrid Memory Cube, flash-based main memories, and high-bandwidth/high-capacity memory, and will discuss the impact of tomorrow’s memory technologies on tomorrow’s applications and operating systems.
Dimecres, 07 Juny, 2017
This talk will cover the technical features of the RISC-V ISA design, which has the goals of scaling from the tiniest implementations for IoT up to the largest warehouse-scale computers, with support for extensive customization. We'll also describe industry-competitive open-source cores developed at UC Berkeley, all written in Chisel, a productive new open-source hardware design language.
Dimarts, 06 Juny, 2017
Dijous, 25 Maig, 2017
In this lecture, organised in collaboration with UPC, Dr Vargas will explore the key challenges and opportunities for data management in the new scientific world, and discuss how a possible data centric artificial intelligence community can best contribute to these exciting domains.
Dijous, 18 Maig, 2017
In the Big Data era, due to the ever-growing graph scale and algorithm complexity, several distributed graph processing frameworks have attracted many interests from both academia and industry. In this talk, will be investigated how to achieve the trade-off between performance and cost for large scale graph processing on the Cloud.
Dimecres, 10 Maig, 2017
Fast content-based searches and complex analytics of the vast amount of data collected via social media, cell phones, ubiquitous smart sensors, and satellites is likely to be the biggest economic driver for the IT industry over the next decade. A cheaper and cooler alternative to large clusters which provides high-performance, high-capacity, scalable random-access flash storage, and allows computation near the data via FPGA-based programmable flash controllers will be presented.
Dimarts, 09 Maig, 2017
In this talk, a structured approach to the management of HPC resilience using the concept of resilience-based design patterns will be presented. A design pattern is a general repeatable solution to a commonly occurring problem. The authors identify the commonly occurring problems and solutions used to deal with faults, errors and failures in HPC systems.