BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CECS - ECPv6.18.4//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-ORIGINAL-URL:https://www.cecs.uci.edu
X-WR-CALDESC:Events for CECS
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:20090308T100000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:20091101T090000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:20100314T100000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:20101107T090000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:20110313T100000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:20111106T090000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/Los_Angeles:20100903T150000
DTEND;TZID=America/Los_Angeles:20100903T160000
DTSTAMP:20231011T183701Z
CREATED:20231011T183701Z
LAST-MODIFIED:20231011T183701Z
UID:14315-1283526000-1283529600@www.cecs.uci.edu
SUMMARY:Framework for Spiking Neural Network Modeling on High-Performance Architectures
DESCRIPTION:Framework for Spiking Neural Network Modeling on High-Performance Architectures\nby Jayram Moorkanikara Nageswaran \nCommittee: \nProf. Nikil Dutt (chair)\nProf. Jeffrey L Krichmar\nProf. Alex Nicolau\nProf. Alex Veidenbaum \nSpiking neural network (SNN) models are emerging as a plausible paradigm for characterizing neural dynamics in the cerebral cortex. Traditionally these SNN models were simulated on large-scale clusters\, super-computers\, or on dedicated VLSI architectures. Alternatively\, Graphics Processing Units (GPUs) can provide a low-cost\, programmable\, and high-performance computing platform for simulation of SNNs. This thesis proposes a systematic framework for modeling and simulation of biologically realistic large-scale spiking neural networks on high-performance graphics processors. \nThe first part of the framework consists of a high-level specification to quickly build arbitrary\, large-scale spiking neural network for different applications. The high-level SNN specification is then converted to a sparse adjacency matrix representation and mapped on the GPUs. Further\, we present a collection of new techniques related to parallelism extraction\, mapping of irregular communication\, and compact adjacency matrix representation for effective simulation of SNNs on GPUs. The last part of the generic framework proposes an evolutionary approach to automate parameter tuning in spiking neural networks. We demonstrate the validity of this framework by applying to three different spike-based computation problems. Significant improvements in modeling\, simulation and parameter tuning time have been achieved by these approaches.
URL:https://www.cecs.uci.edu/event/framework-for-spiking-neural-network-modeling-on-high-performance-architectures/
CATEGORIES:PhD Defenses
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/Los_Angeles:20100907T150000
DTEND;TZID=America/Los_Angeles:20100907T160000
DTSTAMP:20231011T183701Z
CREATED:20231011T183701Z
LAST-MODIFIED:20231011T183701Z
UID:14314-1283871600-1283875200@www.cecs.uci.edu
SUMMARY:Leveraging Profile-Selected Execution Patterns for Optimized Code Execution in Resource-Constrained Systems
DESCRIPTION:Leveraging Profile-Selected Execution Patterns for Optimized Code Execution in Resource-Constrained Systems\nby Carmen Badea \nLocation: Bren Hall 3013 \nDate and Time: September 7\, 2010 11:00am \nCommittee:\nProf. Alex Nicolau (Chair)\nProf. Alex Veidenbaum\nProf. Nikil Dutt \nAbstract:\nEmbedded systems and related devices have become ubiquitous in today’s world and are expected to proliferate even further. The execution performance for such devices is of paramount importance and is often perceived as the limiting factor w.r.t. both the capabilities of the system\, as well as the user perception. Furthermore\, unlike traditional general-purpose computing systems\, these devices exhibit further inherent constraints due to their nature: memory and energy consumption. \nIn order to overcome the intrinsic drawbacks of Java execution for resource-constrained systems\, we propose to use dynamic profile-selected\, optimized execution patterns in order to build a lightweight Java execution environment\, targeting resource-constrained systems\, that achieves competitive Java runtime performance\, while having negligible impact on memory footprint\, as well as reducing the overall energy consumption. \nMore recently\, with the advancements in the web application technology\, the browser has emerged as the favorite platform for deploying many applications\, in particular for embedded devices such as smartphones. As such\, we investigated whether the use of frequent execution patterns selected as directed by dynamic profiling would be beneficial for browser execution performance. With the use of collected dynamic profiling data indicating the most frequent patterns of execution we were able to develop a browser prototype that delivers significant runtime performance speedups.
URL:https://www.cecs.uci.edu/event/leveraging-profile-selected-execution-patterns-for-optimized-code-execution-in-resource-constrained-systems/
CATEGORIES:PhD Defenses
END:VEVENT
END:VCALENDAR