BEGIN:VCALENDAR PRODID;X-RICAL-TZSOURCE=TZINFO:-//Calagator//EN CALSCALE:GREGORIAN X-WR-CALNAME:Calagator METHOD:PUBLISH VERSION:2.0 BEGIN:VTIMEZONE TZID;X-RICAL-TZSOURCE=TZINFO:America/Los_Angeles BEGIN:DAYLIGHT DTSTART:20080309T020000 RDATE:20080309T020000 TZOFFSETFROM:-0800 TZOFFSETTO:-0700 TZNAME:PDT END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT CREATED;VALUE=DATE-TIME:20080829T185647Z DTEND;TZID=America/Los_Angeles;VALUE=DATE-TIME:20080902T113000 DTSTART;TZID=America/Los_Angeles;VALUE=DATE-TIME:20080902T103000 DTSTAMP;VALUE=DATE-TIME:20080829T185647Z LAST-MODIFIED;VALUE=DATE-TIME:20080829T185647Z UID:http://calagator.org/events/1250455662 DESCRIPTION: \;\nTitle: GpuGen: Bringing the Power of GPUs into t he Haskell World \;\n \;\nSpeaker: Sean Lee \;\n P rogramming Languages &\; Systems \;\n UNSW\, Sydney \ ;\n \;\nDate: Tuesday\, September 2nd. \;\n 10.30a m \;\n \;\nLocation: Galois\, Inc. \;\n 421 SW 6th Ave. Suite 300 \;\n (3rd floor of the Commonwealth Building ) \;\n Portland\, Oregon \;\n \;\nAbstract: \;\n&# 13\;\nFor the last decade\, the performance of GPUs has out-grown CPUs\, and \;\ntheir programmability has also improved to the level where t hey can be \;\nused fo general-purpose computations. Nonetheless\, GP U programming is \;\nstill limited only to those who understand the h ardware architecture and \;\nthe parallel processing. This is because the current GPU programming \;\nsystems are based on the specialized parallel processing model\, and \;\nrequire low-level attention in m any aspects such as thread launching \;\nand synchronization. \;\n \;\nThe need for a programming system which provides a high-level 3\;\nabstraction layer on top of the GPU programming systems without los ing \;\nthe performance gain arises to facilitate the use of GPUs. In stead of \;\nwriting a programming system from the scratch\, the deve lopment of a \;\nHaskell extension has been chosen as the ideal appro ach\, since the \;\nHaskell community has already accumulated a signi ficant amount of \;\nresearch and resources for Nested Data Paralleli sm\, which could be \;\nadopted to provide a high-level abstraction o n GPU programming and even \;\nto broaden the applicability of GPU pr ogramming. In addition\, the \;\nForeign Function Interface of Haskel l is sufficient to be the \;\ncommunication medium to the GPU. \;\ n \;\nGpuGen is what connects these two dots: GPUs and Haskell. It co mpiles \;\nthe collective data operations such as scan\, fold\, map\, etc\, which \;\nincur most computation cost\, to the GPU. The design of the system\, the \;\nstructure of the GpuGen compiler\, and the c urrent development status are \;\nto be discussed in the talk. \;\ n \;\nBiographical details: \;\n \;\n Sean Lee is a PhD candi date at the UNSW\, Sydney\, working \;\n in the Programming Languag es &\; Systems Group. This summer \;\n he's been interning at Nv idia in Santa Clara\, working on \;\n programming GPUs with Haskell . \;\n \;\nAbout the Galois Tech Talks. \;\n \;\n Galois ( http://galois.com) has been holding weekly technical \;\n seminars for several years on topics from functional programming\, \;\n form al methods\, compiler and language design\, to cryptography\, and \;\ n operating system construction\, with talks by many figures from the& #13\;\n programming language and formal methods communities. \;\n&# 13\;\n The talks are open and free. If you're planning to attend\, dro pping \;\n a note to is appreciated\, but not required. \;\n If you're interested in giving a talk\, we're always looking for new \;\n speakers.\n\nTags: galois\, technology\, functional programming\, pdxfunc\n\nImported from: http://calagator.org/events/1250455662 URL:http://groups.google.com/group/pdxfunc/browse_thread/thread/33da952d0 02bb1c7 SUMMARY:Galois Tech Talk: GpuGen: Bringing the Power of GPUs into the Has kell World LOCATION:Galois\, Inc: 421 SW 6th Ave. Suite 300\, Portland OR 97204 US SEQUENCE:0 END:VEVENT END:VCALENDAR