Loading…
This event has ended. View the official site or create your own event → Check it out
This event has ended. Create your own
The schedule is still fluid at the moment as the final approved sessions are added. We appreciate your patience.
View analytic
Thursday, May 22 • 12:00pm - 12:50pm
In-Place Computing for Big and Complex Data - Yicheng Huang, MacroData

Sign up or log in to save this to your schedule and see who's attending!

In order to address real-time analytic problems for big data, we’ve developed an unconventional technology, In-Place Computing (IPC). It defines an abstraction that data objects live (save) and work (compute) in one flat and infinite address space. To approximate an IPC system in 64-bit Linux, we introduce a key notion, Macro Data Structure (MDS), which looks just like any regular data structure except it is in-place, persistent, relocatable and big enough to hold entire set of big data. The challenge is to design a mechanism for efficient MDS implementation to reduce page faults and improve performance as the data size exceeds that of the swap space. The mechanism covers issues for (1) memory arrangements for better data locality and (2) access patterns in a hierarchical MDS. The mechanism delivers substantial performance advance for an IPC implementation.

Speakers
JC

Jocelyn Chen

MacroData
Started in late 2011 by two computer science researchers, Dr. Hseush from Columbia and Dr. Huang from Cornell, the BigObject team was first incubated inside eBizprise Inc., striving to solve complicated supply and demand forecast issues for emerging countries. In the past two years, these two researchers have been revisiting the fundamentals of computer science and endeavored to discover a more efficient and effective way to handle big and... Read More →


Thursday May 22, 2014 12:00pm - 12:50pm
Neptune

Attendees (4)