Sciweavers

TREC
2007

Exegy at TREC 2007 Million Query Track

14 years 1 months ago
Exegy at TREC 2007 Million Query Track
Exegy’s submission for the TREC 2007 million query track consisted of results obtained by running the queries against the raw data, i.e., the data was not indexed. The hardwareaccelerated streaming engine used to perform the search is the Exegy Text Miner (XTM), developed at Exegy, Inc. The search engine’s architecture is novel: XTM is a hybrid system (heterogeneous compute platform) employing general purpose processors (GPPs) and field programmable gate arrays (FPGAs) in a hardware-software co-design architecture to perform the search. The GPPs are responsible for inputting the data to the FPGAs and reading and postprocessing the search results that the FPGAs output. The FPGAs perform the actual search and due to the high degree of parallelism available (including pipelining) are able to do so much more efficiently than the GPPs. For the million query track the results for a particular query were obtained by searching for the exact query string within the corpus. This brute forc...
Naveen Singla, Ronald S. Indeck
Added 07 Nov 2010
Updated 07 Nov 2010
Type Conference
Year 2007
Where TREC
Authors Naveen Singla, Ronald S. Indeck
Comments (0)