XML is playing crucial roles in web services, databases, and document representing and processing. However, the processing of XML document has been regarded as the main performance bottleneck especially for the processing of very large XML data. On the other hand, multi-core processing gains increasingly popularity both on the desktop computers and server computing machines. To take full advantage of multi-cores, we present a novel hybrid parallel XML processing model, which combines data-parallel and pipeline processing. It first partitions the XML by chunks to perform data parallel processing for both XML parsing and schema validation, then organize and execute them as a two stage pipeline to exploit more parallelism. The hybrid parallel XML processing model has shown great overall performance advantage on multi-core platform as indicated by the experiment performance results.
Understanding Scalability and Fine-Grain Parallelism of Synchronous Data Parallel Training
