Hi Ying,
Thank you for your question!
One of targeted usages for Intel® DAAL is the performance acceleration of Big Data analytics. This is where performance and efficiency matter. Since DAAL optimizes the entire workflow, from data acquisition to data transformation to data analysis, training, and prediction, it is very important for Big Data analytics domain.
Indeed, in typical Big Data analysis environment you cannot simply focus on optimization of numerical computation side. Every aspect of data management (remember of really BIG data volumes) may be essential for the overall solution performance. That's why DAAL provides IA-optimized building blocks for data extraction (via Data Sources such as CSV data files or SQL queries), data streaming into memory, effective in-memory data manipulations (via Numeric Tables) and finally variety of optimized and threaded numerical algorithms.
Intel DAAL package provides some distributed computing code samples how to use the library with MPI*-based High-Performance Data Analytics environments, with Spark* and Hadoop*. With these samples you can start using Intel® DAAL in most commonly used Big Data environments in minutes and seeing how native code library accelerates every piece of your Big Data infrastructure.
Thank you,
Sergey Maidanov
