GPUs in the News - data mining
To continue our series about HPC (high-performance computing) in general and GPUs (graphics-processing units) in particular, I will write today about a real-life application of GPU technology in business.
NVIDIA blog has an interesting mini-case study about how one German company is using GPUs for business analysis applications. Jedox AG produces a software called Palo Suite that uses a technique called OLAP (short for online business analytical processing) to make large databases manageable for data mining, which is the science of finding meaningful analyses from large sets of information.
The information age has given scientists, researchers and analysts what they have always sought - abundant data. One of the challenges of the 21st century, it seems to me, is not collecting more data (although that will continually be done) but learning how to make sense of it. Indeed, leaders in industry, science and government have already been working extensively on this problem (if you are interested in reading in more detail about data analysis in the informational age, I recommend the Special Report in The Economist called Data, data everywhere).
Jedox's software, Palo Suite, seeks to solve the data mining problem (that is, the problem that even supercomputers lack the processing power to analyze enormous databases) by condensing data gathered from a business into a meaningful subset.
As the NVIDIA post describes, OLAP compresses a large data set of three variables into a ""cube"", with each variable representing one of the three dimensions. With this compression, Palo Suite is able to reduce the storage space required for the data to only 1/10th of the original number.
Then, Palo Suite runs analyses of the data on a GPU system to find meaningful patterns and insights that would help the business better understand their sales patterns.
By organizing data in meaningful ways and analyzing it with the help of affordable computing technology such as GPU servers, businesses of the 21st century will be even informed about their role in the marketplace.
GPUs in the News - data mining
To continue our series about HPC (high-performance computing) in general and GPUs (graphics-processing units) in particular, I will write today about a real-life application of GPU technology in business.
NVIDIA blog has an interesting mini-case study about how one German company is using GPUs for business analysis applications. Jedox AG produces a software called Palo Suite that uses a technique called OLAP (short for online business analytical processing) to make large databases manageable for data mining, which is the science of finding meaningful analyses from large sets of information.
The information age has given scientists, researchers and analysts what they have always sought - abundant data. One of the challenges of the 21st century, it seems to me, is not collecting more data (although that will continually be done) but learning how to make sense of it. Indeed, leaders in industry, science and government have already been working extensively on this problem (if you are interested in reading in more detail about data analysis in the informational age, I recommend the Special Report in The Economist called Data, data everywhere).
Jedox's software, Palo Suite, seeks to solve the data mining problem (that is, the problem that even supercomputers lack the processing power to analyze enormous databases) by condensing data gathered from a business into a meaningful subset.
As the NVIDIA post describes, OLAP compresses a large data set of three variables into a ""cube"", with each variable representing one of the three dimensions. With this compression, Palo Suite is able to reduce the storage space required for the data to only 1/10th of the original number.
Then, Palo Suite runs analyses of the data on a GPU system to find meaningful patterns and insights that would help the business better understand their sales patterns.
By organizing data in meaningful ways and analyzing it with the help of affordable computing technology such as GPU servers, businesses of the 21st century will be even informed about their role in the marketplace.