MUNICH: GRAPHICS OUTFIT Nvidia has announced Rapids, an open source GPU-acceleration platform that it claims will make data crunching 50 times faster compared with CPU-only systems.
"Data has become a very important part of every company, and it's become a critical aspect of how companies do business," said Jeff Tseng, head of product at Nvidia's newly-formed AI Infrastructure team. "It's a logical next step for us."
Developed over the past two years, Rapids - based on Python and built on top of Apache Arrow - takes aim at data scientists and large businesses interested in harnessing the power of GPUs to analyse massive amounts of data and make accurate predictions at "unprecedented speed".
For example, Rapids can help businesses to more-easily detect credit card fraud, forecast retail inventory and to better understand consumer behaviour, Nvidia said.
"For 35 years, we have relied on Moore's Law. And now, it has essentially stopped," said Nvidia founder and CEO Jensen Huang during Wednesday's GTC keynote. "Moore's Law has ended at a time when demand for computing enhancement continues to grow.
"GPU-accelerated computing is now recognised as the path forward", he addd.
Early benchmarks shown off by Nvidia, which ran Rapids using the XGBoost machine learning algorithm on an Nvidia DTX-2 system, shows a 50 times speed improvement compared to a CPU-only system. This, the company claims, will allow data scientists to reduce typical training times from days to hours, or from hours to minutes, depending on the size of their dataset.
Rapids has some big-name support at launch, ranging from the likes of as Anaconda, BlazingDB and Databricks to open source pioneers IBM, HPE and Oracle.
"The compute world today requires powerful processing to handle complex workloads like data science and analytics - it's a job for Nvidia GPUs," swooned Clay Magouyrk, senior vice president of Software Development at Oracle Cloud Infrastructure.
"Rapids is accelerating the speed at which this processing and machine learning training can be done. We are excited to support this new suite of open-source software natively on Oracle Cloud Infrastructure and look forward to working with Nvidia to support Rapids across our platform, including the Oracle Data Science Cloud, to further accelerate our customers' end-to-end data science workflows.
Nvidia has made access to Rapid's open source suite of libraries available at http://www.rapids.ai, where the code is being released under the Apache license. µ
We should be shocked, but...
But the search giant has now squashed the bug
But it's not yet available here in Blighty