IBM And Xilinx Boost Power Chips To Further Expand Into Big Data And Machine Learning

By Vincent Lanaria, Tech Times | November 17, 12:02 AM

Like Follow Share Tweet Reddit Subscribe
IBM Xilinx Partnership

IBM and Xilinx announced a partnership, and the two companies are working on integrating Xilinx FPGAs with Power systems. The pair intends to accelerate workloads such as big data analytics and machine learning, to name a few.
(Photo : Xilinx)

IBM and chip maker Xilinx team up to bring IBM's Power processors to more servers, looking to garner a bigger portion of the data center market than Intel.

Last year, the Big Blue initiated its Open Power program in an effort to boost Power by licensing the design for use by other companies. Now, with this partnership, IBM will be able to take things up a notch.

The pair will work with everything, including infrastructure, middleware and software. They'll be focusing on integrating Xilinx field-programmable gate arrays (FPGA) with Power-based systems to allow faster processing rates to accelerate a wide variety of workloads, such as machine learning, big data analytics, genomics, network-functions virtualization and high-performance computing.

"The combination of IBM and Xilinx provides our clients not only with a new level of accelerated computing made possible by the tight integration between IBM POWER processors and Xilinx FPGAs, but also gives them the ability to benefit directly from the constant stream of innovation being delivered by the rapidly expanding OpenPOWER ecosystem," Ken King, general manager of Open Power, says in a press release.

On top of that, IBM intends to both develop and qualify Xilinx accelerator boards into IBM Power Systems servers. On the other hand, Xilinx plans to roll out Power variants of its leading software named SDAccel Development Environment, including libraries for the community of Open Power developers.

In a bid to provide a faster computing performance, the two companies will persist in working on IBM's Coherent Accelerator Processor Inferface (CAPI), which delivers Accelerate Compute, Accelerate Storage and Accelerate Networking.

Seeing that Xilinx chips will be compatible with CAPI, they will be able to access memory caches on Power processors, which will considerably increase performance.

IBM used to be the sole maker of Power servers, mainly using only IBM technologies. Now, Big Blue has opened up, allowing third parties to build Power products, which is probably because the Unix market is in a downturn.

IBM has collaborated with Nvidia before to integrate GPUs with the Watson supercomputer, which is a Power system. Also, both E4 Computer Engineering and Penguin Computing are making their own high-performance computers with Power processors.

Like Follow Share Tweet Reddit Subscribe
© 2016 Tech Times, All rights reserved. Do not reproduce without permission.

Real Time Analytics