Published On: Mon, May 7th, 2018

Microsoft launches Project Brainwave, the low training acceleration platform

Microsoft currently announced during a Build discussion a preview launch of Project Brainwave, a height for using low training models in a Azure cloud and on a corner in genuine time.

While some of Microsoft’s competitors, including Google, are betting on tradition chips, Microsoft continues to gamble on FPGAs to accelerate a models, and Brainwave is no exception. Microsoft argues that FPGAs give it some-more coherence than conceptualizing tradition chips and that a opening it achieves on customary Intel Stratix FPGAs is during slightest allied to that of tradition chips.

Last August, a association initial minute some aspects of Brainwave, that consists of 3 graphic layers: a high-performance distributed architecture; a hardware low neural networking engine that has been synthesized onto a FPGAs; and a compiler and runtime for deploying a pre-trained models.

Microsoft is attaching a FPGAs right to a altogether information core network, that allows them to turn something same to hardware microservices. The advantage here is high throughput and a vast latency rebate since this design allows Microsoft to bypass a CPU of a normal server to speak directly to a FPGAs. Indeed, Microsoft argues that Brainwave offers 5x reduction latency than Google’s TPUs.

When Microsoft initial announced BrainWave, a program smoke-stack upheld both a Microsoft Cognitive Toolkit and Google’s TensorFlow frameworks.

Brainwave is now in preview on Azure and Microsoft also promises to move support for it to Azure Stack and a Azure Data Box appliance.

About the Author

Leave a comment

XHTML: You can use these html tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>