White Papers

Executive summary
4 Deep Learning using iAbra stack on Dell EMC PowerEdge Servers with Intel technology | Document ID
Executive summary
Why
Enterprise needs solutions to business problems from deep learning not more technical challenges. Many
challenges exist in providing an integrated platform for deep learning development especially in the IoT era.
From semiconductor parts to interconnects and servers through to software libraries and domain-centric data
science every layer in the system has important consequences to overall solution business effectiveness.
Therefore Dell EMC, iAbra and Intel have combined and optimized best in breed components to offer an end
to end off the self-solution, which can kick start enterprise AI programs with solutions not problems, getting
you from domain specific data to deploy-able solution in the datacenter or edge
What
The solution based around Dell EMC PowerEdge C6420 high density server offers pre-configured scalability
of human and server resources, portability between datacenter and edge, GUI workflow from sample creation
through to neural network design evolution then deployment. iAbra's Deep Neural Network algorithm with
auto ML built in from the core takes advantage of latest Intel CPUs with AI extensions, interconnects &
FPGA's combined with Dell EMC PowerEdge C6420 high density to overcome the performance limitations to
practical power efficient self-optimizing AI and model deployment portability
How
Dell EMC, iAbra and Intel solution provides a new paradigm for Deep Learning training and inference, built
from the ground up for enterprise needs. Providing your enterprise a pre integrated clustered stack for training
and inference kick starts programs by removing uncertainty and time spent choosing myriad of DIY
components up and down the solution stack, clear and certain workflow to finding the optimal network to your
business problem removes program risks and costs, horizontal cluster scalability provides project speed up
with increased cluster size, and clear defined path to deploy-ability with model portability fidelity removes the
risks to productization.