Think combining HPC and AI workloads is a challenge? Wait until you try to converge flash and HDD
(2021/06/10)
- Reference: 1623344407
- News link: https://www.theregister.co.uk/2021/06/10/hpe_regcast_event/
- Source link:
Webcast We all know HPC and AI datasets are going to be massive. But they’re also very different — the former built from large, sequential files, while the latter from much more random info.
The underlying storage architectures are quite different too — HDD and Infiniband for the former, Gbe and NVMe flash for the latter.
But with workloads increasingly converging, organisations face a big problem. Tailor systems towards flash, and prices will go through the roof. Focus on HDD, and you could be building in a bottleneck that will throttle the arrays of CPUs and GPUs you have crunching through your most pressing problems.
It’s a fundamental problem, and perfectly capable of stopping you solving much harder and more important problems.
So, whether you’re pondering juicing up your HPC workloads with a little machine learning, or want to work in some modelling to GPU-powered AI work, you should join our webcast, [1]Spend Less on HPC/AI Storage , on June 17 at 0800 PDT (1100 EDT, 16:00 BST.)
Our broadcast expert Tim Phillips will be conversing with HPE’s Uli Plechschmidt, who will [2]explain why you should be spending less on HPC/AI storage — and more on CPU/GPU compute.
They’ll pick through what sticking with your existing architectures could cost — in terms of both cold hard cash, and in innovation. And they’ll explain exactly what parallel HPC/AI storage could mean for you and your workloads, and how to build infrastructure that meets your needs and doesn’t cost the Earth.
Joining this session is a model of simplicity. Just [3]drop your details in here , and we’ll update your calendar and nudge you on the day. In the meantime, just relax, knowing that help with your storage conundrums is on its way.
Sponsored by HPE
[1] https://whitepapers.theregister.com/paper/view/11704/spend-less-on-hpcai-storage?td=promo2
[2] https://whitepapers.theregister.com/paper/view/11704/spend-less-on-hpcai-storage?td=promo2
[3] https://whitepapers.theregister.com/paper/view/11704/spend-less-on-hpcai-storage?td=promo2
The underlying storage architectures are quite different too — HDD and Infiniband for the former, Gbe and NVMe flash for the latter.
But with workloads increasingly converging, organisations face a big problem. Tailor systems towards flash, and prices will go through the roof. Focus on HDD, and you could be building in a bottleneck that will throttle the arrays of CPUs and GPUs you have crunching through your most pressing problems.
It’s a fundamental problem, and perfectly capable of stopping you solving much harder and more important problems.
So, whether you’re pondering juicing up your HPC workloads with a little machine learning, or want to work in some modelling to GPU-powered AI work, you should join our webcast, [1]Spend Less on HPC/AI Storage , on June 17 at 0800 PDT (1100 EDT, 16:00 BST.)
Our broadcast expert Tim Phillips will be conversing with HPE’s Uli Plechschmidt, who will [2]explain why you should be spending less on HPC/AI storage — and more on CPU/GPU compute.
They’ll pick through what sticking with your existing architectures could cost — in terms of both cold hard cash, and in innovation. And they’ll explain exactly what parallel HPC/AI storage could mean for you and your workloads, and how to build infrastructure that meets your needs and doesn’t cost the Earth.
Joining this session is a model of simplicity. Just [3]drop your details in here , and we’ll update your calendar and nudge you on the day. In the meantime, just relax, knowing that help with your storage conundrums is on its way.
Sponsored by HPE
[1] https://whitepapers.theregister.com/paper/view/11704/spend-less-on-hpcai-storage?td=promo2
[2] https://whitepapers.theregister.com/paper/view/11704/spend-less-on-hpcai-storage?td=promo2
[3] https://whitepapers.theregister.com/paper/view/11704/spend-less-on-hpcai-storage?td=promo2