Cracking AI’s storage bottleneck and supercharging inference at the edge


Want smarter ideas in your inbox? Sign up for our weekly newsletters to get what is important for businesses, information and security leaders. Subscribe now


EU applications are increasingly increasing operations of more enterprises, and even promoting patient care through developed medical footage to protect against wildlife and even promote nature promotion.

During ventureebeat Transform 2025Greg Matson, Products and Marketing, Solidig and Roger Cummings, CEO of the summit: AIO, M12 in M12 in M12 in M12 in M12, using the health care works on stage in the scene.

https://www.youtube.com/watch?v=oxjdhmscsi3a

The Monai’s frame is a progress in medical views, faster, more reliable and secure. Progress in storage technology is what allows researchers to build, repeat this frame and allow them to innovate quickly. Peak: AIO, electricity, performer and more than two million bodies of the node in a node, the power of the power, performer and a high-capacity warehouse.

“The enterprise’s infrastructure is developing rapidly, depending on where they are in the Pipeline, according to special use, it is necessary to adapt to the storage apparatus,” Matson said. “The type of work we talk to Monai as well as the type of use of a training cluster needs a good thing with a very high level of state education, which is very high.

To improve the AI ​​beyond

For peak performance beyond, it is very important to avoid reducing a node to bring close to information. And what the key removes the problem. It can be done by making a memory of the AI ​​infrastructure to scalize with information and metadata. The intimacy of accounts increases sharply calculation.

“All huge placements, large green area data centers for AI, using many specific hardware designs to be able to consider the information as close as possible, you see large green area information centers for the EU,” Matson said. “Their information centers are a very high level of state, at a very high speed, a very high speed, the edge of the same technology and the enterprise in a microcosmos.”

All in solid standing is critical to buyers of AI systems to ensure you get the most performance from your system. This allows you to bring a large number of information and allows incredible processing power in a small system on the edge.

The future of AI device

“To do this, it is important to provide open, expandable and memory speed solutions using the latest and largest technology.” “This is a company to ensure that this openness, this slope and the scale of this need for the need for this openness. I think you will see the economies.”

For general training and result information pipeline and in the result, equipment needs will continue to grow, it is a very high speed solution that is very high speed SSD or power effective.

“I would say that this can change a very high-speed product, which has been a petabyte over several years in a few years in a few years in a few years,” he said. Matson. “You see how the Big GPU seller will determine the next warehouse architecture in the system, which is very closely, in a large SSD in HBM, we will continue this in both directions in the next five or 10 years.”



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *