A common phenomenon is that a company with a new chip technology will focus on the technology features, trying to convince you to get out of your comfort zone and go through the disruption of changing how you do things in order to gain the technology benefits of their product.
We believe in going in the opposite direction. We started with a deep understanding of Enterprise software, and created technology that respects the monumental complexity of that ecosystem and respects the customer’s healthy aversion of doing anything to disrupt the complex, interlocking source code they create. We recognize that the software development costs dwarf hardware costs, and so we created a technology that provides a powerful advantage -- 12 times more compute -- but on your existing software.
Other companies make claims, but when you scratch the surface, it’s the same story -- you have to pick from among a few models in TensorFlow that they accelerate, or you have to learn tools with a steep learning curve, and even then, they only apply to some parts of some applications.
With Intensivate acceleration, you get a true server, running every line of source in your own code, and every open source package out there, at the same or better performance than the industry standard. But you get 12 times more servers for your money. They’re considered accelerators because they need cluster ready applications or applications with lots of threads to give full advantage. So you won’t run legacy ERP applications on it, you still need your trusty standard servers for that. But on apps where you really need more compute, that’s where we drop in, to give you twelve times more for your money.