dividiti.com - news about our technology and initiatives to bring industry, academia and end-users together for collaborative and reproducible co-design of the whole software and hardware stack of efficient, reliable and cheap computer systems for emerging workloads such as deep learning.
The complexity of engineering computer systems has been growing exponentially (along with the other things exponential growth is actually good for e.g. transistor count). "Quality, cost to deliver, time to deliver – pick any two", an old adage goes. But the rising complexity means that even the ability to pick any two is becoming elusive. Both hardware and software take longer and cost more to produce, only to achieve quality that users will view as merely acceptable.
We thus believe the time is ripe for accelerating computer systems' R&D – from hardware design to programming tools to software development. We aim for liberating engineers from complex, time-consuming and error-prone tasks, therefore allowing them to focus on what they enjoy most – creative problem solving and innovation.