Blog post

Dev update: MGM API and Architecture

The system consists of the following elements:– Central server– Docker registry compatible image registry managed by Megamind– Computing nodes– Developers– End users Central server. Megamind Central server consists of the following parts: Server REST API. Provides a top-level API for task management and connecting computing nodes. The task in this context is a docker image …

Dev update: MGM API and Architecture Read More »

Supercomputers for NNs to be replaced by a distributed network

Russian programmers have developed a platform for distributed training of large neural networks. It is adapted for a network of many computers of different power, any of which can exit the process at any time. As with distributed scientific computing projects such as Folding@home, this approach, with the help of many volunteers, can create a …

Supercomputers for NNs to be replaced by a distributed network Read More »

Development log

Everybody can agree that the lockdown slowed down all kinds of businesses. Well, it hardly affected Megamind: we continue service development as planned.  According to the Roadmap we plan to release Beta in Q4 2020 and become operational in 2021. At the moment we are working on the platform core: in particular load balancer, API/SDK, …

Development log Read More »

Architecture and technological solutions of MEGAMINÐ

We designed Megamind with a big picture in mind: a marketplace of distributed computing resources orchestrated by the central authority. The authority divides computing tasks into batches and sends them to the most appropriate pieces of hardware which are connected to the network. At the moment we are developing API/SDK to enable a variety of …

Architecture and technological solutions of MEGAMINÐ Read More »