Bitfusion Documentation


Bitfusion provides elastic AI infrastructure for faster, more efficient development and deployment of deep learning workloads. The software deploys into any data center or cloud or can be hosted and managed by Bitfusion, and it bundles and supports all the major deep learning frameworks such as Tensorflow, MXNet, Caffe, Torch, etc. Our platform is powered by our core virtualization engine, a transparent middleware layer that combines multiple systems into a single elastic compute cluster that supports sharing, scaling and management of heterogeneous compute resources.

Get Started

Search results for "{{ search.query }}"

No results found for "{{search.query}}". 
View All Results

Bitfusion Core

Bitfusion Core is a transparent virtualization layer that combines multiple systems into a single elastic compute cluster. By using Core, you can share, scale and manage heterogeneous compute resources.


GPU Superpowers
Co-processor compute virtualization enables powerful new capabilities including on demand elastic GPUs, seamless multi-node scaling, and automatic high availability.

Runs in Userspace
Bitfusion Core runs in userspace, ensuring it can run securely, on almost any OS, and without requiring any changes to existing hypervisors or cloud infrastructure.

No Code Changes
As completely transparent middleware, Bitfusion Core requires zero application changes. That’s why including Core in the AI Platform makes for a perfect combination.

Bitfusion Core uses a client-server architecture where servers provide the GPU resources for the cluster, and clients are where end user applications are run.

  • Bitfusion Application Instance (Client): The machines where the end user will be running their application. It can be a GPU instance, but it is not required that it be.
  • Bitfusion GPU Instance (Server): The machines which provide GPU resources to the cluster.

There are many flexible configurations which are possible using Bitfusion Core. However, the most common are: One-to-Many, Many-to-One, and Many-to-Many.

Bitfusion Core