Covalent’s serverless HPC architecture allows you to easily scale jobs from your laptop to your HPC/Cloud
How it Works
Construct
- Decorate any python function to convert it into a task / workflow
- Work locally and with the IDE of your choice, including Jupyter notebooks
- Workflows can be cross-language and cross-platform
import covalent as ct
@ct.electron
def spacer(word): return word+" "
@ct.electron
def join(*a): return "".join(a)
@ct.lattice
def workflow(name):
return join(spacer("hello"),spacer(name)) +" !!"
id=ct.dispatch(workflow)(name="Shore")
ct.get_result(id,wait=True)
Execute
- Prototype your workflows locally and then run them anywhere
- Switch between hardware types and locations with a single line of code
- On-demand cloud bursting within a single hybrid workflow
Monitor
- Visually engage with your workflows directly with Covalent’s browser-based UI
- Get live status updates for tasks, workflows and hardware
Collect Results
- Automatically add checkpoints to your workflows
- Keep track of and organize your experiments using the Covalent UI
- Automatically package experiments that can easily be shared/reproduced
Covalent supports a growing number of languages and executors
Searching for a specific plugin? Create your own or get started contributing.
Features
Open source
Covalent is open source.
Get the code here.
It’s pythonic!
Reproducibility
Covalent automatically stores and saves the run of every experiment in a reproducible format
PennyLane integration
Gain access to features such as smart scheduling and automatic resource allocation when running quantum tasks with PennyLane
*Coming Soon
Little-to-no overhead
No need to learn any new syntax or mess around with YAML
Browser-based UI
Monitor the state of your workflows (electrons + lattices) in real time
Resources
Stay up to date
FAQs
Yes. Covalent is an Open Source-first platform that anyone can use for free including self-hosting in their own cloud.
Covalent is used by data scientists and computational scientists across industries and disciplines, wherever there is a necessity for running experiments on advanced computing hardware such as GPUs, CPUs, QPUs and others.
Yes, we encourage people to host Covalent Open Source on their own cloud. Learn how here.
While most other workflow tools are focused on the “when” (when to schedule an event), Covalent is focused on the “where” (what type of hardware, and in what modality – cloud vs. on-prem).
Gain access to features such as smart scheduling and automatic resource allocation when running quantum tasks with PennyLane. No need to learn any new syntax or mess around with YAML
The best places to ask questions about Covalent are on our Github Discussions board or in our Slack channel.
Join the Covalent community
