JuliaHub was built for technical computing and is geared toward helping teams interested in using scientific methods to model and simulate real-world phenomena. JuliaHub also helps developers build apps, transform data, and includes security and compliance tools for the enterprise. Last month, the team at JuliaHub revamped our standard tier offering and re-launched a free version of the platform that allows anyone to use up to 20 hours of compute time on the default machine for free every single month. This means that all of JuliaHub’s main features are available to use for free on our standard machine (r6i.large) of 2vCPU with 16GB RAM. This 20 hours resets each month and each additional hour on the standard machine is priced at just $0.15 per hour for batch jobs.
So what’s the big deal about our free offering?
Let’s start by explaining what makes the Julia ecosystem so special.
A Future Based in Scientific and Industrial AI
Prominent technologists and entrepreneurs Bill Gates and Peter Diamandis believe that current advances in machine learning and AI are enabling breakthroughs in traditional science. These two fields are coming together to make massive contributions to industries such as energy, finance, pharmacometrics, climate modeling, food systems, and more - providing hope to those who believe that humanity’s future can be even more abundant than its past.
Many of these scientific breakthroughs are now possible because machine learning and artificial intelligence can model concepts that previously took too much time. For example, in this article from Nature, Julia is enabling biologists to discover new science.
As ML/AI continue to allow machines to accelerate the simulation of real-world phenomena; scientists and engineers today are using new AI-based tools to figure out novel ways to manipulate and build upon knowledge of our physical world. Engineering new materials, new proteins for pharmaceuticals, and potentially new forms of energy are just a few of the breakthroughs that are possible and this is precisely what the Julia community has been building with a robust ecosystem of scientific machine learning capabilities free to use in JuliaHub.
Features for High Performance
One core pillar of JuliaHub is that the platform gives users access to different machines with memory and processing capabilities not available on your average laptop. This means that JuliaHub enables high-performance computing (HPC). HPC is about using supercomputers (virtual and in the cloud) to maximize the performance of a program or process. The Julia language’s distributed computing capabilities such as parallelism, multi-threading, and multi-processing, on JuliaHub allows anyone to perform complex calculations at extremely high speeds.
Thus, using the Julia language on JuliaHub is like getting a supercomputer at your fingertips.
Let’s look at a few features that make JuliaHub special:
-
CloudStation and Compute: Paid users can get access to 8 types of machine configurations including a V100 GPU with 8vCPUs.
-
JuliaHub Projects (for Collaboration): A powerful way to encapsulate and persist files, folders, datasets, and code across any application. Share your work with different team members.
-
JuliaIDE in VSCode: The full Julia REPL, package manager, 9,000+ packages, and more features inside VSCode. Extensible with more package features being built every day.
-
Pluto Interactive Notebooks: Run Julia code in a highly specialized notebook (similar to Jupyter) and get models, plots, and graphs to validate scientific research.
-
Julia Ecosystem Products: JuliaHub includes machine learning and artificial intelligence products such as JuliaSim: a next-generation cloud-based simulation platform.
-
Enterprise features: for paid users, JuliaHub includes reproducible jobs, traceability and audit information, package encryption (security), and compliance features*
*JuliaHub is SOC 2 compliant, TLS 1.2+ encryption; both client-facing & server-side, AES-256 bit encryption.
20 Free Hours Each Month. Usage-based Pricing After
JuliaHub is free to use on default machines for 20 thread hours every month with no credit card required.
Should you want to access higher-level machines (such as 32 virtual CPUs) - you just provide your credit card information and we bill you directly for usage past the free hours. Every month, your free hours reset so you can mix and match your usage of JuliaHub and launch custom jobs and applications based on your needs.
How long can I use the free compute?
On the default machine (2vCPUs, 16GB Memory):
- 20 hours (per thread) every month
A thread-hour is one Linux kernel thread made available to your launched job for an hour. If you are only using 1 thread, this means 20 hours of free compute. If you use 10 threads, your free hours would be spent in 2 hours.
Once you provide us your credit card information, you can access higher machines and you still get free hours included. For example:
- 4vCPUs, 16GB Memory
-
-
- 10 hours (threads per hour), paid usage after
-
- 8vCPUS, 128GB Memory
-
-
- 5 hours (threads per hour), paid usage after
-
If you’d like to see every available compute option in JuliaHub - here it is:
- 2 vCPUs, 8 GB
- 4 vCPUs, 16 GB
- 4 vCPUs, 32 GB
- 8 vCPUs, 32 GB
- 32 vCPUs, 128 GB
- V100vGPU, 8 vCPUs, 61 GB
- K80 vGPU, 5 vCPUs, 61 GB
Once you go over the free hours, the cost of using JuliaHub on the standard subscription is:
- $0.15/per thread hour for batch jobs
- $0.25/per thread hour for interactive
Setting a Spending Limit with CloudStation
When you start a job, you can set a limit to the amount of time (by hours) that this job will be active. Once the time limit is hit, the job will stop and spin down the application or product you are using. The max limit is 36 hours for the standard subscription, but higher tiers can be set with unlimited time. Admins can also set limits to users on their instance of JuliaHub.
For more information on the subscription types, including features included, visit this page.
JuliaHub is full of free features
Accessing high-performance computing capabilities in JuliaHub has never been easier. With access to multiple machines, over 9,000+ open source packages (including industry-leading mathematical and scientific machine learning), an integrated development environment, Pluto interactive notebooks, Git-based team collaboration features for version control…and more features being added every quarter. To follow along with new features, you can see our release notes.
JuliaHub’s managed infrastructure lets your teams do what they do best while we manage the high-performance computing technology stack. Start for free with JuliaHub today.