Nvidia H100's VMs are now available to start right away on Paperspace by DigitalOcean. Try it out by going to start an H100 VM. ⚡️

What you’ll have access to:

• Nvidia H100: Access H100 machines in either an H100x1 (single chip) or H100x8 (entire H100 host) configuration with everything working straight out of the box on the update ML in a Box VM template
• Faster Networking: By the end of January, all users on H100x8 configurations will have access to 3.2TB interconnect speeds for intense multi-node training runs

On-Demand Access & Guaranteed Access

Start H100 on-demand

Start Now

Guarantee access to H100s
(term discounts & multi-node training available)

Schedule a Call

Other Improvements

Simplified Container Registry Experience

• Redesign of the container registry experience to provide field validation/checks and enhanced management of existing private container registries
• New experience in adding containers into Gradient deployments to ensure successful deployment starts

Endpoint Security on Gradient Deployments

• When creating a deployment, you can choose whether to set the endpoint to public or protected. A protected endpoint is secured with basic access authentication encoded tokens
• When a user tries to access a protected endpoint they are required to supply a username/password

Improved Cloud Networking

• Internet access now supports sustained 2Gbps connectivity inbound and outbound per VM

Paperspace has been working to provide users with faster storage speeds, and we are excited to announce that NVMe storage is now available in NY2! VM and shared drive creates are now backed by NVMe when capacity is available, and you can expect read/write speeds to improve by as much as 2.5x compared to resources created before March 2023.

Over the last several months, we have also been rapidly expanding the cloud, specifically with Nvidia Ampere Series GPUs (A4000, A6000, A100). You can also join the waitlist to get early access to Nvidia H100 machines. Waitlist

Fixes & Improvements

• Gradient Deployments now have liveness and readiness health checks, so you can monitor the health of your containers. Learn more here.
• Fixed several architecture bugs related to shutting down notebooks, which will bring capacity back online faster for users.
• Deployment optimizations were made to more intelligently add and remove replicas when spec edits are submitted.
• The Core AMS region now has 10Gb internet uplinks.
• We updated our payment modal to better accommodate international users.

We improved Compute Limit management for team administrators.  Compute Limits assist teams in monitoring and restricting monthly compute spend by providing fine-grained control over compute usage.  You can now create and edit email alerts and absolute maxes for both teams and team members from the billing page.  Administrators will receive an email when an alert or max amount is reached, and users will be notified in-app if they are blocked from accruing additional compute.  We hope these changes will give administrators more insight and control over their monthly spend.

For more information about Compute Limits, please visit our documentation.

Fixes & Improvements

• Added the ability to edit team names
• Improved invoice clarity by adding a section to represent Gradient storage costs
• Improved visibility into Core machine status by more accurately rendering real-time machine state
• Fixed a bug where some users could not scroll on streaming Core machines
• Fixed a bug where non-US countries were not selectable during signup onboarding

Over the past few weeks, we've shipped a number of improvements to improve your product experience.

• Improved private network performance and stability in AMS1: A major upgrade has been completed for the network infrastructure in AM1 resulting in increased performance and stability of the network.
• Improved reliability of machine starts: Starting machines will now encounter fewer issues due to improved automatic remediation of event errors.
• Improved availability of GPUs: Resolved an issue where specific machine types GPU availability were lower than expected.
• Prevent users from accidentally creating machines from deleted custom templates: There was an issue that allowed users to create machines from deleted templates resulting in inaccessible machines. Users will now see an error message if they try to create a machine from a deleted template.
• Improved management of inaccessible machines: Users can now shut down and deactivate machines that are inaccessible from the list views and details view.
• All new Linux machines have a dynamic public IP by default: A public IP is currently required to access a Linux machine except on a private network. By default, a newly created machine will have a dynamic public IP assigned.
• Improved summary of billing charges on machine create: We have improved the machine create billing summary to better reflect the total cost for creating and running a machine. Prices are broken down into monthly and active charges to differentiate the costs of running machines vs having machines active.
• Improved network performance and reliability: We’ve made a change to our QOS policy that now results in greater network availability and reliability for all machines.

New docs come to Paperspace! announcement 

We're excited to introduce an entirely new unified docs experience for Paperspace! 

After maintaining several different systems for documenting different parts of the product, we're eager to announce that Paperspace docs are now available in a single location with a new unified theme and organizational structure!

You can now find Core documentation, Gradient documentation, and general Account Management documentation all in one place!

If you need a place to start, we recommend starting with the Core overview or the Gradient overview -- you'll be able to launch right into tutorials, guides, and reference materials designed to help you succeed with Paperspace.

Have an idea for how to improve Paperspace documentation further? Please send us a note with any comments or suggestions!

We're excited to announce some brand new Core experiences! Let's jump right into what's new.

All-new Linux SSH experience announcement 

We've reconfigured the Linux machine create experience to optimize for connecting to Linux machines via SSH.

We feel that a direct connection to a Linux machine is a fantastic experience. We'll still support Linux VMs in the browser, but if you get a chance, give SSH a try -- it's so easy to connect!

Managing machines just got a lot better announcement 

We've also released a substantial cleanup of the machines settings page in Core which has made it easier than ever to access and manage machine settings. 

Let's say for example we wanted to create a snapshot of our new machine -- easy!

Or let's say we wanted to update our machine name and adjust the autoshutdown timer? Also easy!

We've also made it easier to do things like assign public IPs, generate templates, and more!

Redesigned account settings improvement

We've also updated the global Paperspace account settings to the latest design system standard. 

You'll now find tabs for Profile, Security, and SSH Keys and in general you should now find it easier to access these important settings.

Dynamic public IP addresses improvement 

• We added support for dynamic public IP addresses which provide public IP addresses at a bare minimum of cost

Capacity upgrades improvement 

Meanwhile, we've also been busy adding plenty of capacity to Paperspace datacenters.

• We onboarded a new fleet of RTX4000 machines to the CA1 region
• We dramatically expanded GPU compute capacity in the NY2 region
• We added nearly 100TB in shared storage across regions
• And don't worry, we didn't forget about Europe! New capacity is coming soon!

Bugfixes fix 

• We fixed a bug that was sometimes causing utilization graphs to display inaccurately

We just made a number of improvements to help Core power users self-serve Paperspace resources. 

With this update, you can now create private networks, spin-up shared drives, and assign public IP addresses to any machines that you manage!

Self-serve private networks improvement 

First up, we're pleased to bring private networks to all Core users. When you create a private network, you create a shared resource pool for your team that is isolated from every other machine and customer on Paperspace.

Once you create a private network, you can add machines and drives to the network to share with team members.

Be sure to read the docs for more info!

Self-serve private storage improvement 

Next up, we've made it easy to share a drive among multiple Core machines. After you create a shared network, you can spin-up a shared drive and attach it to the network in a matter of seconds!

For more information on shared drives, check out the docs!

Self-serve public IPs improvement 

Finally, we've made it a lot easier to claim and assign public IP addresses! While previously it was possible to assign a machine to a public IP after the machine was created, we've now streamlined the process to make it more visible at the team level.

To claim a public IP, simply visit the Public IPs tab in the console and claim the address. (Note that Public IPs are region-specific.)

To assign the new public IP to a machine, all we need to do is use the Assign feature to select the machine we want to expose to the public web. That's all there is to it!

If you get stuck please read the docs to learn more or reach out to us with any questions. 

Bugfixes fix 

• We resolved a troublesome issue that resulted in erroneous invoices being sent to a small number of users
• We decreased errors related to over-provisioning on the Paperspace public cluster
• We improved the strategy for guaranteeing hot nodes and faster startup times on the Paperspace public cluster
• We fixed a number of small issues related to Windows 10 BYOL machines

We're pleased to announce a series of new GPU-backed instances available on both Core and Gradient featuring NVIDIA's Ampere microarchitecture!

Introducing all-new Ampere instances! announcement 

Announced in mid-2020, Ampere is the codename for NVIDIA's latest line of GPU accelerator cards. Competition for these cards has been fierce and we're happy to bring you four flavors of Ampere, anchored by the top-of-the-line A100.

In addition to the instances listed, we've also introduced 2-way, 4-way, and 8-way configurations for these cards. 

The full table of instances on Paperspace has been updated in the docs. In general, any instance made available on Core will arrive in Gradient shortly thereafter.

Multi-GPU also comes to Windows machines improvement  

One thing you might have noticed already is that multi-GPU instances in Core are no longer exclusive to Linux. You can now spin-up any multi-GPU instance on a Windows machine!

Check out the Paperspace console to get started. 

Model-backed deployments in Gradient Deployments improvement  

We added an important feature to Gradient Deployments: model-backed deployments! 

It's now possible to inject a model at deployment runtime which means Gradient is now able to fetch a model from the Gradient model registry directly. Models can also be referenced from an external S3 bucket.

For more information, read the docs or reach out if you'd like a demo!

State persistence bugs in Gradient Notebooks improvement  

We made substantial improvements to the way that application and cell state is managed in Gradient Notebooks. 

Previously, if you navigated away from a notebook while a cell was running and then returned to the notebook, the cell would sometimes lose its state. We're happy to have implemented a substantial fix to this issue and a number of other issues influencing state management.

If you have feedback for us, please drop us a line!

We've added RTX4000 and RTX5000 instances to CA1 and NY2 regions, as well as multi-GPU instances for Linux, and new low-cost CPU-only instances for Windows!

Introducing RTX4000 and RTX5000 announcement 

We're pleased to announce RTX4000 and RTX5000 instances are now generally available! 

These cards are based on NVIDIA's Turing microarchitecture and are more than 40% faster than their Pascal series counterparts.

Try RTX

Multi-GPU instances now available on Linux! announcement 

You can now access multi-GPU instances across all regions when selecting Linux as your OS!

P5000x2 instances start at $1.56/hr while P6000x2 instances start at $2.20/hr. 

Try multi-GPU

Low-cost Windows instances now available improvement 

We solidified CPU-only offerings for Windows instances and now provide C5 - C10 instances at an affordable hourly rate.

For just $0.08/hr you can run a full Core VM in the cloud!

Other Improvements

• We improved our backend error monitoring capabilities giving us substantially more insight into performance degradation and remediation
• We accelerated our equipment purchasing plan to provide new hardware faster to meet demand
• We re-wrote some business logic around storage capacity to be able to deliver much faster upgrades

Last month we introduced a product selector to the console. 

Today we've launched a second phase of console redesign, adding product and team selectors to the top navigation bar and removing the sidebar entirely.

You can now find all entities that used to appear in the sidebar on the horizontal subheading bar.

We hope you enjoy the breathing room. 

Full blogpost here.