Get productive with cluster computing in R

Bramblecloud gives you pre-configured cloud clusters to speed up R computations. It works right out of the box, providing you with additional computing power when you need it without the hassle of setting up the cloud service infrastructure first. Our clusters are hosted on Amazon’s AWS platform and you can reliably integrate spot instances into your R cluster to save costs. Special rates for university affiliates are available. Here’s an overview of Bramblecloud’s features:

Speed up computations

Use cloud computing power to get your job done faster.

Easily save 90% of computation time with parallelized algorithms.

Easy setup

No need to waste time on configuring your cluster. Moving your computations to the cloud is just a few clicks away.
Get an RStudio window into the cluster where you can run your computations

Pay only what you use

There’s no subscription fee or hidden costs. You only pay for the time that you use.
Bramblecloud is pay-per-hour. Just like software as a service should be.

High data security

Automatic firewall management to keep your computations secure.
Your data is safe with us.

Cost control

Automatically stop idle clusters to prevent costs from accumulating when the job is done.
Get your full cost overview live from the dashboard.

Use spot instances

We let you add spot instances to your R cluster. You name the price.
Bramblecloud takes care of administering spot instances, so that you can speed up your computation even more for little money.

Complete your simulations in minutes instead of hours

Macro-parallelization makes sense in a lot of simulations like Monte Carlo, evolutionary modelling, adaptive mesh refinement, lattice simulations and molecular dynamics, numerical solvers for differential equations or optimization problems. If you can run it on the 4 CPU cores of your desktop, then you can also run it on 256 cores in the cloud, giving you dramatic speed-ups!

The chart on the right shows the computation times of a simple Monte Carlo integration of sin(x) from 0 to π using 1 trillion simulations. This embarassingly parallel toy example exhibits almost perfect speed gains from parallelization. Like that, you can avoid running computations over night and get more done in less time.

And the best: using spot instances you can lower the cost of large clusters. The computations with 128 and 256 cores were done using 1 permanent machine and 3 / 7 spot instances, respectively.

Computation time in hours w.r.t. number of CPUs


Try now: 2 hours free*

Small Instance US/EU/Asia

$040per instance/hour
  • 2 cores
  • 3.75 GB RAM

Large Instance USA

$229per instance/hour
  • 32 cores
  • 60 GB RAM

Large Instance EU

$254per instance/hour
  • 32 cores
  • 60 GB RAM

Large Instance Asia

$277per instance/hour
  • 32 cores
  • 60 GB RAM

Spot Instance US/EU/Asia

You Name Itper instance/hour
  • 32 cores
  • 60 GB RAM
All prices are without tax. Depending on your location, additional VAT or sales tax charges may apply.
We have got special discounts for students and researchers affiliated with universities. Please ask about it here!
* New customers get ¢240 net credit, enough to test our small machine for 6 hours.

Latest news

Lower prices

Good news for our customers: we decided to slash our prices for large instances, effective immediately! Our large US instance you can now get for $2.29 instead of $2.67. That’s a 14% reduction, and we’ve […]

Updated servers

R 3.1.1 has been released recently, so we’ve updated our machines. That also includes all the pre-installed machine learning packages. See the list of R-packages here.

Latest R and dplyr

We’ve updated our servers and are now running the latest R version 3.3.0 and the latest RStudio. Due to popular demand, we’ve also added dplyr to the list of packages that come right out of […]

Subscribe to our Newsletter