At least 7 python packages were infected:
But this raises important questions about trust and allowing anyone to build on HPC clusters.
What do you think?
A wonderful resource both for getting your data out there and as a repository to draw from:
Clearly a major improvement in performance.
However, some cleanup is needed… open an new tab and be annoyed:
ugly and annoying, neh? looks like safari with extra adds!
Click on the gear (upper right) and be presented with:
Deselect all the you want to be rid off:
And a clean day is had by all your new tabs and windows!
With the very rapid uptake of R in many fields, its limitations in distributed or large system architecture has been a hurdle in a number of fields (eg: Bioinformtics, Environmental Sciences, Economics). Oak Ridge National Laboratories (ORNL) has picked up on this challenge.
They have been working onbgdR. There is a clear explanation of their priorities and plans in the April 2017 issue of HPCWire.
These extensions are seemlessly available on all nodes of nscc.colby.edu today.
A good review of some of the current cloud incumbents using linpack:
Comparative benchmarking of cloud computing vendors with High Performance Linpack
Penguin computing has also made an analysis of the POD (Penguin on Demand) service, which strictly speaking is not a cloud service but an on demand dedicated service.
This is my presentation to ITS about the current state of HPC here at Colby.
The intention with this blog space within the nsc site is to jot down quick notes.
If the note proves useful, I will add the content within the regular portion of the site, such as in FAQs, software or HPC.