One of the world’s largest supercomputers lived for only 10 minutes

DMCA / Correction Notice
- Advertisement -

There was a time when supercomputers were only available to a handful of organizations, mostly governments, public research facilities, and scientific bodies. The rise of cloud computing and the widespread availability of sophisticated cloud workload management (CWM) tools have significantly lowered the barrier to entry.

- Advertisement -

Only last week, Yellowdog, a CWM specialist based in Bristol, United Kingdom, assembled a virtual supercomputer using its proprietary platform, and at its peak, which lasted about 10 minutes, it generated an army of more than 3.2 million vCPUs. collected.

While it was nowhere near as powerful as the Fugaku, it was enough to propel it into the top 10 of the world’s fastest supercomputers by at least a few minutes.


(PSA: By the way, we’re going to update our black friday web hosting deals And black friday website builder deals Pages at least once per day until Cyber ​​Monday)

short term

- Advertisement -

This provision, which was done on behalf of a pharmaceutical company, helps a popular drug discovery application run as a single cluster. The raw cost of the project is approximately $65,000 on the back of the envelope calculations.

This is accounting for 33,333 instances of the AWS 96-core c5.24xLarge instance, one of the instances used during the run (essentially similar to a Baremetal server or dedicated server) at $1.6013 per hour. So this accounts for $53,376 per hour or $57,824 for the entire length of the session (65 minutes in total).

AWS’s Colin Bridger said, “With access to this on-demand supercomputer, researchers were able to analyze and screen 337 million compounds in 7 hours. To replicate that, it took two months using their on-premises system.” Will take.”


The extraordinary thing is that this type of firepower is available to anyone who can afford it. And it’s based on the kind of hardware that drives our cloud computing world: web hosting, website builders, cloud storage, email services, and more.

CWM platforms have evolved over the years to incorporate algorithms and machine learning capabilities to choose the best source of computation, regardless of its origin or type.

For example, a cloud provider may have a minimum cost spot compute, but this would not be selected if the customer-determined areas were not available, or if a sufficient number of servers of the instance types actually required were not available from that cloud provider. , in this case another source of computation will be chosen.

- Advertisement -

Stay on top - Get the daily news in your inbox

Recent Articles

Related Stories