Retiring Sherlock 1.0

The day has come to say goodbye to Sherlock 1.0.

After almost 4 years of service, the Sherlock 1.0 environment has definitely been retired. Starting today, when you connect to the Sherlock 1.0 login nodes (with ssh, you will automatically be redirected to the new Sherlock 2.0 environment.

One Sherlock again

Even more significantly, it means that Sherlock is whole again, and that the Sherlock 1.0/2.0 distinction is not relevant anymore. There's only one Sherlock again, and this effectively concludes the transition process that we initiated at the beginning of 2017.

Looking back

Started in early 2014, with about 120 compute nodes, Sherlock was the first campus-wide high-performance computing cluster, providing computing resources to everyone doing research at Stanford. With tremendous support from the Stanford research community, it now features more than 1,200 compute nodes, largely owned by PIs and Faculty members.

Today, Sherlock is used by more than 3,200 users in 550 PI groups. Its 22,000+ CPU cores and 650 GPUs provide about 1.5 PFlops of computing power and make it one of the top academic supercomputers in the world.

We hope it will continue to grow and thrive, to keep serving the Stanford research community and support amazing science!