What HPC people say

What they really mean

It’s all about the science

I’m going to put up a few science slides first to give me an excuse to talk about our supercomputers for the next hour

Our nation needs this big supercomputer to compete with other nations

Our centre needs a big supercomputer to beat our competing centre in our own country

We must collaborate internationally to get to exascale

We collaborate internationally because then project partners aren’t competing for the same funding agency

We need funding for operations not hardware capital

Don’t you dare reduce my hardware funding, but we do need money for the electricity too

People are the most important asset in HPC

I’m sure this is the right thing to say, but I still want money for hardware

We need to invigorate the HPC supplier ecosystem

I want more money to spend at my HPC centre

The national economy will benefit if we drive HPC usage in industry

I want the government to buy me another supercomputer and fund my business development team

There is a shortage of HPC talent

I have no interest in paying the market rate for existing HPC talent or contractors and so I want the government to develop more cheap recruits for me

Software licences are too expensive

I want to spend money on hardware

TOP500 ranking doesn’t matter

My supercomputer is at the top of the list anyway, so I am free to preach

TOP500 ranking doesn’t matter

My supercomputer would not make a high position on the list, so I will ‘diss’ the list

We need to keep the size of our supercomputer confidential for competitive reasons

We’ve always kept the size of our supercomputer a secret, but no-one really knows why

Our new initiative is OpenThing

‘Thing’ is limited to the technology or partners we choose but we are hoping if we call it ‘Open’ no-one will notice

We have a much richer user base now than traditionally – data analytics, interactive, python, …

I prefer the old days when we had users we knew and they were all doing big Fortran+MPI jobs

We need to collaborate with industry to make this HPC vision succeed

We got less money from NSF/DOE/Horizon2020 than we were wanted so we are hoping we can shake down industry for the rest

We have a great sense of self-deprecating humour as a HPC community

We find this blog post offensive and don’t recognise ourselves at all