Improving the Assignment server (AS)

It’s interesting running a research group at a University — it’s kinda like being the coach of a college football team in a sense — your team keeps on changing, as people are only there for a few years (3-5 years in our case, although granted our people are PhD’s).  So, it can be challenging when really great people leave (such as V Vishal, who wrote a lot of the behind the scenes code such as our analysis backend or the GPU core code), but one nice aspect is that new people bring new ideas.

I recently had a meeting with some newer members of the group (Dan Ensign, Dr. Peter Kasson, and Dr. Vincent Voelz) and they had some great ideas for improving the AS.  It will take us some time to implement the changes, but AS mods are on the first time in a while.  Our goal is to have the AS assign WU’s to donors based on what would maximize points for their hardware.  The idea being that if we can make sure that points are well aligned with what we need scientifically, more points = more science and it’s a "win-win." 

This AS change won’t be easy (and we’re still not sure if it’s even really doable), as there are so many different hardware configs possible, but we’re going to analyze AS data to see if it is.  The new hardware detection in V6 will play a critical role here too.

So, no promises (as this won’t be easy), but I thought people would be curious to know that this is very much something we’re working on.