We had an astonishing 837 responses to the MPI User Survey. Many thanks to all of you who filled out the survey!
The MPI Forum minions are busy analyzing the data — there’s a lot! We’ll have more definitive results later, but for now, see below the jump for a few quickie facts from the results.
First, some boring stats:
- 1,401 people started the survey; 838 people completed it (~60% completion rate)
- 100 more people looked at the cover sheet of the survey and never went any further it (chickens!)
- Average time to take the survey was 14 minutes.
Now, a few fun facts from the responses of the survey:
- 56% of the respondents classified themselves as “MPI application developer” or “Academic educator researcher” — excellent, that is a great demographic for our respondents
- On a scale of 1-5, 46% of the respondents rated their understanding of the MPI standard as a “3”, or “I have a good understanding of some parts of the MPI standard”
- Only 17% of respondents indicated that they expected to be able to upgrade to an MPI-3 implementation and run their existing/legacy MPI applications without recompiling them. That makes things quite a bit easier for us on the Forum!
- On the other side, 74% of respondents expect to be able to upgrade to an MPI-3 implementation and only need to recompile legacy MPI applications with no source code changes.
- An unsurprising 70% of respondents don’t use MPI-2 dynamic processes at all (MPI_COMM_SPAWN and friends).
What do all these percentages mean?
I don’t really know yet — but we’re working on much more detailed analysis. More to come in this area!