Unfortunately I see rather few people doing serious optimizing knowledgeably at all in research computing. (I know how and where it's done well, of course.) T

Unfortunately I see rather few people doing serious optimizing knowledgeably at all in research computing. (I know how and where it's done well, of course.) That has global effects on productivity of users, though some of the biggest gains can come from more global optimization, like in the use of MPI.