[Mpi-forum] MPI-next

N.M. Maclaren nmm1 at cam.ac.uk
Sat Feb 22 04:49:15 CST 2014

On Feb 22 2014, Rolf Rabenseifner wrote:
>as input for your discussions I'll give you some background about the 
>MPI-2.1/2.2/3.0 book selling which includes some lessens learned,
>which may influence your decisions on MPI-3.1 and 4.0,
>see conclusions at the end.

An alternate view:

>- About 30% of MPI programmers who physically get in touch with the book,
>  buy the book (it is the cheapest useful screen extension they can buy ;-)
>  ==> I expect a similar rate in the MPI Forum

Yes, indeed, but NOT among the MPI-using community as a whole.

Few people have much difficulty actually using the MPI facilities, once
over the initial learning curve, and rarely look at ANY reference on it.
Their ongoing problems are almost all due to the management of the data
and its distribution (surprise, surprise!)

A high proportion of them program from multiple locations, usually work
and home, but often have multiple offices, and some do design and/or
coding at external meetings.

I doubt that even 30% of the MPI-using community look at the actual
standard more than a couple of times a year and, for those people, it
isn't just the price but the shelf space and weight of a 1,000 page
book (which is what MPI 4.0 will be).

In Cambridge, I doubt that more than 5% of the MPI-using community
would buy the book even the offer and book were delivered to their desk.
But we have always been one of the most distributed and network-centric
universities in the world.  It might be higher elsewhere.

Nick Maclaren.

More information about the mpi-forum mailing list