MPI_LOCK semantics P.S.

Marc Snir (snir@watson.ibm.com)
Mon, 31 Mar 1997 20:10:52 -0400

Messages sorted by: [ date ][ thread ][ subject ][ author ]
Next message: Marc Snir: "A new version of the chapter is at the ftp server at Wisconsin"
Previous message: Marc Snir: "semantics of lock/unlock"

Another reason for the stronger semantics for MPI_LOCK that I propose is
the NO_CHECK assertion: the user asserts that nobody else is executing a
conflicting access. When? For how long? If we do not give a time
boundwhen the rmeote window access is bound to occur, how can we make such
an assertion?

Next message: Marc Snir: "A new version of the chapter is at the ftp server at Wisconsin"
Previous message: Marc Snir: "semantics of lock/unlock"