LAM/MPI General User's Mailing List Archives

From: Angel Tsankov (fn42551_at_[hidden])
Date: 2005-02-04 15:22:26

• Next message: Josh Hursey: "Re: LAM: Negative rank or communicator size?"
• Previous message: Curry Matthew J CRBE: "RE: LAM: Negative rank or communicator size?"
• In reply to: Curry Matthew J CRBE: "RE: LAM: Negative rank or communicator size?"
• Next in thread: Josh Hursey: "Re: LAM: Negative rank or communicator size?"

> Because they do return error codes, if something is wrong. I thought I
> remembered reading something in the standard about changing the
> behavior on an error, i.e., you can make it not do an MPI_Abort on error,
> IIRC.

But they do return error codes anyway! So why put an extra error value in a
variable that is intended for a rank or communicator size?

• Next message: Josh Hursey: "Re: LAM: Negative rank or communicator size?"
• Previous message: Curry Matthew J CRBE: "RE: LAM: Negative rank or communicator size?"
• In reply to: Curry Matthew J CRBE: "RE: LAM: Negative rank or communicator size?"
• Next in thread: Josh Hursey: "Re: LAM: Negative rank or communicator size?"