March gcc 3.0 and 3.1 Bootstraps Fail 34% of Time

Richard Kenner kenner@vlsi1.ultra.nyu.edu
Tue Apr 3 04:22:00 GMT 2001


        >> 43% success 60% success

        Nathan> This is really bad.

    Yes, it is.

    It is, in fact, entirely unacceptable.

For the record, I agree.  Although I feel that trying for a goal of a
very small number for the mainline (but not the branch) is
unrealistic, I too feel these numbers are too high.  My suggestion
would be a goal in the 10-15% range.



More information about the Gcc mailing list