Some numbers on gcc compiling itself

[Giovanni Bajo]

Steven Bosscher wrote:

>>> insn-recog.i 28.62 41.57 45.25
>>> insn-attrtab.i 27.58 37.69 36.66
>>
>> Can you please file two bug reports for these two files, attacching them?
>
> No because there are already two bug reports about this,
> see PR17895 and PR15524 for example.  For these two files
> the problem is the quadratic behavior in several places
> for SWITCH_EXPRs.

Law has a patch for 15524 which he's going to commit AFAICT. Did you do your
measurement before or after his patch? If before, can you check if his patch
solves all the problems or not?

Giovanni Bajo