This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: GCC inline parameters (PR 10160 testcase)

> What kind of performance difference does this give us?  I fear that
> this change will speed up compilation but slow down execution and
> would therefore like to see some performance data of compiled code
> also.

I'm not in a position where I can make these measurements.  Thus,
it would be greatly appreciated if others could do this.  I can say
that I haven't seen any appreciable change in bootstrap times of
GCC with the patch.

I am certain that there isn't one set of parameters that will give the
best performance of compiled code in all situations.  I believe the
default inline parameters should be set to values that enable building
code like lyx and pooma in a reasonable amount of time, utilizing a
reasonable amount of memory.  Where you draw the line is obviously
somewhat arbitrary but users that want the last 10% in execution
performance can always use the available options to do this.

The average user of gcc doesn't want to undertake a research project
into why a build failed.  For example,


J. David Anglin                        
National Research Council of Canada              (613) 990-0752 (FAX: 952-6602)

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]