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: [PATCH] Reduce GC overhead of the C++ lexer buffer


On Mon, 19 Jun 2006, Mark Mitchell wrote:

> > At the moment we have
> > 
> > cp/parser.c:292 (cp_lexer_new_main)  0: 0.0%  82369152:33.3%  0: 0.0% 
> > 15784576:37.2%  7
> > 
> > which is due to the fact that the initial lexer buffer size is no where
> > near a power-of-two value and we keep gc-reallocating the vector, doubling
> > its size.
> > 
> > With the following patch, this overhead is removed nearly completely
> > 
> > cp/parser.c:266 (cp_lexer_new_main)                       0: 0.0%         
> > 36: 0.0%          0: 0.0%          4: 0.0%          1
> > 
> > Comments?
> This is a good result -- but what test case did you use?

Everything with many functions.  In this case it was tramp3d-v4.

> Also, can you explain why it's important to use a power of two here?

Because all tokens are read up front, and if the buffer doesn't fit, its
size is doubled, hence the non-power-of-two-ness increases .  As the whole
buffer lives in garbage collected space and ggc-page.c makes sure to only
allocate power-of-two space, this wastes huge amounts of memory (never
touched, though) if the input size is not a power of two.


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