This is the mail archive of the gcc-regression@gcc.gnu.org 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] | |
> On Sat, 16 Jul 2005, Jan Hubicka wrote: > > > Joseph, > > any idea why the patch increased memory by up to 18%? > > It should if anything have reduced memory consumption by causing fewer > types to be allocated - you'll need to give more details of what sort of > objects there are more of being allocated and where they are being > allocated. > > One possibility might be that the optimizers are working better now they > have fewer integer types with the same width, and this causes them to > allocate more memory - but if so, the places where the optimizers care > about which of multiple types with the same width is being used would be > worth tracking down, as this could also apply with e.g. int and long, not > just with bit-field types. I am attaching logs from today and yesterday. Looks like tree optimizers are actually working worse, at least more RTL instructions are produced so about every pass consume more memory. Honza > > -- > Joseph S. Myers http://www.srcf.ucam.org/~jsm28/gcc/ > jsm@polyomino.org.uk (personal mail) > joseph@codesourcery.com (CodeSourcery mail) > jsm28@gcc.gnu.org (Bugzilla assignments and CCs)
Attachment:
combineO3.rep
Description: Text document
Attachment:
combineO3-yesterday.rep
Description: Text document
Attachment:
combineO2.rep
Description: Text document
Attachment:
combineO2.rep.bz2
Description: Binary data
| Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
|---|---|---|
| Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |