This is the mail archive of the gcc-patches@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]

Re: PATCH RFA: Add -static-libstdc++ option to g++


This is an alloc of 1.7G, on a 32-bit box, and your kernel is refusing to
allocate the memory.  I guess you could argue that a tool should not need
more memory than the program it's debugging, but that's a stretch.  But
for only a 260M program this is odd.

Valgrind definitely need at least 9/8 more memory (i.e. +112.5%) just to track validity of the bytes. (Plus all the metadata about the blocks being allocated).

1.7G is a lot though.

According to the Valgrind mailing lists, this can happen with large
statically-allocated data.  Is there a lot of statically-allocated data
in that program?

If it's Fortran, I'd bet on it.

Paolo

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