[Python-checkins] r46499 - in python/trunk: Include/pyport.h Include/unicodeobject.h Objects/stringobject.c

[Neal Norwitz]

On 5/28/06, Fredrik Lundh <fredrik at pythonware.com> wrote:
>
> >> needforspeed: added Py_MEMCPY macro (currently tuned for Visual C only),
> >> and use it for string copy operations.  this gives a 20% speedup on some
> >> string benchmarks.

How about extending this to list and tuple at least (for memset)?
Frame init does a memset, while tuple init does a loop.  I'm sure
there are other places that could make a diff.  getargs, obmalloc,
maybe.

Note that in non-optimizing builds, memset/memcpy are much faster than
the corresponding loops.  At least that was my experience on Linux/gcc
when profiling.

n