A (very) slight speed improvement for iterating over bytes

[gvanrossum]

My mentee @xvxvxvxvxv noticed that iterating over array.array is
slightly faster than iterating over bytes. Looking at the source I
observed that arrayiter_next() calls getitem(ao, it->index++) wheras
striter_next() uses the idiom (paraphrased)

item = PyLong_FromLong(seq->ob_sval[it->it_index]);
if (item != NULL)
    ++it->it_next;
return item;

I'm not 100% sure but I think that the second version has less
opportunity for the CPU to overlap the index++ operation with the
rest of the code (which in both cases involves a call). So here I am
optimistically incrementing the index -- if the PyLong_FromLong() call
fails, this will leave the iterator pointing at the next byte, but
honestly I doubt that anyone would seriously consider resuming use of
the iterator after that kind of failure (it would have to be a
MemoryError). And the author of arrayiter_next() made the same
consideration (or never ever gave it a thought :-).

With this, a loop like

for _ in b: pass

is now slightly faster than the same thing over an equivalent array,
rather than slightly slower (in both cases a few percent).

[corona10]

@gvanrossum

I measured that 2-3% speed improvement on follow script.

Benchmark

./python.exe -m pyperf compare_to master.json pr-21705.json
Mean +- std dev: [master] 443 us +- 10 us -> [pr-21705] 432 us +- 10 us: 1.02x faster (-2%)

Script

import pyperf

runner = pyperf.Runner()
runner.timeit(name="bench bytes iter",
              stmt="""for _ in datas: pass""",
              setup = """datas = b'a'*9999 """
)

[gvanrossum]

@corona10 Thanks! I'll take it, but I'll wait a bit until for else to pipe up. Maybe there's a reason that I've forgotten we need to be extra careful with not incrementing the index until we're sure we have a result.

I also have a question about pyperf -- I'm not sure how I get JSON output from the script you show.

[corona10]

@gvanrossum
You can get the JSON file by executing the attached script with the option :)

./python.exe bench_iter.py -o pr-21705.json

For more detail: https://github.com/psf/pyperf

[gvanrossum]

Thanks! That is truly magical -- apparently pyperf parsers the command line even when it is not the __main__ program. I didn't see anything about this in the pyperf docs.

[gvanrossum]

(And I now see that the pyperf's README file has this up front -- but I didn't think to look there, I only looked on readthedocs. :-( )

[bedevere-bot]

@gvanrossum: Please replace # with GH- in the commit message next time. Thanks!