Re: Balanced trees

References	(8 earlier) <CAGGBd_qU3Zp3A4pymnDQfWynWZwFVrdHJpG=U0WZTap4HiymdA@mail.gmail.com> <lffv32$mqo$1@ger.gmane.org> <CAN1F8qU=2K6ysbpnu-JoUtfWfTToRsOaRek0dSQzjPx_sYzPKQ@mail.gmail.com> <CAMMy=OsYUzULRTRHwytiTszd44f3anz43=2DtGALHR3nBGe8JQ@mail.gmail.com> <CAN1F8qUqVnhjxXEUnub36GEwK_DD+ukKK2GwA5wYku0s=wpdgA@mail.gmail.com>
From	Daniel Stutzbach <stutzbach@google.com>
Date	2014-03-17 18:01 -0700
Subject	Re: Balanced trees
Newsgroups	comp.lang.python
Message-ID	<mailman.8233.1395104535.18130.python-list@python.org> (permalink)

Show all headers | View raw

[Multipart message — attachments visible in raw view] - view raw

On Mon, Mar 17, 2014 at 5:08 PM, Joshua Landau <joshua@landau.ws> wrote:

> Thanks.  First, I want to state that there are two aspects to my
>  claim. The first is that these benchmarks to not represent typical
> use-cases. I will not go too far into this, though, because it's
> mostly obvious.
>

I would love to have include macro-benchmarks.  I keep waiting for the PyPy
benchmark suite to get ported to Python 3...

> "Create from an iterator" gives me relatively different results when I
> run it (Python 3).
>

The graphs were originally created to compare vanilla Python with a Python
modified to use blist as the built-in list type.  I think I used Python
3.1, but I'm not certain.  As I recall, the built-in type has a few small
advantages over any third-party extension type, so that might be what
you're seeing.  Alternately, something may have changed between Python
versions.

> "Delete a slice" is fudged from its inclusion of multiplication, which
> is far faster on blists. I admit that it's not obvious how to fix
> this.
>

I could move the initialization into the timed part, similar to what I did
for sort (see below).  That has downsides too, of course, but it might be
an improvement.

> "First in, first out (FIFO)" should be "x.append(0); x.pop(0)".
>

Wow, I mangled that one badly.

> "Last in, first out (LIFO)" should use "pop()" over "pop(-1)",
> although I admit it shouldn't make a meaningful difference.
>

I like pop(-1) because it's explicit rather than implicit.  I agree it
shouldn't make a meaningful difference.

> "Sort *" are really unfair because they put initialisation in the
> timed part

That's a limitation of timeit.  The setup step is only executed once.  If I
put the initialization there, every sort after the first one would be
sorting a pre-sorted list.  If you compare the "Create form an iterator"
and "Sort a random list", you'll see that the initialization cost is
dwarfed by the sorting cost for n > 15 or so.

> and all have keys.

If you use classes with __lt__ methods instead of keys, the cost is
dominated by the calls to __lt__.  You're right that I should include both,
though.

> >>> python -m timeit -s "from random import choice; import blist; lst =
> blist.blist(range(10**0))" "choice(lst)"
> 1000000 loops, best of 3: 1.18 usec per loop
>
> >>> python -m timeit -s "from random import choice; import blist; lst =
> blist.blist(range(10**8))" "choice(lst)"
> 1000000 loops, best of 3: 1.56 usec per loop
>
> Lower size ranges are hidden by the function-call overhead.
> Perhaps this effect is to do with caching, in which case the limits of
> the cache should be explained more readily.
>

That's definitely a cache issue, which is always a risk with
micro-benchmarks.  I see growth even for the built-in list:

gnusto:~$ python -m timeit -s "from random import choice; lst =
list(range(10**0))" "choice(lst)"
1000000 loops, best of 3: 0.349 usec per loop
gnusto:~$ python -m timeit -s "from random import choice; lst =
list(range(10**8))" "choice(lst)"
1000000 loops, best of 3: 0.634 usec per loop

I agree it's more interesting to pick items randomly instead of always
querying the same index.  The overhead of choice() is kind of a problem,
though.  Since I'm only plotting up to 10**5, I'd expect these to look more
or less flat.

Thanks for all of the feedback.  I filed a bug with myself to improve the
metrics:
https://github.com/DanielStutzbach/blist/issues/64

-- 
Daniel Stutzbach

Thread

Re: Tuples and immutability Duncan Booth <duncan.booth@invalid.invalid> - 2014-03-07 09:33 +0000
  Re: Tuples and immutability Ben Finney <ben+python@benfinney.id.au> - 2014-03-07 22:04 +1100
  Re: Tuples and immutability Chris Angelico <rosuav@gmail.com> - 2014-03-07 22:11 +1100
  Re: Tuples and immutability Peter Otten <__peter__@web.de> - 2014-03-07 12:38 +0100
  Re: Tuples and immutability Chris Angelico <rosuav@gmail.com> - 2014-03-07 22:45 +1100
  Re: Tuples and immutability Alister <alister.ware@ntlworld.com> - 2014-03-07 11:51 +0000
    Re: Tuples and immutability Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-07 11:23 -0700
  Re: Tuples and immutability Roy Smith <roy@panix.com> - 2014-03-07 08:37 -0500
  Re: Tuples and immutability Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2014-03-08 15:17 +1300
    Re: Tuples and immutability Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-07 21:17 -0700
      Balanced trees (was: Re: Tuples and immutability) Marko Rauhamaa <marko@pacujo.net> - 2014-03-08 10:34 +0200
        Re: Balanced trees (was: Re: Tuples and immutability) Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-08 04:03 -0700
          Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-08 13:26 +0200
        Re: Balanced trees (was: Re: Tuples and immutability) Dan Stromberg <drsalists@gmail.com> - 2014-03-08 11:58 -0800
        Re: Balanced trees Mark Lawrence <breamoreboy@yahoo.co.uk> - 2014-03-08 20:37 +0000
          Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-08 23:21 +0200
            Re: Balanced trees Roy Smith <roy@panix.com> - 2014-03-08 17:22 -0500
              Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-09 11:17 +0200
            Re: Balanced trees Dan Stromberg <drsalists@gmail.com> - 2014-03-08 17:31 -0800
              Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-09 11:27 +0200
                Re: Balanced trees Dan Stromberg <drsalists@gmail.com> - 2014-03-09 14:20 -0700
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-09 23:32 +0200
                Re: Balanced trees Dan Stromberg <drsalists@gmail.com> - 2014-03-09 14:37 -0700
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-09 23:43 +0200
                Re: Balanced trees Dan Stromberg <drsalists@gmail.com> - 2014-03-09 15:08 -0700
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-10 00:24 +0200
                Re: Balanced trees Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-09 18:04 -0600
                Re: Balanced trees Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-10 03:24 +0000
                Re: Balanced trees Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-10 03:24 +0000
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-10 08:16 +0200
                Re: Balanced trees Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-10 08:53 +0000
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-10 11:41 +0200
                Re: Balanced trees Ned Batchelder <ned@nedbatchelder.com> - 2014-03-10 06:57 -0400
                Re: Balanced trees Rustom Mody <rustompmody@gmail.com> - 2014-03-10 09:01 -0700
                Re: Balanced trees Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-11 02:02 +0000
                Re: Balanced trees Roy Smith <roy@panix.com> - 2014-03-10 22:20 -0400
                Re: Balanced trees Chris Angelico <rosuav@gmail.com> - 2014-03-11 13:29 +1100
                Re: Balanced trees Mark Lawrence <breamoreboy@yahoo.co.uk> - 2014-03-10 09:51 +0000
                Re: Balanced trees Roy Smith <roy@panix.com> - 2014-03-10 09:59 -0400
                Re: Balanced trees Chris Angelico <rosuav@gmail.com> - 2014-03-11 03:20 +1100
                Re: Balanced trees Chris Angelico <rosuav@gmail.com> - 2014-03-11 03:24 +1100
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-10 19:08 +0200
                Re: Balanced trees Chris Angelico <rosuav@gmail.com> - 2014-03-11 04:17 +1100
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-10 19:34 +0200
                Re: Balanced trees Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-13 12:40 -0600
                Re: Balanced trees Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-13 23:57 +0000
                Re: Balanced trees Dan Stromberg <drsalists@gmail.com> - 2014-03-13 20:12 -0700
                Re: Balanced trees Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-14 05:02 +0000
                Re: Balanced trees Roy Smith <roy@panix.com> - 2014-03-10 19:24 -0400
                Re: Balanced trees Chris Angelico <rosuav@gmail.com> - 2014-03-11 10:27 +1100
                Re: Balanced trees Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-11 01:26 +0000
                Re: Balanced trees Chris Angelico <rosuav@gmail.com> - 2014-03-11 12:45 +1100
                Re: Balanced trees Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-10 21:38 -0600
                Re: Balanced trees Chris Angelico <rosuav@gmail.com> - 2014-03-11 15:28 +1100
                Re: Balanced trees "Rhodri James" <rhodri@wildebst.org.uk> - 2014-03-12 00:57 +0000
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-11 12:12 +0200
                Re: Balanced trees alex23 <wuwei23@gmail.com> - 2014-03-12 10:13 +1000
                Re: Balanced trees Alister <alister.ware@ntlworld.com> - 2014-03-11 09:25 +0000
                Re: Balanced trees Antoon Pardon <antoon.pardon@rece.vub.ac.be> - 2014-03-12 10:08 +0100
                Re: Balanced trees Tim Chase <python.list@tim.thechases.com> - 2014-03-10 11:33 -0500
                Re: Balanced trees Chris Angelico <rosuav@gmail.com> - 2014-03-11 03:39 +1100
                Re: Balanced trees Dan Stromberg <drsalists@gmail.com> - 2014-03-10 18:05 -0700
                Re: Balanced trees Roy Smith <roy@panix.com> - 2014-03-10 22:13 -0400
                Re: Balanced trees Dennis Lee Bieber <wlfraed@ix.netcom.com> - 2014-03-10 19:57 -0400
        Re: Balanced trees Joshua Landau <joshua@landau.ws> - 2014-03-15 01:13 +0000
          Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-18 00:05 +0200
            Re: Balanced trees Dan Stromberg <drsalists@gmail.com> - 2014-03-18 12:26 -0700
              Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-18 22:55 +0200
                Re: Balanced trees Dan Stromberg <drsalists@gmail.com> - 2014-03-18 14:45 -0700
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-19 00:03 +0200
                Re: Balanced trees Dan Stromberg <drsalists@gmail.com> - 2014-03-18 15:21 -0700
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-19 01:11 +0200
                Re: Balanced trees Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-19 01:15 +0000
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-19 10:49 +0200
                Re: Balanced trees Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-19 13:42 +0000
                Re: Balanced trees Marko Rauhamaa <marko@pacujo.net> - 2014-03-19 15:54 +0200
                Re: Balanced trees Roy Smith <roy@panix.com> - 2014-03-19 10:06 -0400
                Re: Balanced trees Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-19 01:15 +0000
                Re: Balanced trees Ethan Furman <ethan@stoneleaf.us> - 2014-03-19 08:15 -0700
                Re: Balanced trees "Rhodri James" <rhodri@wildebst.org.uk> - 2014-03-20 02:16 +0000
                Re: Balanced trees Dan Stromberg <drsalists@gmail.com> - 2014-03-19 19:34 -0700
                Re: Balanced trees Chris Kaynor <ckaynor@zindagigames.com> - 2014-03-18 18:02 -0700
            Re: Balanced trees Daniel Stutzbach <stutzbach@google.com> - 2014-03-18 13:18 -0700
        blist in standard library  (was Re: Balanced trees) Mark Lawrence <breamoreboy@yahoo.co.uk> - 2014-03-15 12:31 +0000
        Re: Balanced trees Daniel Stutzbach <stutzbach@google.com> - 2014-03-17 14:16 -0700
        Re: Balanced trees Joshua Landau <joshua@landau.ws> - 2014-03-18 00:08 +0000
        Re: Balanced trees Daniel Stutzbach <stutzbach@google.com> - 2014-03-17 18:01 -0700
        Re: Balanced trees Joshua Landau <joshua@landau.ws> - 2014-03-18 07:46 +0000
      Re: Tuples and immutability Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2014-03-09 13:40 +1300
        Re: Tuples and immutability Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-08 19:39 -0700
          Re: Tuples and immutability Marko Rauhamaa <marko@pacujo.net> - 2014-03-09 11:35 +0200
          Re: Tuples and immutability Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2014-03-10 11:03 +1300
            Re: Tuples and immutability Terry Reedy <tjreedy@udel.edu> - 2014-03-09 19:00 -0400
            Re: Tuples and immutability Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-09 17:42 -0600
              Re: Tuples and immutability Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-10 02:37 +0000
                Re: Tuples and immutability Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-10 02:35 -0600
                Re: Tuples and immutability Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-10 09:13 +0000
                Re: Tuples and immutability Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2014-03-11 18:15 +1300
              Re: Tuples and immutability Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2014-03-11 18:03 +1300
                Re: Tuples and immutability Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-11 04:39 -0600
                Re: Tuples and immutability Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-11 16:46 +0000
                Re: Tuples and immutability Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2014-03-12 10:23 +1300
                Re: Tuples and immutability Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-11 17:06 -0600
                Re: Tuples and immutability Steven D'Aprano <steve+comp.lang.python@pearwood.info> - 2014-03-12 23:20 +0000
                Re: Tuples and immutability Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-12 19:35 -0600
                Re: Tuples and immutability Terry Reedy <tjreedy@udel.edu> - 2014-03-12 22:09 -0400
      Re: Tuples and immutability Gregory Ewing <greg.ewing@canterbury.ac.nz> - 2014-03-09 13:45 +1300
        Re: Tuples and immutability Ian Kelly <ian.g.kelly@gmail.com> - 2014-03-08 19:55 -0700

csiph-web