all messages for Emacs-related lists mirrored at yhetil.org
 help / color / mirror / code / Atom feed
From: Ihor Radchenko <yantar92@gmail.com>
To: pareto optimal <pareto.optimal@mailfence.com>
Cc: emacs-orgmode@gnu.org
Subject: Re: Poor org-babel-tangle-file performance with more than 100 trivial noweb-references
Date: Wed, 26 Jan 2022 19:43:15 +0800	[thread overview]
Message-ID: <87sftalo30.fsf@localhost> (raw)
In-Reply-To: <87k0enx5v6.fsf@localhost>

[-- Attachment #1: Type: text/plain, Size: 1168 bytes --]

Ihor Radchenko <yantar92@gmail.com> writes:

> Most of the CPU time is spent in org-babel-tangle-collect-blocks, which
> is basically another regexp search for all the source blocks in buffer.
> The scaling is still slightly non-linear - maybe your source block
> regexp is too complex:

After further investigation I found that it was not the problem with
source block regexp. The code was doing an extra backward regexp search
to find current headline. When there are no headlines in the Org file,
that search created quadratic scaling.

After caching the backwards search, the performance is further improved:
| N blocks | runtime | # of GCs |
|----------+---------+----------|
|       10 |   0.204 |        0 |
|       20 |   0.047 |        0 |
|       40 |   0.171 |        0 |
|       80 |   0.063 |        0 |
|      160 |   0.096 |        0 |
|      320 |   0.155 |        0 |
|      640 |   0.651 |        0 |
|     1280 |   0.419 |        0 |
|     2560 |   0.799 |        0 |
|     5120 |   1.628 |        0 |
|    10240 |   3.306 |        0 |
|    20480 |   5.633 |        0 |
|    40960 |  11.415 |        0 |

41k blocks in 12sec.

Graphical comparison:

[-- Attachment #2: benchmark1.png --]
[-- Type: image/png, Size: 8212 bytes --]

[-- Attachment #3: Type: text/plain, Size: 53 bytes --]


The scaling becomes strictly linear after this fix:

[-- Attachment #4: benchmark2.png --]
[-- Type: image/png, Size: 9648 bytes --]

[-- Attachment #5: Type: text/plain, Size: 3678 bytes --]


Best,
Ihor

-----
Code used to generate the final set of the graphs:

#+name: nocache
| N blocks | runtime | # of GCs |
|----------+---------+----------|
|       10 |   0.027 |        0 |
|       20 |   0.049 |        0 |
|       40 |   0.121 |        0 |
|       80 |   0.391 |        0 |
|      160 |   1.426 |        0 |
|      320 |   6.765 |        0 |
|      640 |  23.972 |        0 |

#+name: cache
| N blocks | runtime | # of GCs |
|----------+---------+----------|
|       10 |   0.030 |        0 |
|       20 |   0.067 |        0 |
|       40 |   0.065 |        0 |
|       80 |   0.081 |        0 |
|      160 |   0.214 |        0 |
|      320 |   0.597 |        0 |
|      640 |   2.225 |        0 |
|     1280 |   9.221 |        0 |
|     2560 |  36.966 |        0 |

#+name: cache-no-self
| N blocks | runtime | # of GCs |
|----------+---------+----------|
|       10 |   0.078 |        0 |
|       20 |   0.075 |        0 |
|       40 |   0.063 |        0 |
|       80 |   0.085 |        0 |
|      160 |   0.095 |        0 |
|      320 |   0.178 |        0 |
|      640 |   0.311 |        0 |
|     1280 |   0.819 |        0 |
|     2560 |   2.302 |        0 |
|     5120 |   8.878 |        0 |
|    10240 |  32.706 |        0 |

#+name: cache-no-self+fix
| N blocks | runtime | # of GCs |
|----------+---------+----------|
|       10 |   0.118 |        0 |
|       20 |   0.106 |        0 |
|       40 |   0.136 |        0 |
|       80 |   0.157 |        0 |
|      160 |   0.139 |        0 |
|      320 |   0.212 |        0 |
|      640 |   0.542 |        0 |
|     1280 |   0.797 |        0 |
|     2560 |   1.533 |        0 |
|     5120 |   4.651 |        0 |
|    10240 |  16.745 |        0 |

#+name: cache-no-self+fix+fix2
| N blocks | runtime | # of GCs |
|----------+---------+----------|
|       10 |   0.204 |        0 |
|       20 |   0.047 |        0 |
|       40 |   0.171 |        0 |
|       80 |   0.063 |        0 |
|      160 |   0.096 |        0 |
|      320 |   0.155 |        0 |
|      640 |   0.651 |        0 |
|     1280 |   0.419 |        0 |
|     2560 |   0.799 |        0 |
|     5120 |   1.628 |        0 |
|    10240 |   3.306 |        0 |
|    20480 |   5.633 |        0 |
|    40960 |  11.415 |        0 |


#+begin_src gnuplot :var d1=nocache :var d2=cache :var d3=cache-no-self :var d4=cache-no-self+fix :var d5=cache-no-self+fix+fix2 :file benchmark1.png
set title 'Tangle code performance timing'
US='u 1:2 w lp ps 2'
set xlabel "N blocks"
set ylabel "Time, sec"
set key top right
plot d1 @US t'Org 9.5.2 stable', d2 @US t'Org 9.6-dev', d3 @US t'Org 9.6-dev no self-verify', d4 @US t'Org 9.6-dev no self-verify + patch', d5 @US t'Org 9.6-dev no self-verify + 2xpatch'
#+end_src

#+RESULTS[e95dafc7253d218d2a9ed19caa75911660e72f77]:
[[file:/home/yantar92/.data/52/0930af-75ae-4d88-ae6a-d8dde39ecc72/benchmark1.png]]

#+begin_src gnuplot :var d1=nocache :var d2=cache :var d3=cache-no-self :var d4=cache-no-self+fix :var d5=cache-no-self+fix+fix2 :file benchmark2.png
set title 'Tangle code performance scaling (normalized)'
US='w lp ps 2'
set xlabel "N blocks"
set ylabel "Time, normalized by time at N=640"
set key top right
set yrange [0:100]
plot d2 u 1:($2/2.225) @US t'Org 9.6-dev', d3 u 1:($2/0.311) @US t'Org 9.6-dev no self-verify', d1 u 1:($2/23.972) @US t'Org 9.5.2 stable', d4 u 1:($2/0.542) @US t'Org 9.6-dev no self-verify + patch', d5 u 1:($2/0.651) @US t'Org 9.6-dev no self-verify + 2xpatch', x*x/870000 t'x^2', [0:10000] x*x/3500000 t'x^{2}', x/2400 t'x^{1}'
#+end_src

#+RESULTS[c69cdd0dfb08f37c73c6a202f415336155a390ab]:
[[file:/home/yantar92/.data/52/0930af-75ae-4d88-ae6a-d8dde39ecc72/benchmark2.png]]

      parent reply	other threads:[~2022-01-26 11:39 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-01-25  5:19 Poor org-babel-tangle-file performance with more than 100 trivial noweb-references pareto optimal
2022-01-25 14:03 ` Sébastien Miquel
2022-01-25 14:25   ` Ihor Radchenko
2022-01-25 14:11 ` Ihor Radchenko
2022-01-25 16:25   ` Tim Cross
2022-01-26 11:43   ` Ihor Radchenko [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87sftalo30.fsf@localhost \
    --to=yantar92@gmail.com \
    --cc=emacs-orgmode@gnu.org \
    --cc=pareto.optimal@mailfence.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this external index

	https://git.savannah.gnu.org/cgit/emacs.git
	https://git.savannah.gnu.org/cgit/emacs/org-mode.git

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.