Re: A speed test on Apple's M1 processor

[Jacques Menu]

Hello Jonas and Jean,

I send you the details privately, not to clutter this list.

JM

> Le 12 sept. 2022 à 08:20, Jonas Hahnfeld via LilyPond user discussion 
>  a écrit :
> 
> On Sun, 2022-09-11 at 11:52 +0200, Jacques Menu wrote:
>> Native lilypond:
>> 
>> jacquesmenu@macmini:/Volumes/JMI_Volume/JMI_Developpement/lilypond/re
>> lease/binaries/lilypond/install/bin > time ./lilypond
>> Fischer_Suite_Sol_M_viola_II_2.23.12.ly
> 
> Based on the path, did you build with the scripts in release/binaries
> of the repository or did you build from scratch yourself? If the
> latter, did you compile the bytecode (make bytecode)?
> 
>> GNU LilyPond 2.23.12 (running Guile 2.2)
>> Processing `Fischer_Suite_Sol_M_viola_II_2.23.12.ly'
>> … … …
>> Preprocessing graphical objects...
>> Finding the ideal number of pages...
>> Fitting music on 54 or 55 pages...
>> Drawing systems...
>> Converting to `Fischer_Suite_Sol_M_viola_II_2.23.12.pdf'...
>> fatal error: failed files: "Fischer_Suite_Sol_M_viola_II_2.23.12.ly"
>> ./lilypond Fischer_Suite_Sol_M_viola_II_2.23.12.ly  7.52s user 0.75s
>> system 98% cpu 8.387 total
> 
> You have a "fatal error" when processing the converted file. Not sure
> where this happens, but maybe check that you are really compiling what
> you think you are...
> 
> Jonas

Re: A speed test on Apple's M1 processor

[Jonas Hahnfeld via LilyPond user discussion]

On Sun, 2022-09-11 at 11:52 +0200, Jacques Menu wrote:
> Native lilypond:
> 
> jacquesmenu@macmini:/Volumes/JMI_Volume/JMI_Developpement/lilypond/re
> lease/binaries/lilypond/install/bin > time ./lilypond
> Fischer_Suite_Sol_M_viola_II_2.23.12.ly

Based on the path, did you build with the scripts in release/binaries
of the repository or did you build from scratch yourself? If the
latter, did you compile the bytecode (make bytecode)?

> GNU LilyPond 2.23.12 (running Guile 2.2)
> Processing `Fischer_Suite_Sol_M_viola_II_2.23.12.ly'
> … … …
> Preprocessing graphical objects...
> Finding the ideal number of pages...
> Fitting music on 54 or 55 pages...
> Drawing systems...
> Converting to `Fischer_Suite_Sol_M_viola_II_2.23.12.pdf'...
> fatal error: failed files: "Fischer_Suite_Sol_M_viola_II_2.23.12.ly"
> ./lilypond Fischer_Suite_Sol_M_viola_II_2.23.12.ly  7.52s user 0.75s
> system 98% cpu 8.387 total

You have a "fatal error" when processing the converted file. Not sure
where this happens, but maybe check that you are really compiling what
you think you are...

Jonas


signature.asc
Description: This is a digitally signed message part

Re: A speed test on Apple's M1 processor

[Jean Abou Samra]

Hi Jacques,

Le 11/09/2022 à 11:52, Jacques Menu a écrit :

Hello folks,

I ran this test with a 7.4Mb, 55 page score, to compare the binaries 
provided by lilypond.org  and the natives ones I 
(finally) built locally on my 8 Gb RAM Mac Mini.
The necessary libraries have been installed using both MacPorts and 
Homebrew, since not all of them are supplied by a single source.



I find this surprising, since both MacPorts and Homebrew provide a 
LilyPond package. What are the dependencies missing on each side? Have 
you tried looking at their respective build definitions? The one for 
Homebrew is here:


https://github.com/Homebrew/homebrew-core/blob/HEAD/Formula/lilypond.rb

and for MacPorts it's here:

https://github.com/macports/macports-ports/blob/master/textproc/lilypond/Portfile

By the way, it's a very good idea (well, it's essential) to build 
LilyPond if you want to contribute to it, but for anyone just looking 
for native Apple Silicon binaries of LilyPond, just install from 
MacPorts or Homebrew. MacPorts is recommended for the time being because 
Homebrew uses version 2.22 with Guile 2. Since we did not support 
byte-compilation of Scheme code at that time, the Homebrew version will 
be slower (and in particular have bad startup time).


Best,
Jean

A speed test on Apple's M1 processor

[Jacques Menu]

Hello folks,

I ran this test with a 7.4Mb, 55 page score, to compare the binaries provided 
by lilypond.org and the natives ones I (finally) built locally on my 8 Gb RAM 
Mac Mini. 
The necessary libraries have been installed using both MacPorts and Homebrew, 
since not all of them are supplied by a single source.

Both versions are 2.23.12.
The speed increase is roughly 20 to 25%, thanks to the avoidance of Intel CPU 
instructions emulation.

I will send the test file privately to anyone interested with pleasure, and you 
can send me larger files for more tests.

A nice day!

JM

--

First convert the LilyPond file to version 2.23.12.

Native convert-ly:

jacquesmenu@macmini:/Volumes/JMI_Volume/JMI_Developpement/lilypond/release/binaries/lilypond/install/bin
 > time ./convert-ly Fischer_Suite_Sol_M_viola_II.ly > 
Fischer_Suite_Sol_M_viola_II_2.23.12.ly
convert-ly (GNU LilyPond) 2.23.12

convert-ly: Processing `Fischer_Suite_Sol_M_viola_II.ly'... 
Applying conversion: 2.20.0, 2.21.0, 2.21.2, 2.23.1, 2.23.2, 2.23.3, 2.23.4, 
2.23.5, 2.23.6, 2.23.7, 2.23.8, 2.23.9, 2.23.10, 2.23.11, 2.23.12


./convert-ly Fischer_Suite_Sol_M_viola_II.ly >   5.80s user 0.04s system 99% 
cpu 5.871 total
jacquesmenu@macmini:/Volumes/JMI_Volume/JMI_Developpement/lilypond/release/binaries/lilypond/install/bin
 > time ./convert-ly Fischer_Suite_Sol_M_viola_II.ly > 
Fischer_Suite_Sol_M_viola_II_2.23.12.ly
convert-ly (GNU LilyPond) 2.23.12

convert-ly: Processing `Fischer_Suite_Sol_M_viola_II.ly'... 
Applying conversion: 2.20.0, 2.21.0, 2.21.2, 2.23.1, 2.23.2, 2.23.3, 2.23.4, 
2.23.5, 2.23.6, 2.23.7, 2.23.8, 2.23.9, 2.23.10, 2.23.11, 2.23.12


./convert-ly Fischer_Suite_Sol_M_viola_II.ly >   5.72s user 0.04s system 99% 
cpu 5.801 total
jacquesmenu@macmini:/Volumes/JMI_Volume/JMI_Developpement/lilypond/release/binaries/lilypond/install/bin
 > time ./convert-ly Fischer_Suite_Sol_M_viola_II.ly > 
Fischer_Suite_Sol_M_viola_II_2.23.12.ly
convert-ly (GNU LilyPond) 2.23.12

convert-ly: Processing `Fischer_Suite_Sol_M_viola_II.ly'... 
Applying conversion: 2.20.0, 2.21.0, 2.21.2, 2.23.1, 2.23.2, 2.23.3, 2.23.4, 
2.23.5, 2.23.6, 2.23.7, 2.23.8, 2.23.9, 2.23.10, 2.23.11, 2.23.12


./convert-ly Fischer_Suite_Sol_M_viola_II.ly >   5.71s user 0.03s system 99% 
cpu 5.748 total


LilyPond-supplied convert-ly. Here, there are variations for some reason:

jacquesmenu@macmini:/Applications/JMI_Applications/LilyPond/lilypond-2.23.12/bin
 > time ./convert-ly Fischer_Suite_Sol_M_viola_II.ly > 
Fischer_Suite_Sol_M_viola_II_2.23.12.ly

convert-ly (GNU LilyPond) 2.23.12

convert-ly: Processing `Fischer_Suite_Sol_M_viola_II.ly'... 
Applying conversion: 2.20.0, 2.21.0, 2.21.2, 2.23.1, 2.23.2, 2.23.3, 2.23.4, 
2.23.5, 2.23.6, 2.23.7, 2.23.8, 2.23.9, 2.23.10, 2.23.11, 2.23.12


./convert-ly Fischer_Suite_Sol_M_viola_II.ly >   7.18s user 0.03s system 99% 
cpu 7.225 total
jacquesmenu@macmini:/Applications/JMI_Applications/LilyPond/lilypond-2.23.12/bin
 > time ./convert-ly Fischer_Suite_Sol_M_viola_II.ly > 
Fischer_Suite_Sol_M_viola_II_2.23.12.ly

convert-ly (GNU LilyPond) 2.23.12

convert-ly: Processing `Fischer_Suite_Sol_M_viola_II.ly'... 
Applying conversion: 2.20.0, 2.21.0, 2.21.2, 2.23.1, 2.23.2, 2.23.3, 2.23.4, 
2.23.5, 2.23.6, 2.23.7, 2.23.8, 2.23.9, 2.23.10, 2.23.11, 2.23.12


./convert-ly Fischer_Suite_Sol_M_viola_II.ly >   7.25s user 0.04s system 99% 
cpu 7.323 total
jacquesmenu@macmini:/Applications/JMI_Applications/LilyPond/lilypond-2.23.12/bin
 > time ./convert-ly Fischer_Suite_Sol_M_viola_II.ly > 
Fischer_Suite_Sol_M_viola_II_2.23.12.ly

convert-ly (GNU LilyPond) 2.23.12

convert-ly: Processing `Fischer_Suite_Sol_M_viola_II.ly'... 
Applying conversion: 2.20.0, 2.21.0, 2.21.2, 2.23.1, 2.23.2, 2.23.3, 2.23.4, 
2.23.5, 2.23.6, 2.23.7, 2.23.8, 2.23.9, 2.23.10, 2.23.11, 2.23.12


./convert-ly Fischer_Suite_Sol_M_viola_II.ly >   15.13s user 0.05s system 99% 
cpu 15.203 total


—

Then apply lilypond to the converted file. ‘… ... ...’ is used for the sake of 
compactness.

Native lilypond:

jacquesmenu@macmini:/Volumes/JMI_Volume/JMI_Developpement/lilypond/release/binaries/lilypond/install/bin
 > time ./lilypond Fischer_Suite_Sol_M_viola_II_2.23.12.ly
GNU LilyPond 2.23.12 (running Guile 2.2)
Processing `Fischer_Suite_Sol_M_viola_II_2.23.12.ly'
… … …
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 54 or 55 pages...
Drawing systems...
Converting to `Fischer_Suite_Sol_M_viola_II_2.23.12.pdf'...
fatal error: failed files: "Fischer_Suite_Sol_M_viola_II_2.23.12.ly"
./lilypond Fischer_Suite_Sol_M_viola_II_2.23.12.ly  7.52s user 0.75s system 98% 
cpu 8.387 total


jacquesmenu@macmini:/Volumes/JMI_Volume/JMI_Developpement/lilypond/release/binaries/lilypond/install/bin
 > time ./lilypond Fischer_Suite_Sol_M_viola_II_2.23.12.ly
GNU LilyPond 2.23.12 (running Guile 2.2)
Processing `Fischer_Suite_Sol_M_viola_II_2.23.12.ly'
… … …
Preprocessing graphical obj

Re: Filtering chunks of the score so to speed up the compiling time

[Paolo Prete]

On Sat, Jun 13, 2020 at 8:39 PM Kieren MacMillan <
kieren_macmil...@sympatico.ca> wrote:

> Hi Paolo,
>
> > In order to speed up the compiling time, I wonder if Lilypond has a
> preprocessor that can filter all the parts of the score which aren't marked
> by a specific tag.
>
> Have you looked into:
>
> 1. \skipTypesetting <
> http://lilypond.org/doc/v2.21/Documentation/notation/skipping-corrected-music
> >
>
>
Hi Kieren,
this is what I was searching for.
However, the doc says that  "it skips all events, including tempo and
instrument changes", but in the proposed example the \tempo instruction is
not skipped.
How can I make it skip *all* events?

Thanks!

Best,
P


>

Re: Filtering chunks of the score so to speed up the compiling time

[Kieren MacMillan]

Hi Paolo,

> In order to speed up the compiling time, I wonder if Lilypond has a 
> preprocessor that can filter all the parts of the score which aren't marked 
> by a specific tag.

Have you looked into:

1. \skipTypesetting 
<http://lilypond.org/doc/v2.21/Documentation/notation/skipping-corrected-music>

2. gridly <https://github.com/openlilylib/gridly>

??

Hope that helps,
Kieren.


Kieren MacMillan, composer (he/him/his)
‣ website: www.kierenmacmillan.info
‣ email: i...@kierenmacmillan.info

Filtering chunks of the score so to speed up the compiling time

[Paolo Prete]

Hello,

In order to speed up the compiling time, I wonder if Lilypond has a
preprocessor that can filter all the parts of the score which aren't marked
by a specific tag.

So, for example, the following code could be compiled only for the chunks
included between \START and \END tags:



\START { \END

% This will be filtered
d' d' d' d'

\START c' c' c' c' \END

% This will be filtered
d' d' d' d'

\START } \END



Is there anything already built-in in Lilypond for this?
Of course it is trivial to make a script for doing the job, but it would be
desirable if Lilypond already had this feature...

Best,
P

Re: Version 2.18 vs.2.19 speed and W10

[Michael Rivers]

Anders Eriksson wrote
> On 2016-09-14 12:07, Phil Holmes wrote:
>> - Original Message - From: "Anders Eriksson" 
>> 

> lilypond@

> 
>> To: "Knut Petersen" 

> Knut_Petersen@

> ; "lilypond-user" 
>> 

> lilypond-user@

> 
>> Sent: Sunday, September 11, 2016 10:55 AM
>> Subject: Re: Version 2.18 vs.2.19 speed and W10
>>
>>
>> I've tried to reproduce this on my Windows 10 machine using Process 
>> Monitor to check file operations.  I get an initial slow run with both 
>> .42 and .43 (I didn't time .42, but .43 was 10.8 seconds).  My second 
>> run with .42 was 0.8 seconds, and 0.4 s with .43.  So I can't help 
>> much further unless I can reproduce the problem.  My Windows version 
>> in 10.0.14393 - what version is the problem appearing on?  Do you have 
>> files in c:\users\yourusername\.lilypond-fonts.cache-2 ?  Do they 
>> update every time LilyPond is run?
>>
> After deleting all the files in the .lilypond-fonts.cache-2 directory it 
> works like normal! Takes a bit longer the first time but not in later
> runs.
> 
> One thing that I noticed was that the latest files where called
> ef9c9ad8cc5857eb63cb3660bc8bd202-i686-mingw32.cache-7.NEW
> 639439a01d569ee739160185e4d7a48e-i686-mingw32.cache-7.NEW
> 
> The were two files called exactly the same minus the .NEW also in the 
> directory.
> 
> 
> 
> 
> Thanks for helping!
> 
> // Anders
> 
> ___
> lilypond-user mailing list

> lilypond-user@

> https://lists.gnu.org/mailman/listinfo/lilypond-user


Same here -- deleting the files in that directory fixed it for me too.



--
View this message in context: 
http://lilypond.1069038.n5.nabble.com/Version-2-18-vs-2-19-speed-and-W10-tp194256p194554.html
Sent from the User mailing list archive at Nabble.com.

___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[Anders Eriksson]

On 2016-09-14 12:07, Phil Holmes wrote:
- Original Message - From: "Anders Eriksson" 
<lilyp...@andis59.se>
To: "Knut Petersen" <knut_peter...@t-online.de>; "lilypond-user" 
<lilypond-user@gnu.org>

Sent: Sunday, September 11, 2016 10:55 AM
Subject: Re: Version 2.18 vs.2.19 speed and W10


I've tried to reproduce this on my Windows 10 machine using Process 
Monitor to check file operations.  I get an initial slow run with both 
.42 and .43 (I didn't time .42, but .43 was 10.8 seconds).  My second 
run with .42 was 0.8 seconds, and 0.4 s with .43.  So I can't help 
much further unless I can reproduce the problem.  My Windows version 
in 10.0.14393 - what version is the problem appearing on?  Do you have 
files in c:\users\yourusername\.lilypond-fonts.cache-2 ?  Do they 
update every time LilyPond is run?


After deleting all the files in the .lilypond-fonts.cache-2 directory it 
works like normal! Takes a bit longer the first time but not in later runs.


One thing that I noticed was that the latest files where called
ef9c9ad8cc5857eb63cb3660bc8bd202-i686-mingw32.cache-7.NEW
639439a01d569ee739160185e4d7a48e-i686-mingw32.cache-7.NEW

The were two files called exactly the same minus the .NEW also in the 
directory.





Thanks for helping!

// Anders

___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[Pierre Perol-Schneider]

Hi Phil,

Thank you for pointing this out!
I've cleared the c:\users\yourusername\.lilypond-fonts.cache-2 file and now
everything seems ok:

\version "2.18.2" { c }
Starting lilypond-windows.exe 2.18.2 [Untitled]...
Processing `c:/users/pierre/appdata/local/temp/frescobaldi-w66wjy/tmprmqi2o/
document.ly'
Parsing...
Interpreting music...
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 1 page...
Drawing systems...
Layout output to `document.ps'...
Converting to `./document.pdf'...
Success: compilation successfully completed
Completed successfully in 1.0".

\version "2.19.47" { c }
Starting lilypond-windows.exe 2.19.47 [Untitled]...
Processing `c:/users/pierre/appdata/local/temp/frescobaldi-w66wjy/tmprmqi2o/
document.ly'
Parsing...
Interpreting music...
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 1 page...
Drawing systems...
Layout output to `./tmp-lilypond-UzOPf4'...
Converting to `document.pdf'...
Deleting `./tmp-lilypond-UzOPf4'...
Success: compilation successfully completed
Completed successfully in 0.8".

Cheers,
Pierre





2016-09-14 12:07 GMT+02:00 Phil Holmes <m...@philholmes.net>:

> - Original Message - From: "Anders Eriksson" <lilyp...@andis59.se>
> To: "Knut Petersen" <knut_peter...@t-online.de>; "lilypond-user" <
> lilypond-user@gnu.org>
> Sent: Sunday, September 11, 2016 10:55 AM
> Subject: Re: Version 2.18 vs.2.19 speed and W10
>
>
>
>>
>> On 2016-09-11 11:38, Knut Petersen wrote:
>>
>>>
>>> 2.19.42.1   4.7
>>>> 2.19.43.1   29.8
>>>>
>>>>
>>> Don't know how to proceed...
>>>>
>>>
>>> Do you use downloaded versions or do you build lilypond from the source
>>> code?
>>>
>> I downloaded from http://download.linuxaudio.org/lilypond/binaries/mingw/
>>
>
>
> I've tried to reproduce this on my Windows 10 machine using Process
> Monitor to check file operations.  I get an initial slow run with both .42
> and .43 (I didn't time .42, but .43 was 10.8 seconds).  My second run with
> .42 was 0.8 seconds, and 0.4 s with .43.  So I can't help much further
> unless I can reproduce the problem.  My Windows version in 10.0.14393 -
> what version is the problem appearing on?  Do you have files in
> c:\users\yourusername\.lilypond-fonts.cache-2 ?  Do they update every
> time LilyPond is run?
>
> --
> Phil Holmes
>
> ___
> lilypond-user mailing list
> lilypond-user@gnu.org
> https://lists.gnu.org/mailman/listinfo/lilypond-user
>
___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[Phil Holmes]

- Original Message - 
From: "Anders Eriksson" <lilyp...@andis59.se>
To: "Knut Petersen" <knut_peter...@t-online.de>; "lilypond-user" 
<lilypond-user@gnu.org>

Sent: Sunday, September 11, 2016 10:55 AM
Subject: Re: Version 2.18 vs.2.19 speed and W10





On 2016-09-11 11:38, Knut Petersen wrote:



2.19.42.1   4.7
2.19.43.1   29.8




Don't know how to proceed...


Do you use downloaded versions or do you build lilypond from the source 
code?

I downloaded from http://download.linuxaudio.org/lilypond/binaries/mingw/



I've tried to reproduce this on my Windows 10 machine using Process Monitor 
to check file operations.  I get an initial slow run with both .42 and .43 
(I didn't time .42, but .43 was 10.8 seconds).  My second run with .42 was 
0.8 seconds, and 0.4 s with .43.  So I can't help much further unless I can 
reproduce the problem.  My Windows version in 10.0.14393 - what version is 
the problem appearing on?  Do you have files in 
c:\users\yourusername\.lilypond-fonts.cache-2 ?  Do they update every time 
LilyPond is run?


--
Phil Holmes 



___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[David Kastrup]

Anders Eriksson <lilyp...@andis59.se> writes:

> On 2016-09-10 22:16, tisimst wrote:
>> Just thought I'd add my two cents. I'm using Windows 8 and I only see a small
>> speed increase between older versions and 2.19.47.
>>
>> Using:
>>
>> \repeat unfold 200 { \tuplet 5/4 { c'4 d' e' f' g' } }
>>
>> and compiling in "publish" mode (i.e., no point-and-click links) I get the
>> following results:
>>
>> 2.18.2 - 6.5 seconds
>> 2.19.36 - 2.7 seconds
>> 2.19.47 - 3.0 seconds
>>
>>
> I have now tested (with the same code) some versions of Lilypond on my
> Windows 10 Anniversary and found this:
> Version Seconds
> 2.18.2.1 6.7
> 2.19.17.1   7.0
> 2.19.27.1   4.0
> 2.19.37.1   4.5
> 2.19.40.1   4.2
> 2.19.42.1   4.7
> 2.19.43.1   29.8
> 2.19.44.1   27.9
> 2.19.46.1   28.2
> 2.19.47.1   28.0
>
> So the problem begins with 2.19.43.

I assume that this is due to changes in GUB or in the font defaults.

-- 
David Kastrup

___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[Anders Eriksson]

On 2016-09-11 11:38, Knut Petersen wrote:



2.19.42.1   4.7
2.19.43.1   29.8




Don't know how to proceed...


Do you use downloaded versions or do you build lilypond from the 
source code?

I downloaded from http://download.linuxaudio.org/lilypond/binaries/mingw/



___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[Anders Eriksson]

On 2016-09-10 22:16, tisimst wrote:

Just thought I'd add my two cents. I'm using Windows 8 and I only see a small
speed increase between older versions and 2.19.47.

Using:

\repeat unfold 200 { \tuplet 5/4 { c'4 d' e' f' g' } }

and compiling in "publish" mode (i.e., no point-and-click links) I get the
following results:

2.18.2 - 6.5 seconds
2.19.36 - 2.7 seconds
2.19.47 - 3.0 seconds


I have now tested (with the same code) some versions of Lilypond on my 
Windows 10 Anniversary and found this:

Version Seconds
2.18.2.1 6.7
2.19.17.1   7.0
2.19.27.1   4.0
2.19.37.1   4.5
2.19.40.1   4.2
2.19.42.1   4.7
2.19.43.1   29.8
2.19.44.1   27.9
2.19.46.1   28.2
2.19.47.1   28.0

So the problem begins with 2.19.43.

Don't know how to proceed...




___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[tisimst]

Just thought I'd add my two cents. I'm using Windows 8 and I only see a small
speed increase between older versions and 2.19.47.

Using:

\repeat unfold 200 { \tuplet 5/4 { c'4 d' e' f' g' } }

and compiling in "publish" mode (i.e., no point-and-click links) I get the
following results:

2.18.2 - 6.5 seconds
2.19.36 - 2.7 seconds
2.19.47 - 3.0 seconds

FWIW,
Abraham



--
View this message in context: 
http://lilypond.1069038.n5.nabble.com/Version-2-18-vs-2-19-speed-and-W10-tp194256p194420.html
Sent from the User mailing list archive at Nabble.com.

___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[Anders Eriksson]

On 2016-09-08 19:03, Michael Rivers wrote:

I'm not sure it's due to a change between v2.19.46 and 47. For most of the
summer, I had been running an older development version and an older version
of Frescobaldi for a while, and they were running fine. I hadn't run
Lilypond in a few weeks, and in the meantime Windows 10 updated to the
"Anniversary edition". When I ran Lilypond again two days ago, it was 20
times slower. Trying to fix that, I updated Lilypond and Frescobaldi to the
newest versions, but that didn't help.

So, while I can't tell exactly what caused the extreme reduction in speed, I
can confirm that Lilypond on the current version of Windows runs
dramatically slower with versions somewhat older than 2.19.47.


I also have noticed the the reduction in speed and one thing I have 
noticed is that every time I compile a .ly file Lilypond will build the 
fonts

Initializing FontConfig...

Adding fontconfig configuration file: 
C:/lilypond-2-19-46-1/LilyPond/usr/share/lilypond/current/fonts/00-lilypond-fonts.conf


Adding fontconfig configuration file: 
C:/lilypond-2-19-46-1/LilyPond/usr/bin/../etc/fonts/fonts.conf


Adding fontconfig configuration file: 
C:/lilypond-2-19-46-1/LilyPond/usr/share/lilypond/current/fonts/99-lilypond-fonts.conf


Adding font directory: 
C:/lilypond-2-19-46-1/LilyPond/usr/share/lilypond/current/fonts/otf


Building font database


and in the 2.19.47.1 version


Initializing FontConfig...

Adding fontconfig configuration file: 
C:/lilypond-2-19-47-1/LilyPond/usr/share/lilypond/current/fonts/00-lilypond-fonts.conf


Adding fontconfig configuration file: 
C:/lilypond-2-19-47-1/LilyPond/usr/bin/../etc/fonts/fonts.conf


Adding fontconfig configuration file: 
C:/lilypond-2-19-47-1/LilyPond/usr/share/lilypond/current/fonts/99-lilypond-fonts.conf


Adding font directory: 
C:/lilypond-2-19-47-1/LilyPond/usr/share/lilypond/current/fonts/otf


Building font database...


I'm running Windows 10 Anniversary Update


___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[Michael Rivers]

I'm not sure it's due to a change between v2.19.46 and 47. For most of the
summer, I had been running an older development version and an older version
of Frescobaldi for a while, and they were running fine. I hadn't run
Lilypond in a few weeks, and in the meantime Windows 10 updated to the
"Anniversary edition". When I ran Lilypond again two days ago, it was 20
times slower. Trying to fix that, I updated Lilypond and Frescobaldi to the
newest versions, but that didn't help.

So, while I can't tell exactly what caused the extreme reduction in speed, I
can confirm that Lilypond on the current version of Windows runs
dramatically slower with versions somewhat older than 2.19.47.



--
View this message in context: 
http://lilypond.1069038.n5.nabble.com/Version-2-18-vs-2-19-speed-and-W10-tp194256p194332.html
Sent from the User mailing list archive at Nabble.com.

___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[Simon Albrecht]

Hi Pierre,

the problem has been reported by Mac users, as a change between v2.19.46 
and .47, in that now it takes so long on every run instead of only the 
very first. Presumably because the font cache is always being rebuilt. 
Maybe it’s the same problem?


Best, Simon

On 06.09.2016 16:59, Pierre Perol-Schneider wrote:

Maybe an simpler compilation:

\version "2.18.2" { c' }

Starting lilypond-windows.exe 2.18.2 [Untitled]...
Processing 
`c:/users/pierre/appdata/local/temp/frescobaldi-zjscau/tmpaf0a6q/document.ly 
'

Parsing...
Interpreting music...
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 1 page...
Drawing systems...
Layout output to `document.ps '...
Converting to `./document.pdf'...
Success: compilation successfully completed
Completed successfully in 1.0".

\version "2.19.47" { c' }

Starting lilypond-windows.exe 2.19.47 [Untitled]...
Processing 
`c:/users/pierre/appdata/local/temp/frescobaldi-zjscau/tmpaf0a6q/document.ly 
'

Parsing...
Interpreting music...
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 1 page...
Drawing systems...
Layout output to `./tmp-lilypond-Jfo5ib'...
Converting to `document.pdf'...
Deleting `./tmp-lilypond-Jfo5ib'...
Success: compilation successfully completed
Completed successfully in 21.2".

Cheers,
Pierre

2016-09-06 16:49 GMT+02:00 Pierre Perol-Schneider 
>:


Hi All,

This morning Win10 forced me to do an update which has last more
than an hour, and then another one few hours later (15mn "only").
Since then, I have a very long run delay with LP v2.19.
For example, the snippet in this thread

:http://lilypond.1069038.n5.nabble.com/Changing-MulitMeasureRest-stencil-to-a-squiggle-td194246.html#a194254



With v2.18.2:
Starting lilypond-windows.exe 2.18.2 [marcMultiRestStencil.ly]...
Processing `C:/Users/Pierre/Documents/LilyPond/0
help/marcMultiRestStencil.ly'
Parsing...
Interpreting music...
Preprocessing graphical objects...
Interpreting music...
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 1 page...
Drawing systems...
Layout output to `marcMultiRestStencil.ps'...
Converting to `./marcMultiRestStencil.pdf'...
Success: compilation successfully completed
Completed successfully in 1.1".

With v2.19.47:
Starting lilypond-windows.exe 2.19.47 [marcMultiRestStencil.ly]...
Processing `C:/Users/Pierre/Documents/LilyPond/0
help/marcMultiRestStencil.ly'
Parsing...
Interpreting music...
Preprocessing graphical objects...
Interpreting music...
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 1 page...
Drawing systems...
Layout output to `./tmp-lilypond-gB6KlR'...
Converting to `marcMultiRestStencil.pdf'...
Deleting `./tmp-lilypond-gB6KlR'...
Success: compilation successfully completed
Completed successfully in 20.6".

so roughly 20 times longer...
Anyone else has the same problem ?

Cheers,
Pierre





___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user



___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[Michael Rivers]

Same problem here. I didn't save any compile times before the update for
comparison, but Lilypond just crawls now. I was compiling a one-page piece
last night, which used to go by in a flash, and now it takes a good 10
seconds.



--
View this message in context: 
http://lilypond.1069038.n5.nabble.com/Version-2-18-vs-2-19-speed-and-W10-tp194256p194257.html
Sent from the User mailing list archive at Nabble.com.

___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Version 2.18 vs.2.19 speed and W10

[Pierre Perol-Schneider]

Maybe an simpler compilation:

\version "2.18.2" { c' }

Starting lilypond-windows.exe 2.18.2 [Untitled]...
Processing `c:/users/pierre/appdata/local/temp/frescobaldi-zjscau/tmpaf0a6q/
document.ly'
Parsing...
Interpreting music...
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 1 page...
Drawing systems...
Layout output to `document.ps'...
Converting to `./document.pdf'...
Success: compilation successfully completed
Completed successfully in 1.0".

\version "2.19.47" { c' }

Starting lilypond-windows.exe 2.19.47 [Untitled]...
Processing `c:/users/pierre/appdata/local/temp/frescobaldi-zjscau/tmpaf0a6q/
document.ly'
Parsing...
Interpreting music...
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 1 page...
Drawing systems...
Layout output to `./tmp-lilypond-Jfo5ib'...
Converting to `document.pdf'...
Deleting `./tmp-lilypond-Jfo5ib'...
Success: compilation successfully completed
Completed successfully in 21.2".

Cheers,
Pierre

2016-09-06 16:49 GMT+02:00 Pierre Perol-Schneider <
pierre.schneider.pa...@gmail.com>:

> Hi All,
>
> This morning Win10 forced me to do an update which has last more than an
> hour, and then another one few hours later (15mn "only").
> Since then, I have a very long run delay with LP v2.19.
> For example, the snippet in this thread :http://lilypond.1069038.n5.
> nabble.com/Changing-MulitMeasureRest-stencil-to-a-
> squiggle-td194246.html#a194254
>
> With v2.18.2:
> Starting lilypond-windows.exe 2.18.2 [marcMultiRestStencil.ly]...
> Processing `C:/Users/Pierre/Documents/LilyPond/0
> help/marcMultiRestStencil.ly'
> Parsing...
> Interpreting music...
> Preprocessing graphical objects...
> Interpreting music...
> Preprocessing graphical objects...
> Finding the ideal number of pages...
> Fitting music on 1 page...
> Drawing systems...
> Layout output to `marcMultiRestStencil.ps'...
> Converting to `./marcMultiRestStencil.pdf'...
> Success: compilation successfully completed
> Completed successfully in 1.1".
>
> With v2.19.47:
> Starting lilypond-windows.exe 2.19.47 [marcMultiRestStencil.ly]...
> Processing `C:/Users/Pierre/Documents/LilyPond/0
> help/marcMultiRestStencil.ly'
> Parsing...
> Interpreting music...
> Preprocessing graphical objects...
> Interpreting music...
> Preprocessing graphical objects...
> Finding the ideal number of pages...
> Fitting music on 1 page...
> Drawing systems...
> Layout output to `./tmp-lilypond-gB6KlR'...
> Converting to `marcMultiRestStencil.pdf'...
> Deleting `./tmp-lilypond-gB6KlR'...
> Success: compilation successfully completed
> Completed successfully in 20.6".
>
> so roughly 20 times longer...
> Anyone else has the same problem ?
>
> Cheers,
> Pierre
>
>
>
___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Version 2.18 vs.2.19 speed and W10

[Pierre Perol-Schneider]

Hi All,

This morning Win10 forced me to do an update which has last more than an
hour, and then another one few hours later (15mn "only").
Since then, I have a very long run delay with LP v2.19.
For example, the snippet in this thread :
http://lilypond.1069038.n5.nabble.com/Changing-MulitMeasureRest-stencil-to-a-squiggle-td194246.html#a194254

With v2.18.2:
Starting lilypond-windows.exe 2.18.2 [marcMultiRestStencil.ly]...
Processing `C:/Users/Pierre/Documents/LilyPond/0
help/marcMultiRestStencil.ly'
Parsing...
Interpreting music...
Preprocessing graphical objects...
Interpreting music...
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 1 page...
Drawing systems...
Layout output to `marcMultiRestStencil.ps'...
Converting to `./marcMultiRestStencil.pdf'...
Success: compilation successfully completed
Completed successfully in 1.1".

With v2.19.47:
Starting lilypond-windows.exe 2.19.47 [marcMultiRestStencil.ly]...
Processing `C:/Users/Pierre/Documents/LilyPond/0
help/marcMultiRestStencil.ly'
Parsing...
Interpreting music...
Preprocessing graphical objects...
Interpreting music...
Preprocessing graphical objects...
Finding the ideal number of pages...
Fitting music on 1 page...
Drawing systems...
Layout output to `./tmp-lilypond-gB6KlR'...
Converting to `marcMultiRestStencil.pdf'...
Deleting `./tmp-lilypond-gB6KlR'...
Success: compilation successfully completed
Completed successfully in 20.6".

so roughly 20 times longer...
Anyone else has the same problem ?

Cheers,
Pierre
___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[Knut Petersen]

On 07.02.2013 13:52, Carlo Stemberger wrote:


$ time make

[...]

real0m14.519s
user0m14.276s
sys0m0.192s


i7-3770K (3rd generation), Debian Wheezy.



cpu/men: Pentium-M Dothan, 1.86 GHz, 2GB
mobo: AOpen i915GMm-HFS
os: openSuSE 12.3, kernel 3.9
lilypond: 16.2

time make
[...]
real1m5.877s
user1m0.884s
sys 0m1.004s

About 4 times slower than your Ivy Bridge system.
Half of that can be accounted to the doubled clock rate.
I suspect that a lot of the rest is caused by general
improvements in cpu architecture. So lilypond seems
not to scale well with the number of available cores.

cu,
 Knut

___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[Phil Holmes]

- Original Message - 
From: Knut Petersen knut_peter...@t-online.de

To: Carlo Stemberger carlo.stember...@gmail.com
Cc: lilypond-user@gnu.org
Sent: Tuesday, April 30, 2013 10:46 AM
Subject: Re: speed



On 07.02.2013 13:52, Carlo Stemberger wrote:


$ time make

[...]

real0m14.519s
user0m14.276s
sys0m0.192s


i7-3770K (3rd generation), Debian Wheezy.



cpu/men: Pentium-M Dothan, 1.86 GHz, 2GB
mobo: AOpen i915GMm-HFS
os: openSuSE 12.3, kernel 3.9
lilypond: 16.2

time make
[...]
real1m5.877s
user1m0.884s
sys 0m1.004s

About 4 times slower than your Ivy Bridge system.
Half of that can be accounted to the doubled clock rate.
I suspect that a lot of the rest is caused by general
improvements in cpu architecture. So lilypond seems
not to scale well with the number of available cores.


Generally, it does.  Read 
http://lilypond.org/doc/v2.17/Documentation/contributor/saving-time-with-the-_002dj-option


--
Phil Holmes 



___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[David Kastrup]

Phil Holmes m...@philholmes.net writes:

 From: Knut Petersen knut_peter...@t-online.de

 About 4 times slower than your Ivy Bridge system.
 Half of that can be accounted to the doubled clock rate.
 I suspect that a lot of the rest is caused by general
 improvements in cpu architecture. So lilypond seems
 not to scale well with the number of available cores.

 Generally, it does.  Read
 http://lilypond.org/doc/v2.17/Documentation/contributor/saving-time-with-the-_002dj-option

Not really.  That just helps when doing _multiple_ files at once.  If
you use -j4, LilyPond initializes itself, splits the file list into 4
parts, then splits itself into 5 copies.  4 of those copies process
their file list independently, and the fifth copy reports when all them
have finished.

You save the time of initialization for three of those jobs.  But that's
a rather small amount.  Other than that, there is no advantage to
splitting the job list yourself and calling parallel copies of LilyPond.

It would be somewhat better if the fifth copy of LilyPond performed the
actual task of a job server, telling each running copy what file to
process next and thus keeping all of them active until the end.  In
practice, I doubt it would make much of a difference.  You can try
experimenting with GNU parallel if you want to try this.

-- 
David Kastrup


___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[Phil Holmes]

- Original Message - 
From: David Kastrup d...@gnu.org

To: lilypond-user@gnu.org
Sent: Tuesday, April 30, 2013 2:20 PM
Subject: Re: speed



Phil Holmes m...@philholmes.net writes:


From: Knut Petersen knut_peter...@t-online.de



About 4 times slower than your Ivy Bridge system.
Half of that can be accounted to the doubled clock rate.
I suspect that a lot of the rest is caused by general
improvements in cpu architecture. So lilypond seems
not to scale well with the number of available cores.


Generally, it does.  Read
http://lilypond.org/doc/v2.17/Documentation/contributor/saving-time-with-the-_002dj-option


Not really.  That just helps when doing _multiple_ files at once.  If
you use -j4, LilyPond initializes itself, splits the file list into 4
parts, then splits itself into 5 copies.  4 of those copies process
their file list independently, and the fifth copy reports when all them
have finished.

You save the time of initialization for three of those jobs.  But that's
a rather small amount.  Other than that, there is no advantage to
splitting the job list yourself and calling parallel copies of LilyPond.

It would be somewhat better if the fifth copy of LilyPond performed the
actual task of a job server, telling each running copy what file to
process next and thus keeping all of them active until the end.  In
practice, I doubt it would make much of a difference.  You can try
experimenting with GNU parallel if you want to try this.

--
David Kastrup



From the snipped posting, I was assuming this was (an amazingly fast) 
compilation of lilypond itself.  From your post, I assume it's using make to 
have lilypond compile some music.


--
Phil Holmes 



___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[David Kastrup]

Phil Holmes m...@philholmes.net writes:

 From the snipped posting, I was assuming this was (an amazingly fast)
 compilation of lilypond itself.  From your post, I assume it's using
 make to have lilypond compile some music.

To quote the original posting:

On my new notebook the lilypond version of Reubke's Psalm94 now
compiles in 20 seconds (including creation of zip package) instead of
100. (3rd generation i5 processor)

-- 
David Kastrup


___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[Janek Warchoł]

Hi,

On Thu, Feb 7, 2013 at 1:22 PM, Andrew Bernard andrew.bern...@gmail.com wrote:
 Greetings,

 On my Linux Mint 14 virtual machine running in Virtualbox on a Macbook Pro
 with a 2.4 GHz i7, I get 30 seconds more or less exactly.

 Looks like those i5's are catching up.

 The Reubke seems like quite a good timing benchmark. And what a wonderful
 piece of engraving.

You mean the LilyPond version?  I'm sorry, but last time i checked it
was full of engraving problems, most notably ugly ties.  LilyPond ties
need serious fixing.

best,
Janek

___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

speed

[Martin Tarenskeen]

Hi,

It was time to buy myself a newer notebook. My Dell D600 is beginning 
to fall apart.


On my new notebook the lilypond version of Reubke's Psalm94 now compiles 
in 20 seconds (including creation of zip package) instead of 100. (3rd 
generation i5 processor)


:-)

--

MT

___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[Andrew Bernard]

Greetings,

On my Linux Mint 14 virtual machine running in Virtualbox on a Macbook 
Pro with a 2.4 GHz i7, I get 30 seconds more or less exactly.


Looks like those i5's are catching up.

The Reubke seems like quite a good timing benchmark. And what a 
wonderful piece of engraving.


cheerio!
Andrew


On 7/02/13 9:25 PM, Martin Tarenskeen wrote:
On my new notebook the lilypond version of Reubke's Psalm94 now 
compiles in 20 seconds (including creation of zip package) instead of 
100. (3rd generation i5 processor)



___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[Martin Tarenskeen]

On Thu, 7 Feb 2013, Andrew Bernard wrote:


Greetings,

On my Linux Mint 14 virtual machine running in Virtualbox on a Macbook Pro 
with a 2.4 GHz i7, I get 30 seconds more or less exactly.


Looks like those i5's are catching up.


I guess VirtualBox is also slowing things down a bit? What if you run the 
native Mac OSX version ?


The Reubke seems like quite a good timing benchmark. And what a wonderful 
piece of engraving.


Indeed!

--

MT

___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[Carlo Stemberger]

Il 07/02/2013 11:25, Martin Tarenskeen ha scritto:


On my new notebook the lilypond version of Reubke's Psalm94 now 
compiles in 20 seconds (including creation of zip package) instead of 
100. (3rd generation i5 processor)




$ time make

[...]

real0m14.519s
user0m14.276s
sys0m0.192s


i7-3770K (3rd generation), Debian Wheezy.

Ciao!

Carlo

--
  .-.  | Registered Linux User #443882|  .''`.
  oo|  | http://linuxcounter.net/ | : :'  :
 /`'\  |   Registered Debian User #9  | `. `'`
(\_;/) | http://debiancounter.altervista.org/ |   `-


___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[David Kastrup]

Carlo Stemberger carlo.stember...@gmail.com writes:

 Il 07/02/2013 11:25, Martin Tarenskeen ha scritto:

 On my new notebook the lilypond version of Reubke's Psalm94 now
 compiles in 20 seconds (including creation of zip package) instead
 of 100. (3rd generation i5 processor)


 $ time make

 [...]

 real0m14.519s
 user0m14.276s
 sys0m0.192s

I am assuming you are not talking about building LilyPond here.

-- 
David Kastrup


___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[Carlo Stemberger]

Il 07/02/2013 13:58, David Kastrup ha scritto:
I am assuming you are not talking about building LilyPond here. 


:D

No, I'm talking about compiling this files:

http://imslp.org/wiki/Special:ImagefromIndex/177628

Ciao!

Carlo

--
  .-.  | Registered Linux User #443882|  .''`.
  oo|  | http://linuxcounter.net/ | : :'  :
 /`'\  |   Registered Debian User #9  | `. `'`
(\_;/) | http://debiancounter.altervista.org/ |   `-


___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[Urs Liska]

Am 07.02.2013 13:58, schrieb David Kastrup:

Carlo Stemberger carlo.stember...@gmail.com writes:


Il 07/02/2013 11:25, Martin Tarenskeen ha scritto:

On my new notebook the lilypond version of Reubke's Psalm94 now
compiles in 20 seconds (including creation of zip package) instead
of 100. (3rd generation i5 processor)


$ time make

[...]

real0m14.519s
user0m14.276s
sys0m0.192s

I am assuming you are not talking about building LilyPond here.


What would you do if he did? ;-/

___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

Re: speed

[David Kastrup]

Urs Liska li...@ursliska.de writes:

 Am 07.02.2013 13:58, schrieb David Kastrup:
 Carlo Stemberger carlo.stember...@gmail.com writes:

 Il 07/02/2013 11:25, Martin Tarenskeen ha scritto:
 On my new notebook the lilypond version of Reubke's Psalm94 now
 compiles in 20 seconds (including creation of zip package) instead
 of 100. (3rd generation i5 processor)

 $ time make

 [...]

 real0m14.519s
 user0m14.276s
 sys0m0.192s
 I am assuming you are not talking about building LilyPond here.

 What would you do if he did? ;-/

Volunteer him for release work.

-- 
David Kastrup


___
lilypond-user mailing list
lilypond-user@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-user

lilypond speed

[Martin Tarenskeen]

Some day soon my still reliable but not very fast 10 year old laptop will 
have to be replaced by a more modern machine.


Will Lilypond benifit much if my next computer will have one of those 
modern multi-core processors like the Intel i3/5/7 ?


I'm just curious.

--

Martin

___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: lilypond speed

[flup2]

On a Mac, speed increase from a Core2Duo to a Corei5 is really noticeable,
about 40% faster. I guess this is about the same on Windows or Linux. I
think you'll see a huge difference. :)

Philippe
-- 
View this message in context: 
http://old.nabble.com/lilypond-speed-tp30929720p30929766.html
Sent from the Gnu - Lilypond - User mailing list archive at Nabble.com.


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: lilypond speed

[Bernardo Barros]

I was thinkig about that too. Parallelization in Lilypond can be
possible. Imagine rendering one page in each core in parallel, or one
system for each core. Also other kinds of optimizations for 'preview'
modes, where you need more speed then optimal quality?  Maybe improved
performance can be a feature request for Lilypond 2.16 ?

___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: lilypond speed

[Phil Holmes]

- Original Message - 
From: Martin Tarenskeen m.tarensk...@zonnet.nl

To: lilypond-user mailinglist lilypond-user@gnu.org
Sent: Tuesday, February 15, 2011 10:42 AM
Subject: lilypond speed




Some day soon my still reliable but not very fast 10 year old laptop will 
have to be replaced by a more modern machine.


Will Lilypond benifit much if my next computer will have one of those 
modern multi-core processors like the Intel i3/5/7 ?


I'm just curious.

--

Martin


If you're using Windows, you definitely won't see an improvement with a 
multi-core machine.  LilyPond is single-threaded, and so can only use a 
single core at a given time.  It may be that your new machine will have more 
memory, and this is definitely beneficial for large scores, where Lily can 
take a lot of memory.  Whether the processor will be faster depends solely 
on the processing power of _one_ of the cores of your new CPU versus the 
power of your old one.



--
Phil Holmes



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Speed tips, again, for extremely large scores?

[Martin Tarenskeen]

On Mon, 1 Feb 2010, Michael Kappler wrote:

I'm also still very interested if there are possibilities to increase 
LilyPond performance further.
My machine is very slow, though and I cannot speak for many people when 
raising performance issues.


Would it be an idea to create a Lilypond Benchmark webpage, small, 
interesting, and useless ;-) This webpage would show a list with the following 
info:

1. Hardware / Processor type
2. Platform / OS
3. Lilypond version 4. The benchmark result: I suggest to give te time needed 
to make the (at least in this mailinglist) already famous Reubke Psalm 94 
score.


The website should have a simple interface where Lilypond users can enter their 
results and show a list sorted by results. On linux it is easy to measure the 
time using the time command.


Would be interesting to see results with relatively old computers as well as 
the newest i7 speeddevils ...


Lilypond does benifit a lot from this fast new processor technology. Speed is 
not Lilypond's strongest point. Other, less powerful but still quite good, 
text-to-musicscore commandline programs like mup (shareware), abcm2ps, and pmw 
are MUCH faster. Even on my ancient Atari computer running at 32 MHz, I could 
run these programs with reasonable speed. But there are several other reasons 
why I prefer and use mostly Lilypond now.


If one day computers will be too fast to give reliable results, we can always 
change the benchmark test by repeating the same test 10 or 100 times :-)


There have been some individual postings on this subject, but a small dedicated 
website would be nice. An extra page on the normal Lilypond homepage maybe ?


I'm not a very capable webpage designer. Any volunteers ?

--

Martin


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Speed tips, again, for extremely large scores?

[Mats Bengtsson]

Martin Tarenskeen wrote:


On Mon, 1 Feb 2010, Michael Kappler wrote:

I'm also still very interested if there are possibilities to increase 
LilyPond performance further.
My machine is very slow, though and I cannot speak for many people 
when raising performance issues.


Would it be an idea to create a Lilypond Benchmark webpage, small, 
interesting, and useless ;-) This webpage would show a list with the 
following info:

1. Hardware / Processor type
2. Platform / OS
3. Lilypond version 4. The benchmark result: I suggest to give te time 
needed to make the (at least in this mailinglist) already famous 
Reubke Psalm 94 score.
You can find some examples of such benchmarking in the mailing list 
archives. Don't forget that RAM size is a major factor when it comes to 
processing time.


  /Mats



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Speed tips, again, for extremely large scores?

[Trevor Daniels]

Mats Bengtsson wrote Monday, February 01, 2010 11:52 AM


Martin Tarenskeen wrote:


On Mon, 1 Feb 2010, Michael Kappler wrote:

I'm also still very interested if there are possibilities to 
increase LilyPond performance further.
My machine is very slow, though and I cannot speak for many 
people when raising performance issues.


Would it be an idea to create a Lilypond Benchmark webpage, 
small, interesting, and useless ;-) This webpage would show a 
list with the following info:

1. Hardware / Processor type
2. Platform / OS
3. Lilypond version 4. The benchmark result: I suggest to give te 
time needed to make the (at least in this mailinglist) already 
famous Reubke Psalm 94 score.


It would be both interesting and a useful check on
whether code additions to new releases have had an
effect on processing speed.  Although for this we
would have to establish one or maybe several standard
configurations so the tests are directly comparable.

Don't forget that RAM size is a major factor when it comes to 
processing time.


Probably the major factor along with cpu speed,
unless you already have enough RAM.

There are several other parameters which will affect
the speed but which are harder to determine.  These
vary depending on the particular computer model, so the
results might still be very variable.  These include

 Processor cache size and speed
 Memory transfer rates
 Hard disk seek/search times
   (affects page rates if RAM is insufficient)
 Whether page file is fragmented (ditto)

Also the effect of other processes would need to be
excluded by doing the test in a newly-booted clean
system, and running the test several times, retaining
just the fastest time.

Trevor




___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Speed tips, again, for extremely large scores?

[Graham Percival]

On Mon, Feb 01, 2010 at 12:37:42PM -, Trevor Daniels wrote:

 Mats Bengtsson wrote Monday, February 01, 2010 11:52 AM

 Martin Tarenskeen wrote:

 On Mon, 1 Feb 2010, Michael Kappler wrote:

 I'm also still very interested if there are possibilities to  
 increase LilyPond performance further.
 My machine is very slow, though and I cannot speak for many people 
 when raising performance issues.

 Would it be an idea to create a Lilypond Benchmark webpage, small, 
 interesting, and useless ;-)

 It would be both interesting and a useful check on
 whether code additions to new releases have had an
 effect on processing speed.  Although for this we
 would have to establish one or maybe several standard
 configurations so the tests are directly comparable.

ARGH!  Bloody mao!  (not directed at you, Trevor)

Like **so many** things in lilypond, this was done ages ago.  For
the past few YEARS, the regtests have included this information.

The problem is that it isn't documented anywhere, the output is
hard to read, and I have a sneaky suspicion that part of it is
broken at the moment but would only require a 10-minute bugfix if
the person knew what he was doing.


You want to improve the situation?  There's two options:

1) start reading the sources to figure out how this benchmarking
works.  Figure out what the output means, whether or not it's
currently working, look at old versions to see what it looked like
back then, etc.

2) start helping me with easy stuff (like writing plain text for
the new website), so that *I* can work on #1.  I have the
technical knowledge and persistance required to solve #1, but
given the lack of other people on more critical things, I doubt
I'll start working on it for at least 3 months.

- Graham


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Tim Reeves]

Frank wrote:
Am Donnerstag, 3. September 2009 schrieb Tim Reeves:

 Mainly for my own curiosity, I compiled the Reubke Sonata score to 
check
 timing:
 WinXP SP3 32-bit, LP 2.13.3, LPT 2.12.869, on Intel C2D E9600 (2.8GHz), 
2
 GB RAM

 5 min 38 seconds.

 A bit slower than the Linux times others got.

W00t, I got only
real5m47.699s
user5m32.306s
sys 0m11.697s
on my linux system (C2D @ 2 GHz), but I'm still on 2.12.1, which gave me 
some 
error messages, though the PDF was created. Perhaps 2.13 is a little 
faster(?)

-


Frank,

I forgot to mention that I also got quite a few warnings [not errors] on 
2.13.3 - I think due to a missing font - but like you still got the output 
file.
Did you also have 2GB of RAM? - I understand that the amount of physical 
memory available makes a big difference in compile times on a large LP 
file.
Perhaps if I ran Linux on my machine it would really scream, but I have to 
have Windows XP for work. At least its not Vista!

Tim


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Valentin Villenave]

On Fri, Sep 4, 2009 at 2:13 AM, Frank Steinmetzgerwar...@gmx.de wrote:
 W00t, I got only
 real    5m47.699s
 user    5m32.306s
 sys     0m11.697s
 on my linux system (C2D @ 2 GHz), but I'm still on 2.12.1, which gave me some
 error messages, though the PDF was created. Perhaps 2.13 is a little
 faster(?)

Actually, I stumbled upon something very odd: though I haven't the
exact numbers, with 2.12 my opera used to compile in ~40 minutes on
Win32, ~25 minutes on Linux64 -- but now that I have upgraded to the
latest git sources (that include Joe's recent work on vertical
spacing), it takes more than... 90 minutes!!!
(the PDF looks nicer though).

I'll investigate this problem a bit more, as I can hardly believe it myself.

Regards,
Valentin


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Graham Percival]

On Fri, Sep 04, 2009 at 10:52:13PM +0200, Valentin Villenave wrote:
 Actually, I stumbled upon something very odd: though I haven't the
 exact numbers, with 2.12 my opera used to compile in ~40 minutes on
 Win32, ~25 minutes on Linux64 -- but now that I have upgraded to the
 latest git sources (that include Joe's recent work on vertical
 spacing), it takes more than... 90 minutes!!!
 (the PDF looks nicer though).

How is that odd?  More complicated algorithms take more time.  I
haven't followed the details of the spacing changes, but I'd
certainly expect them to take longer.

Cheers,
- Graham


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Valentin Villenave]

On Fri, Sep 4, 2009 at 11:05 PM, Graham
Percivalgra...@percival-music.ca wrote:
 How is that odd?  More complicated algorithms take more time.  I
 haven't followed the details of the spacing changes, but I'd
 certainly expect them to take longer.

I do too, but -- let me do the math -- a _360%_ increase, really? :-)

Regards,
Valentin


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Joe Neeman]

On Fri, 2009-09-04 at 22:05 +0100, Graham Percival wrote:
 On Fri, Sep 04, 2009 at 10:52:13PM +0200, Valentin Villenave wrote:
  Actually, I stumbled upon something very odd: though I haven't the
  exact numbers, with 2.12 my opera used to compile in ~40 minutes on
  Win32, ~25 minutes on Linux64 -- but now that I have upgraded to the
  latest git sources (that include Joe's recent work on vertical
  spacing), it takes more than... 90 minutes!!!
  (the PDF looks nicer though).
 
 How is that odd?  More complicated algorithms take more time.  I
 haven't followed the details of the spacing changes, but I'd
 certainly expect them to take longer.

The algorithms shouldn't really be more complicated, just differently
organized. If you can figure out which commit caused the problem, that
would be helpful.

Cheers,
Joe




___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

RE: Lilypond Speed

[Nick Payne]

 -Original Message-
 From: lilypond-user-bounces+nick.payne=internode.on@gnu.org
 [mailto:lilypond-user-bounces+nick.payne=internode.on@gnu.org] On
 Behalf Of Nick Payne
 Sent: Wednesday, 2 September 2009 7:02 PM
 Cc: 'lilypond'
 Subject: RE: Lilypond Speed
 
  -Original Message-
  From: lilypond-user-bounces+nick.payne=internode.on@gnu.org
  [mailto:lilypond-user-bounces+nick.payne=internode.on@gnu.org] On
  Behalf Of Michael David Crawford
  Sent: Wednesday, 2 September 2009 1:39 AM
  Cc: lilypond
  Subject: Re: Lilypond Speed
 
  It would be interesting to see what the performance is on different
  operating systems, but running on the exact same hardware.
 
 Ask and it shall be done. I took the opportunity while testing builds
 of Win2008 x86 and x64 to install LP and run the same test. I found the
 results rather surprising. All tests (building Reubke's sonata on the
 94th psalm using LP 2.13.3) conducted on the same machine:
 
 Debian 4 amd64: 4min 4sec
 
 Win2008 SP2 x64: 5min 58sec
 
 Win2008 SP2 x86: 6m 3sec
 
 That's quite a substantial performance hit for the Windows version
 compared to Linux. I don't have a server build of x86 Linux but if I
 get the chance tomorrow I'll see if Ubuntu 9.04 x86 has the Perc
 drivers needed to install on the server and run the test with that.
 
More timings (build of same score, same machine):

Ubuntu 9.04 x86: 5min 25sec

Ubuntu 9.04 amd64: 4min 1sec

So it looks as though, for any sort of substantial score, turnaround time will 
be markedly reduced by using 64-bit Linux.

Nick



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Bertalan Fodor (LilyPondTool)]

Please don't start this discussion :)

Pierre Couderc wrote:

Mmm,
It may be too that linux applications are by nature of the OS quicker 
than Windows ones...


Tim McNamara a écrit :


On Sep 3, 2009, at 7:04 AM, Nick Payne wrote:

So it looks as though, for any sort of substantial score, turnaround 
time will be markedly reduced by using 64-bit Linux.


Thanks for all the comparisons, Nick, that was very interesting.  If 
I understand the history of LilyPond correctly, it is a Linux 
application ported to Windows and OS X.  So it's not surprising that 
it's faster on the various Linux OSes.  Maybe someday I'll get around 
to buying an Intel Mac and can install all three OSes on their own 
partitions and do a comparison that way.


I use LilyPond on a 1.33 gHz PPC iBook with 1 GB RAM, called from 
Carbon Emacs (GNU Emacs 22.2.1 (powerpc-apple-darwin8.11.0, Carbon 
Version 1.6.0)) and while LP is no speed demon, there are a *lot* of 
calculations and processes happening and I don't expect a score to be 
rendered in 3 seconds.



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user




___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user






___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Frank Steinmetzger]

Am Donnerstag, 3. September 2009 schrieb Bertalan Fodor (LilyPondTool):
 Please don't start this discussion :)

Then how about a discussion about top posting? ;-) ;-) *duckandhide*
-- 
Gruß | Greetings | Qapla'
What do you call a dead bee? - A was.


signature.asc
Description: This is a digitally signed message part.
___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Tim Reeves]

Mainly for my own curiosity, I compiled the Reubke Sonata score to check 
timing:
WinXP SP3 32-bit, LP 2.13.3, LPT 2.12.869, on Intel C2D E9600 (2.8GHz), 2 
GB RAM

5 min 38 seconds.

A bit slower than the Linux times others got.

I do have a Vista machine at home (wife's PC) I could check it on if 
someone is interested, but I'd have to update the LP to make it 
meaningful.

Tim Reeves


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Frank Steinmetzger]

Am Donnerstag, 3. September 2009 schrieb Tim Reeves:

 Mainly for my own curiosity, I compiled the Reubke Sonata score to check
 timing:
 WinXP SP3 32-bit, LP 2.13.3, LPT 2.12.869, on Intel C2D E9600 (2.8GHz), 2
 GB RAM

 5 min 38 seconds.

 A bit slower than the Linux times others got.

W00t, I got only
real5m47.699s
user5m32.306s
sys 0m11.697s
on my linux system (C2D @ 2 GHz), but I'm still on 2.12.1, which gave me some 
error messages, though the PDF was created. Perhaps 2.13 is a little 
faster(?)
-- 
Gruß | Greetings | Qapla'
The first time you’ll get a Microsoft product that doesn’t suck
will be the day they start producing vacuum cleaners.


signature.asc
Description: This is a digitally signed message part.
___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Jonathan Wilkes]

--- On Sat, 8/29/09, Bertalan Fodor (LilyPondTool) lilypondt...@organum.hu 
wrote:

 From: Bertalan Fodor (LilyPondTool) lilypondt...@organum.hu
 Subject: Re: Lilypond Speed
 To: Jonathan Wilkes jancs...@yahoo.com
 Cc: lilypond-user@gnu.org
 Date: Saturday, August 29, 2009, 6:40 AM
 Jonathan Wilkes wrote:
  It sounds like there is a wide discrepancy depending
 on machine/os/etc. 
  Does anyone have any insight into how I could decrease
 this time on my winxp machine?
  
  I feel like if I could get it down to something close
 to one second, it would be a lot easier to learn Lilypond.
  
  Also: is there a way using LilypondTool to set
 Lilypond to run every time I enter a barcheck?
    
 Sure, as you can set a shortcut for running LilyPond, see
 UtilitiesGlobal optionsShortcuts, you can set the
 shortcut to be |
 
 Bert

Hi Bert,
 It seems that when I set | to be a shortcut for Run Lilypond, it 
no longer prints the character; it only runs the shortcut.  Do you know 
if there's a workaround for this?

Thanks,
Jonathan





___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

RE: Lilypond Speed

[Nick Payne]

 -Original Message-
 From: lilypond-user-bounces+nick.payne=internode.on@gnu.org
 [mailto:lilypond-user-bounces+nick.payne=internode.on@gnu.org] On
 Behalf Of Michael David Crawford
 Sent: Wednesday, 2 September 2009 1:39 AM
 Cc: lilypond
 Subject: Re: Lilypond Speed
 
 It would be interesting to see what the performance is on different
 operating systems, but running on the exact same hardware.
 
Ask and it shall be done. I took the opportunity while testing builds of 
Win2008 x86 and x64 to install LP and run the same test. I found the results 
rather surprising. All tests (building Reubke's sonata on the 94th psalm using 
LP 2.13.3) conducted on the same machine:

Debian 4 amd64: 4min 4sec

Win2008 SP2 x64: 5min 58sec

Win2008 SP2 x86: 6m 3sec

That's quite a substantial performance hit for the Windows version compared to 
Linux. I don't have a server build of x86 Linux but if I get the chance 
tomorrow I'll see if Ubuntu 9.04 x86 has the Perc drivers needed to install on 
the server and run the test with that.

Nick



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Michael David Crawford]

Peter Chubb wrote:

Han-Wen More importantly: LilyPond is single-threaded, so the number
Han-Wen of cores is irrelevant.


While LilyPond may be single threaded, in general the underlying 
operating system is multithreaded.  It might be the case that a system 
call LilyPond depends on can get executed in a multithreaded way.


--
Michael David Crawford
mich...@geometricvisions.com
http://www.geometricvisions.com/ -- Creative Commons Lilypond Scores


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Han-Wen Nienhuys]

On Tue, Sep 1, 2009 at 10:52 AM, Michael David
Crawfordmich...@geometricvisions.com wrote:


 Peter Chubb wrote:

 Han-Wen More importantly: LilyPond is single-threaded, so the number
 Han-Wen of cores is irrelevant.

 While LilyPond may be single threaded, in general the underlying operating
 system is multithreaded.  It might be the case that a system call LilyPond
 depends on can get executed in a multithreaded way.

LilyPond almost does not interact with the OS except for reading and
writing a couple of files. It's CPU bound.

-- 
Han-Wen Nienhuys - han...@xs4all.nl - http://www.xs4all.nl/~hanwen


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Michael David Crawford]

Han-Wen Nienhuys wrote:

While LilyPond may be single threaded, in general the underlying operating
system is multithreaded.  It might be the case that a system call LilyPond
depends on can get executed in a multithreaded way.


LilyPond almost does not interact with the OS except for reading and
writing a couple of files. It's CPU bound.


That may be true, but it also places a load on the system in ways that 
aren't obvious.  For example the virtual memory system, paging in 
executable code, mapping shared libraries.


It was mentioned that LilyPond uses a lot of stack.  The stack is 
extended automatically on most VM systems - when you try to access a 
region of stack memory that's not mapped, there is a page fault, and the 
page fault handler creates the necessary mapping.


It would be interesting to see what the performance is on different 
operating systems, but running on the exact same hardware.


--
Michael David Crawford
mich...@geometricvisions.com
http://www.geometricvisions.com/ -- Creative Commons LilyPond Scores


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Nick Payne]

As I have just had a rather powerful evaluation server to play around with
for a few days while I tested our various Windows and Linux server builds on
it, I thought I'd also take the opportunity to compare the build speed of a
reasonably substantial score. I used Reinhold's setting of Reubke's sonata
on the 94th psalm. I tested on three machines, all running the same version
of Lilypond:

1. Dell GX620 workstation, Pentium D dual-core 3.0GHz CPU, 1Gb RAM, Ubuntu
9.04 x86: 10min 11sec
2. Dell GX745 workstation, Pentium D dual-core 3.4GHz CPU, 2Gb RAM, WinXP
SP3: 9min 22sec
3. PowerEdge R710 server, dual quad-core Xeon 5560 2.8GHz CPUs, 24Gb RAM,
Debian 5 amd64: 4min 4sec

Number of CPUs seemed irrelevant as only a single CPU was getting flogged on
each machine while the build was in progress. I saw pretty much the same
percentage difference in build time on shorter scores as well - eg a four
page score built in 16sec on the GX620 workstation and 8sec on the server.

Nick



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Peter Chubb]

 Nick == Nick Payne njpa...@internode.on.net writes:

Nick As I have just had a rather powerful evaluation server to play
Nick around with for a few days while I tested our various Windows
Nick and Linux server builds on it, I thought I'd also take the
Nick opportunity to compare the build speed of a reasonably
Nick substantial score. I used Reinhold's setting of Reubke's sonata
Nick on the 94th psalm. I tested on three machines, all running the
Nick same version of Lilypond:

Nick 1. Dell GX620 workstation, Pentium D dual-core 3.0GHz CPU,
Nick1Gb RAM, Ubuntu 9.04 x86: 10min 11sec 
Nick 2. Dell GX745 workstation, Pentium D dual-core 3.4GHz CPU,
Nick2Gb RAM, WinXP SP3: 9min 22sec 
Nick 3. PowerEdge R710 server, dual quad-core Xeon 5560 2.8GHz CPUs,
Nick24Gb RAM, Debian 5 amd64: 4min 4sec


I think you'll find the main difference is in size of L2/L3 cache,
and amount of RAM.  Lily (like many object-oriented programs) tends to
have quite a deep stack, and to use lots of memory --- which it
visits in what looks to the processor like random orders --- so small
caches generate lots of cache misses, which slows things down.  If you
run out of RAM and have to swap, things get even worse.

Xeon 5560: 256k L2, 8M L3 cache (which is almost as fast as the Pentium D's L2 
cache)
Pentium D: 1M L2 cache, no L3 cache.
--
Dr Peter Chubb  www.nicta.com.aupeter DOT chubb AT nicta.com.au
http://www.ertos.nicta.com.au   ERTOS within National ICT Australia


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Han-Wen Nienhuys]

On Mon, Aug 31, 2009 at 8:03 PM, Peter
Chubblily.u...@chubb.wattle.id.au wrote:

 I think you'll find the main difference is in size of L2/L3 cache,
 and amount of RAM.  Lily (like many object-oriented programs) tends to
 have quite a deep stack, and to use lots of memory --- which it
 visits in what looks to the processor like random orders --- so small
 caches generate lots of cache misses, which slows things down.  If you
 run out of RAM and have to swap, things get even worse.

More importantly: LilyPond is single-threaded, so the number of cores
is irrelevant.

-- 
Han-Wen Nienhuys - han...@xs4all.nl - http://www.xs4all.nl/~hanwen


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Peter Chubb]

 Han-Wen == Han-Wen Nienhuys hanw...@gmail.com writes:

Han-Wen On Mon, Aug 31, 2009 at 8:03 PM, Peter
Han-Wen Chubblily.u...@chubb.wattle.id.au wrote:

 I think you'll find the main difference is in size of L2/L3 cache,
 and amount of RAM.  Lily (like many object-oriented programs) tends
 to have quite a deep stack, and to use lots of memory --- which it
 visits in what looks to the processor like random orders --- so
 small caches generate lots of cache misses, which slows things
 down.  If you run out of RAM and have to swap, things get even
 worse.

Han-Wen More importantly: LilyPond is single-threaded, so the number
Han-Wen of cores is irrelevant.

That doesn't explain why going from the Core Duo to the Xeon
dropped the time from 11 minutes to 4 minutes.  The reason, as I said,
is the increased cache size.

--
Dr Peter Chubb  www.nicta.com.aupeter DOT chubb AT nicta.com.au
http://www.ertos.nicta.com.au   ERTOS within National ICT Australia


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Lilypond Speed

[Jonathan Wilkes]

Hello,
 I'm curious how long it takes for other people to run lilypond on the 
following simple score:

\relative c' {
  c4 d e fis
}

I'm on winxp sp3, pentium 1.7Ghz with 512mb ram and it consistently takes 
7 seconds to complete, whether I do it on the command line or in 
LilypondTool.

-Jonathan


  


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Jethro Van Thuyne]

On Fri, 28 Aug 2009, Jonathan Wilkes wrote:


I'm curious how long it takes for other people to run lilypond on the
following simple score:


It took me 4,474 seconds


on winxp sp3, pentium 1.7Ghz with 512mb ram


on Debian, Intel Core2Duo T7250 @ 2.00GHz with 2GB ram.

Jethro.



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[James E. Bailey]

On 28.08.2009, at 18:45, Jonathan Wilkes wrote:


Hello,
 I'm curious how long it takes for other people to run lilypond  
on the

following simple score:

\relative c' {
  c4 d e fis
}

I'm on winxp sp3, pentium 1.7Ghz with 512mb ram and it consistently  
takes

7 seconds to complete, whether I do it on the command line or in
LilypondTool.

-Jonathan



Using time on osx 10.4.11 I get:
with generating the pdf:
real0m1.482s
user0m1.206s
sys 0m0.255s

just generating a ps:
real0m1.307s
user0m1.091s
sys 0m0.201s


James E. Bailey



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Federico Bruni]

Jethro Van Thuyne wrote:

On Fri, 28 Aug 2009, Jonathan Wilkes wrote:


I'm curious how long it takes for other people to run lilypond on the
following simple score:


It took me 4,474 seconds



How can you be so precise? :-)

Frescobaldi does not give me timing information.
If I type in a terminal lilypond -V file.ly the only information about 
timing is:

elapsed time: 0.01 seconds

It seems too much fast, anyway it's within 1 second.

Ubuntu, Intel Core2Duo T7500 @ 2.20GHz with 2GB ram


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Kieren MacMillan]

Hi Jonathan,


I'm curious how long it takes for other people to run lilypond on the
following simple score:

\relative c' {
  c4 d e fis
}


About 1.5 seconds on my MacBook 667GHz G5 w/1GB RAM.
Kieren.


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Jethro Van Thuyne]

It took me 4,474 seconds
on Debian, Intel Core2Duo T7250 @ 2.00GHz with 2GB ram.


And with a version statement it takes 0,941 seconds...

Jethro.


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[James E. Bailey]

On 28.08.2009, at 19:35, Federico Bruni wrote:


Jethro Van Thuyne wrote:

On Fri, 28 Aug 2009, Jonathan Wilkes wrote:
I'm curious how long it takes for other people to run lilypond on  
the

following simple score:

It took me 4,474 seconds


How can you be so precise? :-)



OSX has a time command that can be run from the command line and will  
output the timing of the process that follows it. I learned that  
because of this query.


James E. Bailey



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Frank Steinmetzger]

Am Freitag, 28. August 2009 schrieb Jonathan Wilkes:
 Hello,
  I'm curious how long it takes for other people to run lilypond on the
 following simple score:

 \relative c' {
   c4 d e fis
 }

 I'm on winxp sp3, pentium 1.7Ghz with 512mb ram and it consistently takes
 7 seconds to complete, whether I do it on the command line or in
 LilypondTool.

Core2Duo T7200 at 1GHz:

Creating a PS:
real0m1.502s
user0m1.418s
sys 0m0.073s

Creating a PDF:
real0m1.808s
user0m1.644s
sys 0m0.138s

And at 2GHz:

Creating a PS:
real0m0.810s
user0m0.763s
sys 0m0.040s

Creating a PDF:
real0m0.973s
user0m0.879s
sys 0m0.083s

(all values +/- 20ms)

Have you tried right after booting into Windows? It is known to have a faible 
for swapping here and there even if the RAM isn't filled up yet.
-- 
Gruß | Greetings | Qapla'
Windows: reboot. Linux: be root.


signature.asc
Description: This is a digitally signed message part.
___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Wilbert Berendsen]

on a Intel(R) Core(TM)2 Duo CPU E7400  @ 2.80GHz :
wilb...@sweelinck:~/ly/test$ echo \\relative c' { c4 d e fis }  test.ly
wilb...@sweelinck:~/ly/test$ time lilypond test   
GNU LilyPond 2.13.1
Verwerken van `test.ly'
Ontleden...
test.ly:0: warning: geen \version uitdrukking gevonden, voeg

\version 2.13.1

toe voor toekomstige compatibiliteit
Vertolken van muziek...
Voorbewerken van grafische objecten...
Solving 1 page-breaking chunks...[1: 1 pages]
Tekenen van systemen...
Opmaakuitvoer naar `test.ps'...
Converteren naar `./test.pdf'...

real0m0.695s
user0m0.612s
sys 0m0.064s


best regards,
Wilbert Berendsen

-- 
Frescobaldi, LilyPond editor for KDE: http://www.frescobaldi.org/
Nederlands LilyPond forum: http://www.lilypondforum.nl/


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Trevor Daniels]

Jonathan Wilkes Friday, August 28, 2009 5:45 PM



Hello,
I'm curious how long it takes for other people to run lilypond 
on the

following simple score:

\relative c' {
 c4 d e fis
}

I'm on winxp sp3, pentium 1.7Ghz with 512mb ram and it 
consistently takes

7 seconds to complete, whether I do it on the command line or in
LilypondTool.


3.5 secs on 1.66GHz Core 2 Duo with 2Gb DDR2
running Windows Vista Home Premium, from start
to end of converting to pdf.  This is the
time after the first run, which took c. 10 secs,
due to having to page-in/load the software (I
had around 12 windows open at the time and hadn't
run LilyPond recently).

Trevor




___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Thomas Scharkowski]

I'm sure it is a little more, but not much ;-)

Thomas 
Intel E6750 @ 2.66Ghz, 2 GM RAM
Windows Xp SP3, LilyPondTool

--
Processing `C:/LilyPondFiles/test/time.ly'
Parsing...
Interpreting music... 
Preprocessing graphical objects...
Solving 1 page-breaking chunks...[1: 1 pages]
Drawing systems...
Layout output to `time.ps'...
Converting to `./time.pdf'...
Processing time:  0  seconds

LilyPond ready.  
--


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Wilbert Berendsen]

Op vrijdag 28 augustus 2009, schreef Federico Bruni:
 Frescobaldi does not give me timing information.
I just implemented this in SVN! ;-)
best regards,
Wilbert Berendsen

-- 
Frescobaldi, LilyPond editor for KDE: http://www.frescobaldi.org/
Nederlands LilyPond forum: http://www.lilypondforum.nl/


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Jonathan Wilkes]

It sounds like there is a wide discrepancy depending on machine/os/etc. 

Does anyone have any insight into how I could decrease this time on my 
winxp machine?

I feel like if I could get it down to something close to one second, it 
would be a lot easier to learn Lilypond.

Also: is there a way using LilypondTool to set Lilypond to run every 
time I enter a barcheck?

Thanks,
Jonathan


--- On Fri, 8/28/09, Thomas Scharkowski t.scharkow...@t-online.de wrote:

 From: Thomas Scharkowski t.scharkow...@t-online.de
 Subject: Re: Lilypond Speed
 To: lilypond-user@gnu.org, Jonathan Wilkes jancs...@yahoo.com
 Date: Friday, August 28, 2009, 8:59 PM
 I'm sure it is a little more, but not
 much ;-)
 
 Thomas 
 Intel E6750 @ 2.66Ghz, 2 GM RAM
 Windows Xp SP3, LilyPondTool
 
 --
 Processing `C:/LilyPondFiles/test/time.ly'
 Parsing...
 Interpreting music... 
 Preprocessing graphical objects...
 Solving 1 page-breaking chunks...[1: 1 pages]
 Drawing systems...
 Layout output to `time.ps'...
 Converting to `./time.pdf'...
 Processing time:  0  seconds
            
         
 LilyPond ready.       
      
 --
 





___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Federico Bruni]

Wilbert Berendsen wrote:

on a Intel(R) Core(TM)2 Duo CPU E7400  @ 2.80GHz :
wilb...@sweelinck:~/ly/test$ echo \\relative c' { c4 d e fis }  test.ly
wilb...@sweelinck:~/ly/test$ time lilypond test   


ok, this is the command I was looking for..
so...

on a Intel Core2Duo T7500 @ 2.20GHz with 2GB ram (Ubuntu)

f...@fede-laptop:/tmp$ time lilypond speed.ly
GNU LilyPond 2.13.3
Processing `speed.ly'
Parsing...
Interpreting music...
Preprocessing graphical objects...
Solving 1 page-breaking chunks...[1: 1 pages]
Drawing systems...
Layout output to `speed.ps'...
Converting to `./speed.pdf'...

real0m1.021s
user0m0.881s
sys 0m0.099s


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Tim Reeves]

Hello,
 I'm curious how long it takes for other people to run lilypond on the 

following simple score:

\relative c' {
  c4 d e fis
}

I'm on winxp sp3, pentium 1.7Ghz with 512mb ram and it consistently takes 
7 seconds to complete, whether I do it on the command line or in 
LilypondTool.

-Jonathan


Using LilypondTool, the first time I compiled it took 27 seconds (!), then 
I  added the version statement (based on someone else's comment that it 
became shorter) and it took 0 seconds (that was the console output - I 
don't know how to get a more precise time - I would assume that means 0.5 
sec) and then I removed the version statement and it took 0 seconds again. 
LP 2.13.3, LPT 2.12.869, WinXP SP3, Core2Duo 2.8GHz, 2GB RAM.


Tim Reeves


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Frank Steinmetzger]

Am Freitag, 28. August 2009 schrieb Jonathan Wilkes:
 It sounds like there is a wide discrepancy depending on machine/os/etc.

 Does anyone have any insight into how I could decrease this time on my
 winxp machine?

Half way through reading that sentence I wanted to say install Linux. *dh*
-- 
Gruß | Greetings | Qapla'
LOL, you said ROFL.


signature.asc
Description: This is a digitally signed message part.
___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Lilypond Speed (Jonathan Wilkes)

[Frederick Dennis]

Dear Jonathan,
It takes me 11 seconds the first time, 4 seconds without a version number
and 3 seconds with a version number.
AMD Sempron 2500+
1.4 GHz, 448MB of RAM
Physical Address Extension
Windows XP Professional
Fred
___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Bertalan Fodor (LilyPondTool)]

Jonathan Wilkes wrote:
It sounds like there is a wide discrepancy depending on machine/os/etc. 

Does anyone have any insight into how I could decrease this time on my 
winxp machine?


I feel like if I could get it down to something close to one second, it 
would be a lot easier to learn Lilypond.


Also: is there a way using LilypondTool to set Lilypond to run every 
time I enter a barcheck?
  
Sure, as you can set a shortcut for running LilyPond, see 
UtilitiesGlobal optionsShortcuts, you can set the shortcut to be |


Bert



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond Speed

[Jonathan Wilkes]

--- On Sat, 8/29/09, Bertalan Fodor (LilyPondTool) lilypondt...@organum.hu 
wrote:

 From: Bertalan Fodor (LilyPondTool) lilypondt...@organum.hu
 Subject: Re: Lilypond Speed
 To: Jonathan Wilkes jancs...@yahoo.com
 Cc: lilypond-user@gnu.org
 Date: Saturday, August 29, 2009, 6:40 AM
 Jonathan Wilkes wrote:
  It sounds like there is a wide discrepancy depending
 on machine/os/etc. 
  Does anyone have any insight into how I could decrease
 this time on my winxp machine?
  
  I feel like if I could get it down to something close
 to one second, it would be a lot easier to learn Lilypond.
  
  Also: is there a way using LilypondTool to set
 Lilypond to run every time I enter a barcheck?
    
 Sure, as you can set a shortcut for running LilyPond, see
 UtilitiesGlobal optionsShortcuts, you can set the
 shortcut to be |
 
 Bert

Ah, of course.  Thanks a lot, Bert.

-Jonathan





___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond speed

[Valentin Villenave]

2009/8/4 Graham Percival gra...@percival-music.ca:
 There you go:
 http://lists.gnu.org/archive/html/lilypond-devel/2005-11/msg00024.html

This is huge! (I suspect I wasn't subscribed to -devel when this was
posted, otherwise I'd have noticed it).

Even though there's clearly no magic recipe to speed up LilyPond
(except multi-threading, but we're nowhere near implementing it), the
server approach could be very, very useful for all kind of purposes.

Is there a way to post this to the LSR? (this is where the newly-added
devel tag might be useful...)

Regards,
Valentin


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond speed

[Graham Percival]

On Tue, Aug 04, 2009 at 12:42:15PM +0200, Valentin Villenave wrote:
 2009/8/4 Graham Percival gra...@percival-music.ca:
  There you go:
  http://lists.gnu.org/archive/html/lilypond-devel/2005-11/msg00024.html
 
 This is huge! (I suspect I wasn't subscribed to -devel when this was
 posted, otherwise I'd have noticed it).

 Even though there's clearly no magic recipe to speed up LilyPond
 (except multi-threading, but we're nowhere near implementing it), the
 server approach could be very, very useful for all kind of purposes.

Why do people never believe me when I say that there's tons of
cool stuff we /could/ do, if only more people helped out?

If the website was finished earlier (i.e. if people contributed
content), then I'd have worked on this during the summer.  If more
people helped out, we could have a much better set of safe
lilypond commands.  The above two points would let us run
multi-threaded web-available lilypond servers for doing multiple
small snippets at once.  If more people helped out, we could have
started+finished GLISS already, and have a stable syntax (for the
commands).  If more people helped out, we might actually have a
*decreasing* list of issues.

There are tons of cool stuff we /could/ do.  In my idle moments, I
make plans of how it would be organized, how the overall work
structure would go, etc.  But there's no point trying to do cool
stuff unless the foundation is solid.  We need more people working
on those foundations.

Cheers,
- Graham

PS yes, I was planning on writing an article about all the cool
stuff we *could* be doing, if only people helped out with the
mundane/routine jobs, as a continuation of my Report contribution.
Of course, there's no point writing the sequel until the first one
is published.



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond speed

[Valentin Villenave]

2009/8/4 Graham Percival gra...@percival-music.ca:

 Why do people never believe me when I say that there's tons of
 cool stuff we /could/ do, if only more people helped out?

That's not my point. My point is to make sure that nothing potentially
cool gets lost.

 Of course, there's no point writing the sequel until the first one
 is published.

Good to see you're still in good shape :)

Regards,
Valentin


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond speed

[hernan gonzalez]

 I read in the main.cc source that the GUILE start-up is very time 
 consumming. I
 wonder if some modification in the code could be done so that the GUILE 
 startup
 occurs once for several compilation cycles, something as (pseudocode)

 Yes, that's been done with the lilypond server.  Search the
 lilypond-devel mailist for details.


Well, I've been searching, but I've found nothing relevant...
http://www.mail-archive.com/search?l=lilypond-devel%40gnu.orgq=guile+server
... just some mention (from yourself) telling this same advice (search
the mailing lists)...
http://www.mail-archive.com/lilypond-de...@gnu.org/msg22084.html

About lilypond server I only found some scattered references to
running in a web server
(and some about the jail options and related stuff), but no mention
about what interest
me, a way of run the lilypond executable (desktop mode) and keep it running,
over the same files in a loop.

Peeking into the sources, I see that it can't easily been done from
the main C entry point,
because the main scm module does not return to the caller, but instead does
a hard exit. Could (should?) this be changed ?

Any other hint or pointer ? I'd guessed that this is a relevant issue
for the typical lilypond user.

Best regards.

Hernán J. González
http://hjg.com.ar/ghibli/musica/


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond speed

[Graham Percival]

On Mon, Aug 03, 2009 at 12:10:18PM -0300, hernan gonzalez wrote:
  I read in the main.cc source that the GUILE start-up is very time 
  consumming. I
  wonder if some modification in the code could be done so that the GUILE 
  startup
  occurs once for several compilation cycles, something as (pseudocode)
 
  Yes, that's been done with the lilypond server.  Search the
  lilypond-devel mailist for details.
 
 Well, I've been searching, but I've found nothing relevant...

There you go:
http://lists.gnu.org/archive/html/lilypond-devel/2005-11/msg00024.html

 Any other hint or pointer ? I'd guessed that this is a relevant issue
 for the typical lilypond user.

No, it's not relevant for the typical lilypond user.  We'd need at
least 20 hours of programming until this became relevant for the
typical lilypond user.

Cheers,
- Graham



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond speed

[Han-Wen Nienhuys]

On Mon, Aug 3, 2009 at 12:19 AM, hernanhgonza...@gmail.com wrote:
 My main frustration with Lilypond is speed. In my setup (Win-XP, P4 3.0Ghz, 1G
 ram) to process a fairly simple scoresheet (2 or 3 pages) it takes about 8
 seconds. That might not seem a great deal, but it is really annoying when one 
 is
 doing lots of retouching (edit one bit, compile, see results, edit again... 
 etc)

 Is there some recipe to speed things up? Are the performance bottlenecks
 identified?

 I read in the main.cc source that the GUILE start-up is very time consumming. 
 I
 wonder if some modification in the code could be done so that the GUILE 
 startup
 occurs once for several compilation cycles, something as (pseudocode)

The startup time consumed by GUILE is less than 0.5 second.  This will
not really make a dent in the processing time.


-- 
Han-Wen Nienhuys - han...@xs4all.nl - http://www.xs4all.nl/~hanwen


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Lilypond speed

[hernan]

My main frustration with Lilypond is speed. In my setup (Win-XP, P4 3.0Ghz, 1G
ram) to process a fairly simple scoresheet (2 or 3 pages) it takes about 8
seconds. That might not seem a great deal, but it is really annoying when one is
doing lots of retouching (edit one bit, compile, see results, edit again... 
etc) 

Is there some recipe to speed things up? Are the performance bottlenecks
identified? 

I read in the main.cc source that the GUILE start-up is very time consumming. I
wonder if some modification in the code could be done so that the GUILE startup
occurs once for several compilation cycles, something as (pseudocode)

main() {
DO_GUILE_START_UP();
do {
COMPILE_FILES()
PROMPT(PRESS 'E' TO END, 'R' TO RECOMPILE THE SAME FILE/S)
} while (GETCHAR()=='R');
exit(0);
}



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Lilypond speed

[Graham Percival]

On Mon, Aug 03, 2009 at 12:19:28AM +, hernan wrote:
 Is there some recipe to speed things up? Are the performance bottlenecks
 identified? 

There are some minor tweaks you can do.  I think they're currently
listed in LM 5.  Speeding up typesetting or something like that.

 I read in the main.cc source that the GUILE start-up is very time consumming. 
 I
 wonder if some modification in the code could be done so that the GUILE 
 startup
 occurs once for several compilation cycles, something as (pseudocode)

Yes, that's been done with the lilypond server.  Search the
lilypond-devel mailist for details.

Cheers,
- Graham


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Speed-up compiling

[Michael Käppler]

Hi all,
I'm suffering from enormous compiling durations on large files. With 
large I mean a file with about 250 measures and seven staves per 
system. The last time I compiled the file completely (without using 
showLastLength) I did it overnight, since after one and a half hour 
Lilypond still was Preformatting graphical elements... (Don't know 
what's the correct term in the english version)
I know that's my laptop is way too old to do such complex task quickly - 
(Athlon XP 2600, 256MB Ram, OpenSUSE 11.1) but are there general 
suggestions which help to speed up the compiling process?


Regards,
Michael


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Speed-up compiling

[Mark Polesky]

0123456789012345678901234567890123456789012345678901234567890123456
Michael Käppler wrote:

 I'm suffering from enormous compiling durations on large files.
 With large I mean a file with about 250 measures and seven
 staves per system. The last time I compiled the file completely
 (without using showLastLength) I did it overnight,  since after
 one and a half hour Lilypond still was Preformatting graphical 
 elements... (Don't know what's the correct term in the english
 version) I know that's my laptop is way too old to do such
 complex task quickly - (Athlon XP 2600, 256MB Ram, OpenSUSE 11.1)
 but are there general suggestions which help  to speed up the
 compiling process?

An hour and a half sounds much, much longer than it should be
taking. Here are some thoughts:

Maybe try using \pointAndClickOff ? I don't know if point-and-click
slows things down, but it certainly increases filesize.
http://lilypond.org/doc/v2.12/Documentation/user/lilypond-program/Point-and-click

Also, you can temporarily comment out any instances of 
keep-inside-line since it's a known time-hog.
http://lilypond.org/doc/v2.12/Documentation/user/lilypond-learning/Avoiding-tweaks-with-slower-processing

Is your computer slow with other programs? Could you have a computer
virus? Maybe 256MB RAM is too small. Maybe you could try an external
laptop RAM stick? RAM sizes of 4,096MB are not uncommon these days.

Hope this helps.
- Mark






___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Speed-up compiling

[Michael Käppler]

Hi Mark,

Maybe try using \pointAndClickOff ? I don't know if point-and-click
slows things down, but it certainly increases filesize.
http://lilypond.org/doc/v2.12/Documentation/user/lilypond-program/Point-and-click
  

I've tested this, rendering only the last 100 measures:
When I ran Lily the first time on that file, it needed 171 s with 
Point-And-Click turned on. Afterwards I turned Point-And-Click off, 
where it finished after 88s. After that I did a second run with 
Point-And-Click turned on - now it finished after just 93s.
I don't think there are significant differences concerning 
Point-And-Click. But obviously every first run that Lily does on a file, 
lasts longer.
I don't know what's the reason for it. A further observation I made 
(which I posted some time ago on -devel) was that when running LilyPond 
with many files (lilypond foo1.ly foo2.ly foo3.ly foo4.ly) it needs 
much(!) more time than running it that way: lilypond foo1.ly  lilypond 
foo2.ly  lilypond foo3.ly  lilypond foo4.ly.
Also, you can temporarily comment out any instances of 
keep-inside-line since it's a known time-hog.

http://lilypond.org/doc/v2.12/Documentation/user/lilypond-learning/Avoiding-tweaks-with-slower-processing
  

I don't use any of this extra-checks in my score.


Regards,
Michael


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Speed-up compiling

[Jonathan Kulp]

On Wed, Jul 29, 2009 at 5:48 AM, Michael Käppler xmichae...@web.de wrote:

 Hi Mark,

 Maybe try using \pointAndClickOff ? I don't know if point-and-click
 slows things down, but it certainly increases filesize.

 http://lilypond.org/doc/v2.12/Documentation/user/lilypond-program/Point-and-click


 I've tested this, rendering only the last 100 measures:
 When I ran Lily the first time on that file, it needed 171 s with
 Point-And-Click turned on. Afterwards I turned Point-And-Click off, where it
 finished after 88s. After that I did a second run with Point-And-Click
 turned on - now it finished after just 93s.
 I don't think there are significant differences concerning Point-And-Click.
 But obviously every first run that Lily does on a file, lasts longer.
 I don't know what's the reason for it. A further observation I made (which
 I posted some time ago on -devel) was that when running LilyPond with many
 files (lilypond foo1.ly foo2.ly foo3.ly foo4.ly) it needs much(!) more
 time than running it that way: lilypond foo1.ly  lilypond foo2.ly 
 lilypond foo3.ly  lilypond foo4.ly.

 Also, you can temporarily comment out any instances of keep-inside-line
 since it's a known time-hog.

 http://lilypond.org/doc/v2.12/Documentation/user/lilypond-learning/Avoiding-tweaks-with-slower-processing


 I don't use any of this extra-checks in my score.


 Regards,
 Michael


 Hi Michael,

Given the specs of your machine, I think the compile time is probably about
what you can expect for such a large score. I have an orchestral score with
181 bars and 18-20 staves per system, and it compiles in about 7-8 minutes
on my laptop (about a year old with 3 GB RAM, dual-core processors, etc) but
takes more than an hour on my old eMac (5 years old, 1 GB ram, 1.25Ghz G4
processor). The processing has always been slower on macs, in my
observations. Even on my nice Intel iMac at work, which has pretty good
specs, it always takes considerably longer than on my Linux laptop.

Jon
-- 
Jonathan Kulp
http://www.jonathankulp.com
___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Speed-up compiling

[Wilbert Berendsen]

Op woensdag 29 juli 2009, schreef Michael Käppler:
 I'm suffering from enormous compiling durations on large files.

The 256 MB RAM is a bit small. In my experience, LilyPond uses enormous 
amounts of memory compiling large scores. Engraving my Reubke score (one 28 
page \score object), the LilyPond process uses more than 600MB. Enlarging the 
amount of memory available will speed up things considerably.

best regards,
Wilbert Berendsen

-- 
Frescobaldi, LilyPond editor for KDE: http://www.frescobaldi.org/


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: Speed-up compiling

[Patrick Horgan]

Michael Kppler wrote:
Hi all,
  
I'm "suffering" from enormous compiling durations on large files. With
"large" I mean a file with about 250 measures and seven staves per
system. The last time I compiled the file completely (without using
showLastLength) I did it overnight, since after one and a half hour
Lilypond still was "Preformatting graphical elements..." (Don't know
what's the correct term in the english version)
  
I know that's my laptop is way too old to do such complex task quickly
- (Athlon XP 2600, 256MB Ram, OpenSUSE 11.1) but are there general
suggestions which help to speed up the compiling process?
  

The one thing you could do is greatly increase your ram to 1GB. You'll
think you have a new computer for $50-$150. Here's why.

Here's the problem.

You're thrashing to disk, which means the amount of memory used by
lilypond is far larger than the available real memory and your system
is spending more time moving information between disk and memory than
it is actually running your programs. Modern operating systems support
virtual memory which means that you can use a larger amount of pretend
memory than the existing real hardware memory. As much of the virtual
memory will be in the hardware memory as possible including the stuff
you're using the most, and the part that doesn't fit will be located in
a special area of your hard drive. That means that when lilypond or
another application wants to use more than 256MB of memory to speed
things up, some of the stuff that should be in hardware memory for
speed gets pushed out to the disk. When lilypond tries to access the
part of memory that's actually on disk, it has to move something else
from hardware memory to disk to make room, and then read the part
you're trying to access from disk back into hardware memory. 

Disks are WAY SLOWER than memory. A common disk these days might take
10ms (.010 seconds), on average, to access a bit of data. Memory would
likely take 10ns (.000 000 010 seconds) to access a bit of data. That
means that getting the data out of memory is a MILLION times faster.
This is, I'm sure, slowing down many other things for you, not just
lilypond. 256MB or even 512MB is not enough for current software on
current operating systems. The truth is even worse than I just
explained, because you never get to use the whole 256MB anyway. Some
of that hardware memory is reserved by the OS for the kernel. It never
gets swapped out. It's locked into physical memory. On my Ubuntu box
right now it's 45MB, although less would be reserved
if there was less available memory--I have 2GB physical memory. Be
thankful you're not using Microsoft Windows. It uses up a LOT more
memory than Linux just to present you with an interface you can use to
run programs. Your system would die trying to start up.

Here's the fix.

Without buying a new computer, you can buy new RAM (Random Access
Memory) and you'll THINK that you have a new computer. It will make
you amazed at the difference! You will dance around and shout Huzzah!
Woman will hold up their babies to be kissed by you. You will get a
promotion and a humanitarian award. Yes! It makes that big a
difference! To find out what kind of RAM you need you can go to the
online memory sites of memory vendors like Kingston or Crucial who have tools you can use
to look up what kind of memory your computer takes. Your laptop with
Athlon XP 2600, is likely to max out at 1GB-1 2GB. If it supports 1
1/2 GB don't do it, just get 1GB. Memory likes to be in amounts that
are powers of two (1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048,
etc.) 1 1/2GB is not any of these and while your hardware may have
special circuitry to support that strange amount, it will actually slow
you down! 

You'll be taking out the 256MB chip and throwing it away. Depending on
the hardware you'll support a maximum size CHIP of either 512MB or
1GB. Two 512MB chips supply as much memory as one 1GB chip. If you
support the 1GB chip, just buy one of those and for around $50 you'll
be happy. If the maximum size you support is 512MB, you'll have to buy
two of those and it will probably cost you $60-$120. It's possible
that your hardware will support expanding to 2GB with two 1GB chips for
around $100-$120. That's up to you, but you wont see a dramatic change
for most applications after 1GB. 2GB will let you run more than one
memory hungry thing at a time and some rare applications can use the
extra memory by themselves (including lilypond with a large enough
input file). It's up to you.

Go to this
CNET video for more tips about how to expand RAM. If you search
around there's lots of videos and tutorials about how to do it. If you
have any more questions just ask.

Best regards,
Patrick

p.s. Memory uses a lot less energy that spinning a disk, so if you have
power saving on which stops spinning your drive when you aren't using
it, you will save much energy and get much longer battery life, and
help the

Re: Speed-up compiling

[M Watts]

On 07/30/2009 04:29 AM, Patrick Horgan wrote:


Here's the fix.

Without buying a new computer, you can buy new RAM (Random Access
Memory) and you'll THINK that you have a new computer. It will make you
amazed at the difference! You will dance around and shout Huzzah! Woman
will hold up their babies to be kissed by you. You will get a promotion
and a humanitarian award. Yes! It makes that big a difference!...


Thanks for this amusing and informative post.

The moral of the story is: Fill your machine with RAM -- it'll go like a 
rocket.



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Lilypond speed on Windows/Linux

[Thomas Scharkowski]

FYI:
I have tested LilyPond 2.7.18 on Windows XP and Kanotix/Debian, same box 
(quite old), same file, same HD, both with jEdit:

Windows: 52 seconds
Linux: 27 seconds

Thomas


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: lilypond speed

[Roman V. Isaev]

On 09/12, Han-Wen Nienhuys wrote:
 Roman V. Isaev wrote:
 
  Well, I have it both ways. It's slow if I use cygwin, it's slow
 if I run it from cmd.exe prompt or drag a .ly file on its icon.
 That's not the question. My question is whether you're using the
 Native 2.6 binary (available from lilypond.org/web/install/ for the 
 windows platform).

Yes. And it was slow when it was 2.5.x.

-- 
 Roman V. Isaev http://www.gunlab.com.ru Moscow, Russia



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

lilypond speed

[Roman V. Isaev]

Why lilypond on windows is VERY slow?! It takes almost 30 seconds to 
complete
something that compiles in less than a second on Fedora Core 4... I'm shocked.
For some reasons I can't use Fedora at home and it's very annoying to wait so
much for a little correction :( I thought lilypond is slow by design, but
when I installed FC4 at work and decided to try unix version I was truly 
shocked...
The difference between windows and unix computer's hardware  does exist, but 
certainly
it's not an order of 100 times

-- 
 Roman V. Isaev http://www.soprano-recorder.ru Moscow, Russia



___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user

Re: lilypond speed

[Han-Wen Nienhuys]

Roman V. Isaev wrote:

Why lilypond on windows is VERY slow?! It takes almost 30 seconds to 
complete
something that compiles in less than a second on Fedora Core 4... I'm shocked.
For some reasons I can't use Fedora at home and it's very annoying to wait so
much for a little correction :( I thought lilypond is slow by design, but
when I installed FC4 at work and decided to try unix version I was truly 
shocked...
The difference between windows and unix computer's hardware  does exist, but 
certainly
it's not an order of 100 times


Are you using Cygwin?  Do you have many fonts? The difference in speed 
between the mingw and FC binaries should be neglible.


--
 Han-Wen Nienhuys - [EMAIL PROTECTED] - http://www.xs4all.nl/~hanwen


___
lilypond-user mailing list
lilypond-user@gnu.org
http://lists.gnu.org/mailman/listinfo/lilypond-user