Speed/latency issues for development in a Wine environment

[Alan W. Irwin]

To introduce myself to this list, I am a complete Wine newbie with no
Windows experience at all, but I do have years of experience developing
software on Linux, and in the interests of testing that software on a free
Windows platform, I have recently been using the combination of Windows
native versions of MinGW and CMake to build and test applications and
libraries under Wine. That is, I run a Windows version of cmake under wine
to configure Makefiles followed by running a Windows version of MinGW make
under Wine (which in turn runs MinGW gcc and other applications under Wine
to build and test the software).

So far this idea is going rather well, but I have run into some issues which
I have been reporting on the wine-users mailing list. I find such
preliminary discussion useful to help prepare a good bug report, but
following a suggestion from Alexandre Julliard, I am moving the discussion
of these issues here.

Since at least some of you have already seen my posts on wine-users, I
will only summarize the issues as I currently see them.

The rest of this post concerns a slow software build issue I have discovered
under Wine, and I will deal with a second issue in a later post.

The issue is that Wine builds of software take roughly a factor of 5 longer
than the equivalent Linux builds.  This issue appears to be caused mostly by
command startup latency.  What I mean by that phrase is

wine cmake --version

(and the "--version" option for many commands you can run under wine) take
roughly 150 ms (with wine-1.2-rc3 built with the -O3 option and running that
wine version with WINEDEBUG='-all' and wineserver -p) to complete while
Linux versions of the same commands require 1 ms to complete.

It is a startup latency in the commands run under wine rather than a startup
latency issue with the wine command itself because that command returns much
faster than 150 ms if you give it an unknown command.

Because many thousands of commands (e.g., "gcc"; several "cmake -E" variants
per gcc command to help report progress; figure out dependencies, etc.; and
various further make commands generated by the initial make command) are
used to complete a typical build (say of cmake itself or PLplot), the
startup latency for each of those commands accumulates and is a major factor
which slows down the build (by roughly a factor of 5) compared to the
equivalent Linux build.  So a reduction of the command startup latency under
Wine by an order of magnitude should largely solve the problem.

I find the command startup latency depends very much on what X server I am
using. Of course, the actual commands I am using (cmake, MinGW make, MinGW
gcc, etc.) to build software under wine have nothing to do with X.  So I am
hoping if I build a stripped down version of wine configured to ignore X and
all other systems (sound, etc.) that are not required for builds, then I
might achieve the reduced command startup latency that I need.

Before I try that, however, is there some less brute-force way to
reduce command startup latency?

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); PLplot scientific plotting software
package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of
Linux Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

---------- Forwarded message ----------
Date: Sat, 12 Jun 2010 14:10:31 -0700 (PDT)
From: Alan W. Irwin <irwin at beluga.phys.uvic.ca>
To: wine-users at winehq.org
Subject: Re: [Wine] Speed/latency issues for development in a Wine environment

On 2010-06-11 18:09-0700 Alan W. Irwin wrote:

The previous timing numbers were done with wine-1.1.42.  Following a
suggestion made off-list to me, I have built wine-1.2-rc3 today (with
CFLAG=-O3) and immediately noticed the start-up latency for simple commands
(e.g., cmake --version) was reduced from about 0.25 s to about 0.15 s. I
don't know how the previous Debian wine-1.1.42 packages were optimized so
that difference may just be due to an optimization difference or an actual
improvement in the Wine code from 1.1.42 to 1.2-rc3. However, to see how -O3
optimized wine-1.2-rc3 does with the full tests, I have redone my timing
tests for I. (cmake), II (make install), and III (make install latency) and
the following reductions have been observed.

> Wine I.
> wine at raven> time wine cmake -G "MinGW Makefiles" \
> "-DCMAKE_INSTALL_PREFIX=z:/home/wine/cmake/install1" \
> z:/home/software/cmake/cmake-2.8.1_patched/ >& cmake.out
> 
> real    14m17.124s
==> real    8m18.041s

> Wine II.
> 
> wine at raven> time wine mingw32-make install >& make_install.out
> 
> real    20m13.421s
==>
real    12m4.198s

> Wine III.
> wine at raven> time wine mingw32-make install >& make_install.out1
> 
> real    1m39.018s
==>
real    0m51.045s

For reference here are the corresponding Linux numbers which are still
much better than the corresponding Wine-1.2-rc3 results.

> Linux I.
> 
> wine at raven> time cmake -G "Unix Makefiles" \
> "-DCMAKE_INSTALL_PREFIX=/home/wine/cmake/install_linux" \
> /home/software/cmake/cmake-2.8.1_patched/ >& cmake.out
> 
> real    0m34.015s
> Linux II.
> 
> wine at raven> time make install >& make_install.out
> 
> real    2m12.805s
> Linux III.
> 
> wine at raven> time make install >& make_install.out1
> 
> real    0m0.869s

To summarize these findings, it appears that changes in startup latency for
simple commands strongly correlates with timing results for build steps I,
II, and III on Wine. This makes sense since we know that several thousand
commands are run for, e.g., build step II so startup latency of hundreds of
milliseconds per command must be a real build performance killer.

So what can be done to reduce the simple startup latency from the current
~150 ms on Wine down much closer to the Linux startup latency for the same
commands which is typically 1 ms?  Simple commands requiring 150 ms to start
must have an awful lot of extra start baggage.  Is all that baggage really
needed for command-line commands such as cmake, mingw32-make, gcc, etc.?

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); PLplot scientific plotting software
package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of
Linux Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________