June 2008 - Wine-devel - list.winehq.org

Wine compiler options benchmarks
by Ben Hodgetts (Enverex) 16 Jun '08

16 Jun '08

Just in-case anyone was ever curious about how well Wine performs with different C/XXFLAGs I did a test today with RC5 to see how much of a difference it makes with 3DMark 2001 SE, nothing major but if someone can think of a better benchmark to try, please let me know (I had hoped to try Oblivion or some such but it has no benchmark feature). Core2Quad Q9450 @ 3.4Ghz || 4GB PC8500 5-5-5-15 || GCC 4.3.1 || Linux 2.6.25 3DMark 2001 SE B300. Wine 1.0-rc5. 28576 -march=native -O3 -pipe -fomit-frame-pointer 28522 -march=native -O3 -pipe -fomit-frame-pointer -mfpmath=sse,387 28511 -march=native -O2 -pipe -fomit-frame-pointer 28427 -march=native -O3 -fomit-frame-pointer -ffast-math -funroll-loops -Wall -pipe 28426 -march=native -O2 -pipe -mfpmath=sse,387 28311 -march=native -O2 -pipe 28270 -march=native -Os -pipe -fomit-frame-pointer 28126 -O3 -pipe -fomit-frame-pointer 28072 Wine Default (-g -O2) 27984 -march=native -O3 -pipe 27646 -march=native -Os -pipe For reference I tried the same with Lame just to show how it would affect other programs: Core2Quad Q9450 @ 3.4Ghz || 4GB PC8500 5-5-5-15 || GCC 4.3.1 || Linux 2.6.25 ~30 minute stereo wave file encoded in Lame Lame-3.97 64bit. real 1m28.739s -O3 -ffast-math -funroll-loops -Wall -pipe (Lame defaults) real 1m29.168s -march=native -O3 -ffast-math -funroll-loops -Wall -pipe real 1m35.633s -march=native -O3 -pipe -ffast-math real 1m35.923s -march=native -O3 -pipe -ftracer -ffast-math real 1m36.346s -march=native -O3 -pipe -ffast-math -falign-functions=64 real 1m36.768s -march=native -O2 -pipe -ftracer -ffast-math real 1m37.242s -march=native -O3 -pipe -ftracer real 1m38.127s -march=native -O3 -pipe -fomit-frame-pointer real 1m38.235s -march=native -O3 -pipe -fvect-cost-model real 1m38.243s -march=native -O2 -pipe -mfpmath=sse,387 real 1m38.278s -march=native -O2 -pipe real 1m38.423s -march=native -O2 -pipe -falign-functions=64 real 1m38.527s -march=native -O2 -pipe -ffast-math real 1m38.573s -march=native -O2 -pipe -falign-functions=128 real 1m38.575s -march=native -O2 -pipe -mfpmath=sse real 1m39.241s -march=native -O2 -pipe -ftracer real 1m39.261s -march=native -O2 -pipe -fomit-frame-pointer real 1m40.946s -march=native -Os -pipe -ffast-math real 1m41.385s -march=native -Os -pipe real 1m41.513s -march=native -Os -pipe -fomit-frame-pointer real 1m41.602s -march=native -Os -pipe -fomit-frame-pointer -mfpmath=sse,387 real 1m46.577s -march=native -O2 -pipe -mfpmath=387 Slightly more influential there. Ben H.

3 2

Wine compiler options benchmarks
by Ben Hodgetts (Enverex) 16 Jun '08

16 Jun '08

Just in-case anyone was ever curious about how well Wine performs with different C/XXFLAGs I did a test today with RC5 to see how much of a difference it makes with 3DMark 2001 SE, nothing major but if someone can think of a better benchmark to try, please let me know (I had hoped to try Oblivion or some such but it has no benchmark feature). Core2Quad Q9450 @ 3.4Ghz || 4GB PC8500 5-5-5-15 || GCC 4.3.1 || Linux 2.6.25 3DMark 2001 SE B300. Wine 1.0-rc5. 28576 -march=native -O3 -pipe -fomit-frame-pointer 28522 -march=native -O3 -pipe -fomit-frame-pointer -mfpmath=sse,387 28511 -march=native -O2 -pipe -fomit-frame-pointer 28427 -march=native -O3 -fomit-frame-pointer -ffast-math -funroll-loops -Wall -pipe 28426 -march=native -O2 -pipe -mfpmath=sse,387 28311 -march=native -O2 -pipe 28270 -march=native -Os -pipe -fomit-frame-pointer 28126 -O3 -pipe -fomit-frame-pointer 28072 Wine Default (-g -O2) 27984 -march=native -O3 -pipe 27646 -march=native -Os -pipe For reference I tried the same with Lame just to show how it would affect other programs: Core2Quad Q9450 @ 3.4Ghz || 4GB PC8500 5-5-5-15 || GCC 4.3.1 || Linux 2.6.25 ~30 minute stereo wave file encoded in Lame Lame-3.97 64bit. real 1m28.739s -O3 -ffast-math -funroll-loops -Wall -pipe (Lame defaults) real 1m29.168s -march=native -O3 -ffast-math -funroll-loops -Wall -pipe real 1m35.633s -march=native -O3 -pipe -ffast-math real 1m35.923s -march=native -O3 -pipe -ftracer -ffast-math real 1m36.346s -march=native -O3 -pipe -ffast-math -falign-functions=64 real 1m36.768s -march=native -O2 -pipe -ftracer -ffast-math real 1m37.242s -march=native -O3 -pipe -ftracer real 1m38.127s -march=native -O3 -pipe -fomit-frame-pointer real 1m38.235s -march=native -O3 -pipe -fvect-cost-model real 1m38.243s -march=native -O2 -pipe -mfpmath=sse,387 real 1m38.278s -march=native -O2 -pipe real 1m38.423s -march=native -O2 -pipe -falign-functions=64 real 1m38.527s -march=native -O2 -pipe -ffast-math real 1m38.573s -march=native -O2 -pipe -falign-functions=128 real 1m38.575s -march=native -O2 -pipe -mfpmath=sse real 1m39.241s -march=native -O2 -pipe -ftracer real 1m39.261s -march=native -O2 -pipe -fomit-frame-pointer real 1m40.946s -march=native -Os -pipe -ffast-math real 1m41.385s -march=native -Os -pipe real 1m41.513s -march=native -Os -pipe -fomit-frame-pointer real 1m41.602s -march=native -Os -pipe -fomit-frame-pointer -mfpmath=sse,387 real 1m46.577s -march=native -O2 -pipe -mfpmath=387 Slightly more influential there. Ben H.

3 2

Re: rpcrt4/tests: Eliminate the test_simple_struct_marshal valgrind error.
by Rob Shearman 16 Jun '08

16 Jun '08

2008/6/16 Dan Hipschman <dsh(a)linux.ucla.edu>: > > Eliminates eight of these (and similar) from the ndr_marshall test: > > Conditional jump or move depends on uninitialised value(s) > at 0x4AEF05E: test_simple_struct_marshal (ndr_marshall.c:668) > by 0x4AEFD91: test_simple_struct (ndr_marshall.c:889) > by 0x4AF6B9A: func_ndr_marshall (ndr_marshall.c:1829) > by 0x4B08FF2: run_test (test.h:449) > by 0x4B09332: __wine_spec_exe_entry (exe_entry.c:36) > by 0x7B870810: start_process (process.c:904) > by 0x47EE4B6: (within /usr/local/google/work/wine/libs/wine/libwine.so.1.0) > > The warning comes from memcmp-ing a struct with padding in it that is not > initialized. http://www.winehq.org/pipermail/wine-patches/2008-June/055831.html -- Rob Shearman

2 1

The upcoming 1.0 and announcing Bordeaux for Linux
by Steven Edwards 16 Jun '08

16 Jun '08

Hello All, I mentioned the other day in the noise about the upcoming 1.0 and the short comings I felt it had. Others have commented quite a bit on the subject already and so I'll be pretty brief. I think the 1.0 process while its been slow has been a good even with its shortcomings. Like with everything its a learning process and hopefully we will be able to tweak the process further for 1.x releases. On the positive side, I think that Dan has done a great job as release manager, maintaining the bug lists, working on winetest, staying on top of all the tips and tricks needed to make Valgrind really useful and maintaining winetricks. I'm also really glad that Alexandre did a lot of refinement to the winetest infrastructure, everything he has done with the Wine infrastructure leading up to this point and frankly for everything he's done for Wine for the past 15 years has been awesome. Though I will be glad when the freeze is over so we can get back to having our patches just silently ignored when he does not like something rather than him having an excuse of the code freeze. =) One of the shortcomings I see as I stated in my email the other day, I think we really missed the target on. I think some Office versions should have been a core requirement of 1.0. Even if it was Office 97 and 2000 working in 98 mode I think we should have been able to provide a nearly perfect experience out of the box. Or nearly out of the box as it were, with a little version tweaking. Office is a core pillar of Microsoft's monopoly and not making this a priority I feel misses the point of Wine for a lot of users. Also I think the process can be refined a bit more for subsequent releases. Some of the developers I spoke with offline were unhappy about not having a public experimental branch that was blessed to continue development in. While they could privately do development and push and pull changes with git, not having a sanctioned experimental branch, I think slowed the overall processes of development down and did not really help with bug fixes enough to be warranted. Perhaps in the future we can maintain parallel lines of development before major releases. A good example is the Samba guys, they've been working on Samba 4 for ever, but the stable branch is still being constantly improved, better, newer stuff is being back ported and development continues on both tracks. Finally I am still not really happy with the results of Winetest. I think we should have had dozens of people that could pass all of the tests months or even years before 1.0 shipped rather than just Alexandre's system being the only one that could do a perfect test run. Maybe we should have just asked him to mirror his setup and provide a list of every package and configuration option he uses so we could base target systems on that. Also I think we should have done something to insure new Wine tests always passed on a fixed version of Windows a lot sooner. James has done a lot of work to make all of the tests pass on Windows 2003 but I think we should have done this back in 2003 and said "every new test has to pass on Windows 2003" as well as Alexandre's box before it gets merged. As far as 1.0 goes, its not perfect but its not horrible and I think we've learned quite a bit and Wine is better now than its ever been. I expect 1.1 or 1.2 whatever number Alexandre wants to call it, will be an exponential improvement given the experience we've gained. And lastly I'd like to announce I am starting a new Free Software company with a long term goal of providing Products and Services around Wine and ReactOS. The Bordeaux Group has today released Bordeaux for Linux which builds on the infrastructure laid out in Wine 1.0, Winetricks, IES4Linux and other tools and tries to wrap them all together in a kind of CrossOver Lite type of environment. The ultimate goal is to help refine Wine more and aide in ReactOS development to create a platform that is generic so that users wanting to migrate away from Windows will have a more broad range of choices, be it Wine on Linux and the Mac, ReactOS, or even CrossOver, PlayOnLinux,WineDoors or some other Wine package. Some of the (limited) work we've already done we hope to be able to fold back in to Winehq. If your interested check it out http://www.bordeauxgroup.com Thanks -- Steven Edwards "There is one thing stronger than all the armies in the world, and that is an idea whose time has come." - Victor Hugo

2 2

[legal] eligibility for contributing code to wine for students / graduates of a course derived from the Microsoft Windows Acadamic Program
by Mark Farnell 16 Jun '08

16 Jun '08

If a student is currently taking, or has previously taken a course derived from the Microsoft Windows Academic Program http://www.microsoft.com/resources/sharedsource/windowsacademic/default.mspx which contains the curriculum resource kit, Windows Research Kernel (containing Windows XP x64 amd Server 2003 SP1 kernel sources) and Project OZ, will they become *ineligible* to contribute code to the wine project in the future? Thanks! Mark

7 9

Valgrind error status on the eve of Wine 1.0
by Dan Kegel 16 Jun '08

16 Jun '08

I've added suppressions for gecko and a few other things, added six more patches of my own and eight patches from Hans, Jon Griffiths, Roy Shea, and Zac Brown. As a result, the error count today (not counting memory leaks) is down 33% from last week; we're down to 218 unique-ish valgrind warnings when running the wine test suite! If you've submitted a patch to fix a warning reported by Valgrind, please check to make sure I've included it. If you have a spare hour or two, please grab one of the remaining errors and see if you can come up with a good fix (i.e. one that doesn't just paper over a problem without understanding it). Once we're down to near zero valgrind warnings, it'll be a lot easier for the average wine developer to check their own patches with valgrind before submission. (Why am I ignoring memory leak warnings? Because I had to focus on something first. Once I get those in shape, I'll focus on memory leaks.) Report showing unique nonleak errors: http://kegel.com/wine/valgrind/logs-2008-06-15.counts.txt Logs: http://kegel.com/wine/valgrind/logs-2008-06-15.log.gz http://kegel.com/wine/valgrind/logs-2008-06-15/ Pending wine patches: http://kegel.com/wine/valgrind/logs-2008-06-15-patches/ http://kegel.com/wine/valgrind/logs-2008-06-15-patches.tgz http://kegel.com/wine/valgrind/logs-2008-06-15-patch.txt Suppression file: http://kegel.com/wine/valgrind/valgrind-suppressions

1 0

re: Appdb has no rc5 choice
by Dan Kegel 16 Jun '08

16 Jun '08

I added the version to bugzilla just now, so the appdb should have it too.

1 0

iTunes ipod support
by Ben Engbers 16 Jun '08

16 Jun '08

Hi, A few days ago, I bought an iPod classic. When I connect it to my laptop (running FC7) the iPod shows up on the desktop. I downloaded iTunessetup.exe from the Apple site and it installed under Wine without problems. iTunes looks good, scanned my system and found all the MP3's, but it doesn't detect the iPod. (Amarok recognizes the iPod but every time I try to synchronize, the iPod database gets corrupted and I have to ask my daughter to reinitialize the iPod :-( so I cant't use that either) How I can I figure out why iTunes does not detect the iPod? Ben

2 1

Re: gdi32: stop font tests crashing under win98
by Dmitry Timoshkov 16 Jun '08

16 Jun '08

"Alistair Leslie-Hughes" <leslie_alistair(a)hotmail.com> wrote: > Changelog: > gdi32: stop font tests crashing under win98 > - if (!*font_name && font_charset == DEFAULT_CHARSET) /* do it only once */ > + /* EnumFontFamilies with the NULL parameter crashs win98 */ > + if (pGetFontUnicodeRanges && !*font_name && font_charset == DEFAULT_CHARSET) /* do it only once */ Why you are testing for pGetFontUnicodeRanges here since it has nothing to do with the test, and disables NT4 as well? There should be a better way to avoid Win9x, for instance using unicode APIs in this, and only this case. -- Dmitry.

1 0

Appdb has no rc5 choice
by Leslie Viljoen 16 Jun '08

16 Jun '08

Hi people I am testing some games with Rc5 and wish to submit results in the AppDb but there is no Rc5 choice. Les

3 2