https://bugzilla.novell.com/show_bug.cgi?id=342190#c10
--- Comment #10 from Robert Jordan <[EMAIL PROTECTED]> 2007-11-20 06:09:19 MST --- I run some tests on the same machine (Windows XP, Intel Core Duo 2, 2GHz): randomfile.bin ============== MS.NET /ziptest.exe randomfile.bin ICSharpCode.SharpZipLib, Version=2.84.0.0, Culture=neutral, PublicKeyToken=1b03e6acf1164f73 SharpZip :00:00:10.7343750 Zlib :00:00:14.4218750 MONO 1.2.6 on Windows: mono neo:~/src/tests $ mono ./ziptest.exe randomfile.bin ICSharpCode.SharpZipLib, Version=2.84.0.0, Culture=neutral, PublicKeyToken=1b03e6acf1164f73 SharpZip :00:00:16.0940000 Zlib :00:00:08.4680000 logfile.txt =========== MS.NET ICSharpCode.SharpZipLib, Version=2.84.0.0, Culture=neutral, PublicKeyToken=1b03e6acf1164f73 SharpZip :00:00:03.9218750 Zlib :00:00:04.1562500 MONO 1.2.6 on Windows mono neo:~/src/tests $ mono ./ziptest.exe logfile.txt ICSharpCode.SharpZipLib, Version=2.84.0.0, Culture=neutral, PublicKeyToken=1b03e6acf1164f73 SharpZip :00:00:07.8440000 Zlib :00:00:02.2650000 zerofile.bin ============ MS.NET mono neo:~/src/tests $ ./ziptest.exe zerofile.bin ICSharpCode.SharpZipLib, Version=2.84.0.0, Culture=neutral, PublicKeyToken=1b03e6acf1164f73 SharpZip :00:00:04.3593750 Zlib :00:00:03.3750000 MONO 1.2.6 on Windows mono neo:~/src/tests $ mono ./ziptest.exe zerofile.bin ICSharpCode.SharpZipLib, Version=2.84.0.0, Culture=neutral, PublicKeyToken=1b03e6acf1164f73 SharpZip :00:00:12.7190000 Zlib :00:00:01.5470000 The first line is the output of Console.WriteLine (typeof (GZipOutputStream).Assembly); => the tests are using the same SharpZipLib version on both runtimes. * randomfile.bin (100MB) is reproducible, since it was generated with a known seed using mono on the 2.0 profile: using System; using System.IO; class Test { static void Main () { Random rng = new Random (42); byte[] bytes = new byte [1024 * 1024]; Stream stm = File.Create ("randomfile.bin"); for (int i = 0; i < 100; i++) { rng.NextBytes (bytes); stm.Write (bytes, 0, bytes.Length); } stm.Close (); } } * logfile.txt (100MB) was taken from an apache log file. * zerofile.bin is the output of dd if=/dev/zero of=zerofile.bin count=100 bs=1048576 The results show that mono's performance degradation is related to the compression rate of the data. See zerofile.bin, where mono performs worst (relatively to MS.NET). I'm aware that this doesn't help much, but someone who knows the GZip algorithm should be able to deduce which parts of the GZip code is performance-critical when zerofile.bin is processed. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug. You are the assignee for the bug. _______________________________________________ mono-bugs maillist - mono-bugs@lists.ximian.com http://lists.ximian.com/mailman/listinfo/mono-bugs