Performance/Achievements Windows Startup

Performance Project performance.openoffice.org
Quick Navigation Project Wiki Home Project Web Home Project Wiki Category Team Team Members Communication Mailing Lists Meeting Minutes Activities ToDo List
About this template

Overview - current achievements for Windows (cold start up)

This page wants to provide information about current achievements regarding improving the cold start up of OpenOffice.org on Windows operating systems.

General conditions for the tests

• Measurements were done using the Performance Monitor tool from Microsoft.
• To better see the benefits of Microsofts Prefetch feature (introduced for Windows XP) all tests were done with prefetch disabled/enabled.
• All start up tests were done using swriter.exe to start an empty Writer document after a reboot.
• After every new installation of OpenOffice.org Windows made a disk defragmentation.

General work for a better cold start up performance

This part wants to provide performance data for the current general achievements. Means that our latest work concentrate on general ways to speed up start up.

Netbook class (slow CPU, sufficient memory)

Test machine

• Netbook Samsung NC10
• 1,6 GHz Intel Atom N270 (Single core/hyperthreading)
• 1 GB RAM
• Fujitsu MHZ2160BH 160GB, 2,5" - 5400 rpm - SATAII - 8MB cache
• Windows XP Home SP3

Optimization method	Test run 1	Test run 2	Test run 3	Mean
OOO310m9	26,7s	26,9s	26,8s	26,8s (100%)
OOO310m9 (prefetch*)	25,6s	25,8s	25,2s	25,5s (95%)
DEV300m45 (flat+non-rebased**)	18,7s	18,4s	18,6s	18,6s (69%)
DEV300m45 (flat+non-rebased**+prefetch*)	18,4s	18,2s	18,3s	18,3s (68%)

Low-end system (limited memory - recommended memory requirement for OpenOffice.org on Windows)

Test machine

• Notebook
• 1,8 GHz Pentium M (Single core)
• 512 MB RAM
• Toshiba MK6026GAX 60 GB, 2.5" - 5400rpm - UDMA100 - 16MB Cache
• Windows XP Professional SP3

This is a typical notebook system with plenty of software installed by the notebook manufacturer. Windows reports about 205MB of physical memory available after start up. Trashing couldn't be noticed during the cold start up tests, even when using Process Monitor for logging.

Optimization method	Test run 1	Test run 2	Test run 3	Mean
OOO310m9	22,2s	21,6s	22,1s	22,0s (100%)
OOO310m9 (prefetch*)	21,0s	20,7s	21,1s	20,9s (95%)
DEV300m45 (flat+non-rebased**)	16,9s	16,2s	16,5s	16,5s (75%)
DEV300m45 (flat+non-rebased**+prefetch*)	16,4s	16,4s	16,3s	16,4s (74%)

Medium range system (plenty of memory)

Test machine

• Desktop computer
• Opteron 175 2,2Ghz (Dual core)
• 4 GB RAM
• Deskstar 7K250 250GB - 7200rpm - SATA - 8MB Cache
• Windows Vista Ultimate 32-Bit (Super Fetch disabled)

Optimization method	Test run 1	Test run 2	Test run 3	Test run 4	Test run 5	Mean
OOO310m9	16,1s	16,1s	16,4s	16,2s	16,2s	16,2s (100%)
OOO310m9 (prefetch*)
DEV300m45 (flat+non-rebased**)	9,9s	11,5s	11,3s	11,6s	10,7s	11,0s (68%)
DEV300m45 (flat+non-rebased**+prefecth*)

(*) Measurements were done with Windows and Prefetch feature enabled.

(**) Current optimizations for a better cold start up performance. You can find more information here:

• Reorder symbols for libraries
• Start up performance and library placement

Conclusion

The optimizations for a better cold start up performance are clearly visible on a wide range of hardware. The speed up differences are heavily dependent on hard disk speed. The test system with the smallest and slowest hard disk scores the lowest speed up. In contrast the system with slowest CPU but a better hard disk scores a higher performance gain. On these single user systems no negative impacts could be noticed during the tests. Even the system with the smallest amount of system memory behave normally.

Individual work on start up performance

This part describes work on OpenOffice.org which tries to speed up specific details. That also means that only minor performance improvement can be seen.

CWS fwk108

This CWS starts the spellchecker libraries on-demand and not when the language status bar control is created. Additionally a fix for the SfxVoidItem has been made in svtools.

Test machine

• Notebook
• 1,8 GHz Pentium M (Single core)
• 512 MB RAM
• Toshiba MK6026GAX 60 GB, 2.5" - 5400rpm - UDMA100 - 16MB Cache
• Windows XP Professional SP3

This is a typical notebook system with plenty of software installed by the notebook manufacturer. Windows reports about 205MB of physical memory available after start up. Trashing couldn't be noticed during the cold start up tests, even when using Process Monitor for logging.

Version	Test run 1	Test run 2	Test run 3	Test run 4	Test run 5	Mean
OOO310m9	22,2s	21,6s	22,1s	---	---	22,0s (100%)
DEV300m45	20,5s	20,1s	20,0s	20,3s	20,2s	20,2s (92%)
DEV300m45 (fwk108)	20,4s	19,7s	19,6s	19,5s	19,8s	19,8s (90%)

Overview performance of MWS builds OOO310/DEV300

Low-end system (limited memory - recommended memory requirement for OpenOffice.org on Windows)

Test machine

• Notebook
• 1,8 GHz Pentium M (Single core)
• 512 MB RAM
• Toshiba MK6026GAX 60 GB, 2.5" - 5400rpm - UDMA100 - 16MB Cache
• Windows XP Professional SP3

This is a typical notebook system with plenty of software installed by the notebook manufacturer. Windows reports about 205MB of physical memory available after start up. Trashing couldn't be noticed during the cold start up tests, even when using Process Monitor for logging.

Version	Test run 1	Test run 2	Test run 3	Test run 4	Test run 5	Mean
OOO310m9	22,2s	21,6s	22,1s	---	---	22,0s (100%)
DEV300m45	20,2s	20,1s	20,4s	20,2s	20,1s	20,2s (92%)
DEV300m46	20,5s	20,1s	20,0s	20,3s	20,2s	20,2s (92%)

Medium range system (plenty of memory)

Test machine

• Desktop computer
• Opteron 175 2,2Ghz (Dual core)
• 4 GB RAM
• Deskstar 7K250 250GB - 7200rpm - SATA - 8MB Cache
• Windows Vista Ultimate SP1 32-Bit (Super Fetch disabled)

Version	Test run 1	Test run 2	Test run 3	Test run 4	Test run 5	Mean
OOO310m9	16,1s	16,1s	16,4s	16,2s	16,2s	16,2s (100%)
DEV300m45	15,8s	15,8s	14,9s	15,9s	14,8s	15,4s (95%)
DEV300m46	15,2s	15,2s	15,5s	15,1s	15,8s	15,4s (95%)
DEV300m47	15,7s	15,5s	15,7s	15,4s	15,3s	15,5s (96%)
DEV300m59	10,4s	10,9s	10,4s	10,7s	10,8s	10,6s (65%)