Test PC setup
For testing, we used a 333MHz Quantex Pentium II with a 8GB EIDE hard disk drive. As a test of WinAlign performance improvements, we chose Netscape Navigator 4.05 because it is a common application and it takes a very long time to load. The long load times meant we could make relatively accurate measurements of load times using a stopwatch. Netscape was set to display a blank page when it launched, so no network activity would affect the results.
How the tests were performed
The basic test consisted of the following procedure:
- Boot the system, and wait for boot-time disk activity to subside.
- Time the launch of Netscape Navigator, from the initial click to the time the blank browser screen appeared.
- Exit Netscape, and wait for disk activity to stop.
- Time a second launch of Netscape as in step two.
This test procedure was performed for the following configurations, with both 16MB and 64MB of RAM:
- Windows 95 OEM Service Release 2.0 (OSR2) with a freshly defragmented FAT16 disk.
- Windows 98 immediately after installation and FAT32 conversion.
- Windows 98 after optimization with Disk Defragmenter. Netscape was run 10 times before optimization so that its dynamic load profile would be recorded for use by Disk Defragmenter.
- Windows 98 after aligning the Netscape Navigator DLL and EXE files, then re-optimizing.
For each configuration, we performed the test procedure three times in exactly the same way, then averaged the results and rounded to the nearest half-second.
When compared to Windows 95 OSR2, the combination of disk optimization and aligned program code in Windows 98 resulted in significant reductions in program launch times. In low-memory tests (16MB of system RAM), the overall improvement was 34 percent; high-memory tests (64MB of RAM) improved by 56 percent.
In both low-memory and high-memory cases, aligning the Netscape program code reduced initial launch times by about 20 percent compared to the best case obtained in Windows 98 with unaligned code. As expected, the second launch time was not reduced significantly from the non-aligned case, since the majority of the program code was still in memory from the first launch.
Click here to view the test results of
"Time to launch Netscape Navigator 4.05 in seconds".
Using WinAlign to align program code in Windows 98 can make a measurable reduction in program launch times. Since the process does carry some risks with it (see the FAQ), the best strategy for using WinAlign is probably to align only the few programs that you use frequently, and only those that take a long time to load. Our tests show that Netscape Navigator is certainly one good candidate.
As vendors begin to see the benefit of aligned program code, more of them will ship their applic ations pre-aligned so that it won't be necessary to run WinAlign. In the meantime, judicious use of WinAlign can help to tweak Windows 98 performance.
Return to: The Expert's Guide to Windows 98 WinAlign