Thanks to a combination of bad journalism and ineffective research on the part of various people, a number of so-called "performance tweaks" for Windows 2000 and XP have been circulated. Some of them were functional at one time but have since been removed in more recent Service Packs, and replaced with more intelligent core logic that tunes the system's memory demands more effectively than manual tweaking.
One of the most infamous of these tweaks is the IoPageLockLimit value, found in HKEY_LOCAL_MACHINE//System//CurrentControl//SetControlSession //ManagerMemory Management. Allegedly, setting this DWORD registry value would allow the user to determine the amount of memory locked by the kernel for I/O operations. On systems with a great deal of memory, such as 512 MB or more, increasing this value would supposedly lead to better performance, since memory was being transferred in bigger batches.
The IoPageLockLimit value was valid in NT 4.0, and in the RTM version of Windows 2000. However, as of Service Pack 1 and all subsequent versions of Windows (i.e., Windows XP, Windows 2003 Server, etc.), all references to this setting were deleted and control over the page-locking function was made totally automatic. Setting this option in anything other than the RTM version of Windows 2000 and all versions of NT 4.0 has no effect. There is no harm in putting this tweak into place, but no benefit, either.
Another wholly mythic tweak is the IRQ8Priority value, found in HKEY_LOCAL_MACHINE//System//CurrentControl//SetControl//PriorityControl. The concept behind this tweak is that by raising the priority of the system CMOS clock the overall performance of the system will improve. This is spurious reasoning at best, since the only two higher priority interrupts (in x86 systems) than the system CMOS clock are the system timer and the keyboard. This setting has been proven to have no effect either. (Another related and equally useless trick involves adding values such as IRQ3Priority, IRQ10Priority, and so on—none of which have ever been substantiated to work.)
About the Author
Serdar Yegulalp is the editor of the Windows 2000 Power Users Newsletter. Check out his Windows 2000 blog for his latest advice and musings on the world of Windows network administrators – please share your thoughts as well!
This was first published in December 2003