Windows 8 testing on AMD
Now lets do a fresh Windows 8 install on an AMD platform and see what happens:
Everything seems to be fine here obviously. No bus frequency changes were made, all timers have the right frequency and time.
We rebooted the system and downclocked the bus frequency from 200 to 180 in Windows 8 using an overclocking tool.
As you can see downclocking didn't affect the RTC timer nor the QPC timer on AMD, everything is fine. But, if you look closer you see the QPC timer has changed since, now at 14.32 MHz (same frequency as the HPET, meaning it uses the HPET as source), previously at 4.12 MHz (DMI?).
We will come back on this strange change on a later stage.
Like with Intel, we booted at lower bus frequency from the BIOS to see what happens:
All timers display the same time and the same frequencies as before, everything normal.
As the QPC seems to use the HPET timer (because of the same frequency) we wanted to see what happens if you disable the latter in the BIOS:
Obviously the HPET timer can't be read anymore so the tool doesn't display anything for the HPET Timer.
Furthermore as the QPC Timer is unable to use the HPET it changed source and is now using the ACPI Timer
which can be seen thanks to the displayed frequency which is the same as the ACPI Timer (3.58 MHz).
Okay now what happens when you enable HPET in BIOS again?
The QPC Timer frequency changed to 14.32 MHz again meaning it makes use of the HPET again.
Before further testing, here is a resume of what we learned from the AMD testing:
- Right after finishing the installation of Windows 8 on an AMD platform, the QPC seem to use the same kind of timer as on Intel. After reboot it uses the HPET if available or ACPI.
- By default, on AMD/Windows 8, the QPC use the HPET as timer source if this one is available, ACPI otherwise.
- RTC and QPC Timers use the same hardware timer source?
- AMD is not affected by the "Windows 8 RTC Bug"