- Joined
- Nov 29, 2001
- Location
- NRW, Germany
Always wished that an an overclocking BIOS à la K7S5A would surface for this pretty exotic mobo -- but since nothing is available and interest in the board is rather low, I started messing around a bit with AMIBCP and CMOS RAM contents.
1. The facts:
The P6S5AT comes with a PLL that can handle FSBs at > 133 MHz and that with sensible PCI dividers. What is used is an ICS 9248-146, datasheet available
here .
When examined with AMIBCP, all three available
BIOS versions show values for FSBs > 133 in CPU setup. You just can't select them, i.e. the interesting FSB speeds only show up in the "result" field.
2. CMOS experiments
Since I couldn't do much with AMIBCP, I decided to save the CMOS RAM contents to disk for every FSB/MEM combination the P6S5AT BIOS has to offer. This is a pretty good DOC describing the procedure for the K7S5A.
Turns out that byte 6F in CMOS stores the FSB settings. Problem is that I can't figure out the relation to the PLL programming bits (blame that on new year's eve -- happy new year to everyone, BTW
).
3. (very) limited early result
For 133 FSB / 133 MEM, CMOS byte 6F has a value of 36 (hex). If you change that to 3F (hex), the board will be on a 150 MHz FSB.
Problem is that I can't boot into windows at that speed -- this is with a Tualtin Celeron 1.0A, 1.65 VCore and Infineon SDRAM that happily does 133 at 2T latency and "Ultra" settings. I'm pretty sure the 1500 MHz is not just some random value the BIOS routine prints, since I can boot to DOS no problem and run the CTCM bench which passes no probs and confirms the 1.5 GHz speed.
4. Call for assitance
If you are feeling adventurous, check out if your components can handle 150 MHz FSB. Perhaps we are facing a limitation of the chipset or the board, or (hopefully) it's just my kit that isn't up to the task.
5. Disclaimer
If you manage to change your FSB to > 133 MHz with this sketchy howto, I hope you'll also manage to blame it on yourself if anything catastrophic happens ;-).
Cheers
,
DSTA
1. The facts:
The P6S5AT comes with a PLL that can handle FSBs at > 133 MHz and that with sensible PCI dividers. What is used is an ICS 9248-146, datasheet available
here .
When examined with AMIBCP, all three available
BIOS versions show values for FSBs > 133 in CPU setup. You just can't select them, i.e. the interesting FSB speeds only show up in the "result" field.
2. CMOS experiments
Since I couldn't do much with AMIBCP, I decided to save the CMOS RAM contents to disk for every FSB/MEM combination the P6S5AT BIOS has to offer. This is a pretty good DOC describing the procedure for the K7S5A.
Turns out that byte 6F in CMOS stores the FSB settings. Problem is that I can't figure out the relation to the PLL programming bits (blame that on new year's eve -- happy new year to everyone, BTW
).3. (very) limited early result
For 133 FSB / 133 MEM, CMOS byte 6F has a value of 36 (hex). If you change that to 3F (hex), the board will be on a 150 MHz FSB.
Problem is that I can't boot into windows at that speed -- this is with a Tualtin Celeron 1.0A, 1.65 VCore and Infineon SDRAM that happily does 133 at 2T latency and "Ultra" settings. I'm pretty sure the 1500 MHz is not just some random value the BIOS routine prints, since I can boot to DOS no problem and run the CTCM bench which passes no probs and confirms the 1.5 GHz speed.
4. Call for assitance
If you are feeling adventurous, check out if your components can handle 150 MHz FSB. Perhaps we are facing a limitation of the chipset or the board, or (hopefully) it's just my kit that isn't up to the task.
5. Disclaimer
If you manage to change your FSB to > 133 MHz with this sketchy howto, I hope you'll also manage to blame it on yourself if anything catastrophic happens ;-).
Cheers
, DSTA
