A7S333 question

  • Thread starter Thread starter Robert Jackson Marley
  • Start date Start date
R

Robert Jackson Marley

Hi,

The board is supposed to support 333 chips...hence the name of the mobo....
however it does not. i have a amd xp 2600 (12.5x166) and it only does
12.5x133....overclocking it to 166 which is possible puts the pci bus out of
whack which causes the pc to not function properly...are there any bios
fixes which will allow it to run at correct speed? no bios fix is available
on asus' site

tia

rob
 
"Robert Jackson said:
Hi,

The board is supposed to support 333 chips...hence the name of the mobo....
however it does not. i have a amd xp 2600 (12.5x166) and it only does
12.5x133....overclocking it to 166 which is possible puts the pci bus out of
whack which causes the pc to not function properly...are there any bios
fixes which will allow it to run at correct speed? no bios fix is available
on asus' site

tia

rob

The manual on page 12 says this:

"SiS745 Chipset: supports AGP 4X/2X mode, 133/100MHz Front Side Bus,
and the fastest 333/266/200MHz memory bus."

so the memory bus has the honour of the 333 in the name.

Everything listed here, as far as I know, goes no faster than FSB266.
Notice how they take care to point out the FSB266 rating needed.

http://www.asus.com.tw/support/cpusupport/cpusupport.aspx
Athlon XP 2600+(266FSB)(Model 8)(Thoroughbred) ALL 1004

Ref: Two processors are rated 2600+, one is FSB266, one is FSB333
The 2600+/266 is hard to find.
http://www.qdi.nl/support/CPUQDISocketA.htm

As for your problem, there are two ways of doing clocking. The
asynchronous way, where the AGP/PCI clock is delivered separate
by the clock gen (referred to as an "AGP/PCI lock"). Or, the older,
synchronous way, where the FSB and the AGP/PCI are related by
simple integer ratios.

As you've noticed, a 166 clock really needs a 1/5th divider, to
keep the PCI bus happy. That would be a 33MHz PCI clock, with
166MHz fed to the processor. Instead, it is probably maxed at a
1/4th divider, and at 166MHz, this means 41.5MHz for PCI. That
clock speed is sufficient to upset the PCI and perhaps even
corrupt an IDE connected disk. The fastest I would try, is
4 x 37.5 = 150MHz, as 37.5 is a reasonable limit for PCI.

The features of a clock gen cannot be changed after the fact.
If all possible operating modes weren't added to the BIOS
within the first few BIOS releases, it is unlikely a miracle
is going to happen.

For other information, ICS clock generator info is at icst.com
and there are Windows utilities for changing the settings,
like clockgen, setfsb, softfsb etc. These programs have to be
designed for individual combinations of SMbus addresses and
the clock generator chips themselves, so only a limited selection
of boards are supported this way. Again, if the needed divider
is not inside the clock generator chip, then a Windows utility
cannot fix that. The utility would only be useful if the 1/5th
divider was available, and for some strange reason the BIOS
didn't actually program it.

HTH,
Paul
 
Back
Top