continuation of previous post

Here are the Speedsys images. I can clearly see that the L2 cache and memory are also slower with the Pentium Overdrive 83MHz than with the 486 DX2 66MHz. The CPU test is much faster with the Pentium Overdrive 83MHz. The video memory speed is also slower with the Pentium Overdrive 83MHz than with the 486 DX2 66MHz.

Pentium Overdrive 83MHZ

486 DX 2 66MHZ

these are my bios settings.

The Pentium Overdrive 83 MHZ performs slower than the 486 DX 2 66MHZ in all tests, except for Quake, where it gets 2 frames more

What else can I do to get the speed of the L2 cache, memory, and video memory to match the speed of the 486DX2 66MHz? Are there perhaps incorrect BIOS settings, or have I misplaced my jumpers? The jumper settings are set for a 486DX2 66MHz and I have a 33,33 MHz oscilator installed. Does anyone know if there's anything I can do about this, or is it because this motherboard doesn't support Pentium Overdrive 83MHz better?

I measured the speed with a meter in the ISA slot with the Pentium Overdrive 83MHz and with the 486 DX2 66MHz. See the pictures.

With the Pentium Overdive 83MHZ

with the 486DX2 66MHZ

It seems like the Pentium Overdive 83MHz significantly reduces the speed to the ISA slot. I have no idea why. Could I have configured something incorrectly in the BIOS or with the jumpers?

That's weird, I wonder if it's setting the ISA clock divider by some dumb timing loop that thinks it must have a 40 or 50mhz bus CPU in there to get the same result it got from the POD.

Edit: yah numbers seem close enough that it has switched from a 1:4 divider to a 1:6

This is a case where opening up all the hidden bios parameters is helpful.

Many old ibm bios operated in a stupid way where it only looked at raw clock speed to “automatically” set FSB parameters .

It’s probably slowing cache and memory as well using unnecessarily slow chipset timings.

rmay635703 wrote on 2026-03-01, 15:22:

This is a case where opening up all the hidden bios parameters is helpful.

Many old ibm bios operated in a stupid way where it only looked at raw clock speed to “automatically” set FSB parameters .

It’s probably slowing cache and memory as well using unnecessarily slow chipset timings.

Can you let me know how I can open hidden bios parameters?

egbertjan wrote on 2026-03-01, 15:50:

rmay635703 wrote on 2026-03-01, 15:22:

This is a case where opening up all the hidden bios parameters is helpful.

Many old ibm bios operated in a stupid way where it only looked at raw clock speed to “automatically” set FSB parameters .

It’s probably slowing cache and memory as well using unnecessarily slow chipset timings.

Can you let me know how I can open hidden bios parameters?

You'll want to have a look at a tool called modbin, see here.

egbertjan wrote on 2026-03-01, 15:50:

rmay635703 wrote on 2026-03-01, 15:22:

This is a case where opening up all the hidden bios parameters is helpful.

Many old ibm bios operated in a stupid way where it only looked at raw clock speed to “automatically” set FSB parameters .

It’s probably slowing cache and memory as well using unnecessarily slow chipset timings.

Can you let me know how I can open hidden bios parameters?

Hi egbertjan,

I agree with rmay635703 that enabling hidden BIOS options for ISA-bus clock would be the best solution. For your v2.06k Award v4.50 BIOS, the MODBIN tool from Award Software is the way to do this.

I used MOBIN on your v2.06k BIOS, but unfortunately it doesn’t have any hidden options for ISA clock control.
Looking in the disassembly listing I have of this BIOS, I only see automatic ISA clock control depending on the detected FSB speed.
As Bitwrangler suspected, the FSB speed detection must be confused by the POD 83 with its x2.5 multiplier and think it is running at more than 40MHz FSB. This would cause the BIOS to set a 1:6 divider for 50MHz FSB.

It will take me some time, but I will look in the BIOS code to see if I can fix this.

Groetjes, Jan

Hi egbertjan,

I believe I’ve found an alternate solution to your POD83/ISA-bus clock issue.

It appears that the old DOS chipset control program CTCHIPZ has a VT82C495.CFG file for the VIA chipset on the FIC 4386-VIO board. This VT82C495.CFG file comes complete with Macros for setting the ISA-bus clock directly.
Here is a copy:

From the DOS prompt (or from AUTOEXEC.BAT), use this command to set the FSB/ISA clock divider to 1:4

CTCHIPZ VT82C495 /BUSCLK4

The VT82C495.CFG file also has Macros for other dividers:
BUSCLK2 -> 1:2 for 16MHz FSB
BUSCLK3 -> 1:3 for 25MHz FSB
BUSCLK4 -> 1:4 for 33MHz FSB
BUSCLK5 -> 1:5 for 40MHz FSB
BUSCLK6 -> 1:6 for 50MHz FSB
BUSCLK8 -> 1:8 (not usable on the 4386-VIO)
BUSCLK10 -> 1:10 (not usable on the 4386-VIO)
BUSCLK12 -> 1:12 (not usable on the 4386-VIO)

Let us know if this solution helps with your issue.

Greetings, Jan

Thank you Jan, right now I am rather buzy, I'll be back in a couple of days on this issue

Great find Jan!

I gotta remember that util for stupid OEM boards.

It is possible that the system will work with a 1:3, depending usually on how fussy network and sound cards are, they are usually the problem children above 10mhz, though it's usually over 12mhz that things get more difficult. The only time it's not those two suspects though is if you've got ancient VGA or I/O cards.