Do 133 MHz-marked Am5x86 chips have the same issue de-turboing on this motherboard, or is it just this chip?

Some mobos disable the cache in de-turbo mode, chances are this is not fully compatible with write-back scheme.

feipoa wrote on 2020-02-12, 11:40:

Do 133 MHz-marked Am5x86 chips have the same issue de-turboing on this motherboard, or is it just this chip?

I don't have another amd chip to test it with. Kingston turbochip 133Mhz does de-turbo fine but it doesn't support write back mode.

I once owned an AMD 486DX4-120 SV8B (L1 write back cache) chip.
On a Gigabyte GA-5486AL motherboard, the de-turbo button also didn't work. With the NV8T variant (L1 write thru cache), the de-turbo function works fine.
I believe I also achieved similar results with my AMD 5x86 CPU as that of my SV8B variant(will check some of my benchmarks later).

Although,with an AMD 486DX4-100, the decrease in speed with the turbo button probably gets you around that of a 386DX-40.

jesolo wrote on 2020-02-12, 18:18:

I once owned an AMD 486DX4-120 SV8B (L1 write back cache) chip. On a Gigabyte GA-5486AL motherboard, the de-turbo button also d […]
Show full quote

I once owned an AMD 486DX4-120 SV8B (L1 write back cache) chip.
On a Gigabyte GA-5486AL motherboard, the de-turbo button also didn't work. With the NV8T variant (L1 write thru cache), the de-turbo function works fine.
I believe I also achieved similar results with my AMD 5x86 CPU as that of my SV8B variant(will check some of my benchmarks later).

Although,with an AMD 486DX4-100, the decrease in speed with the turbo button probably gets you around that of a 386DX-40.

For me with 486 the slowdown with the turbo button seems to always depend on the motherboard. I have seen slowdowns of between just 25% slowdown and even close to 75% slowdown. It feels like the older the motherboard is, the more it slows down. At least that is how it has been with the motherboards I have and the newest 486 motherboard from 1996 slows down the least.

I also have a DX2 System who freezes when i press turbo button.. some hardware and drivers simply don`t like the frequency change

matze79 wrote on 2020-02-13, 07:59:

I also have a DX2 System who freezes when i press turbo button.. some hardware and drivers simply don`t like the frequency change

I did bit of testing. I changed the cpu type jumpers on my motherboard from amd 5x86-p75 to normal amd dx4. When the jumpers have been set to amd dx4, turbo does absolutely nothing. The cpu is at full speed no matter if turbo is on or off. I also tried the cpu multiplier jumper at both 3x and 4x. No freezes like this but no slow down either so I have no idea what the motherboard is doing. It slowed intel 33Mhz and 66Mhz cpus just fine when I tested the motherboard with those before.
Anyone has any ideas of what is going on?

DX2 133Mhz 😀
2x66Mhz

PCI 1: 1/2

But it is not stable, not going to detect harddisk / freeze.

3x50Mhz detects hdd but does not load any data from harddisk
2x50Mhz does boot but has problems reading files correctly
3x40Mhz 120Mhz is stable.

EDIT: Cyrix 5x86 2x50Mhz also not booting, Stuck in BIOS Screen / Detection / Slowed Down Cache not helping

Cyrix 5x86 2x50Mhz - BIOS Auto Setting / Booting - 3WS DRAM.. should find faster RAM iguess.

Video Card is Viper 330. As PCI is running 33/25Mhz Speed it should not be a issue.
The CPU does not even get hot, its just warm!

Did not yet mess around with Settings, this is just quick try.

The System still seems to has other Troubles.,
I inserted 2x 32Mb FPM and it reconizes only 48Mb RAM.
So one Bank is missing, Other RAM same Fun.
I guess this is mechanical Problem i will replace the Banks also..
This Board was heavy abused and i`m glad it works at all.
With 1x 32Mb its stable with normal Settings.

Woho 2x66Mhz booting 486DX2 2x66Mhz without L2 Cache, 3WS RAM

Running DOOM Benchmark!

2134 / 1819 realtics

Some progress, so CPU is stable!

2ws / cache 3-2-2-2

Try to stick with a single stick of 60 ns FPM memory and you might be able to get away with 1 ws of memory read and 0 ws on memory write. 3-2-2-2 for cache is expected. What are you running your IBC DEVSEL at? From memory, I think that is what that BIOS setting is called. For 66 MHz FSBs, it should be on SLOW - not fast, not medium.

Slow, see BIOS Shot on the Image.

I can get it POST with 2x83Mhz but it won`t boot.
3x60Mhz - 180Mhz same, goes until Drive Detection then it woes.
Maybe a slight Voltage increase would get it running at 3x60.

Possibly. Try 4 V, but I doubt it would be stable enough to boot Windows. That's my experience anyway.

Is it worth to Upgrade to 1Mb Cache ?
FPM RAM did not help with the Waitstates.

Quake is only stable with highest waitstates, but ram write can be 0WS, only reads need WS.

The 1024K mod is fairly simple. But I did make it so my MB-8433UUD is jumpered for removable 512K and 1024K double-banked, but I haven't modified the manual for this condition. It is quite a bit more work, but the result is more versatile.

With the RAM read at 3 ws, 1024k should help. But it would be nice to figure out why you can't run at 1 WS with FPM. I run a Cyrix at 2x66 on this board at 1 WS read. Have you tried different RAM sticks? Have you tried a Cyrix at 2x66?

What voltage are you using to the CPU? Try 3.6 V. Some AMD's I've tried don't handle 3.45 V well on some boards and 3.6 V solved the issue, though that was at 160 MHz.

I using 3.45V, but i think something else is wrong.

As i also experience RAM Problems with multiple RAM Modules and 33Mhz Bus and CPU @ Stock Clock.
This was to be expected the lower 2 RAM Slots are not holding any Module..
I got used PS/2 Slots and will try to replace them today!

I have several of these boards and one such board has an issue with the first SIMM slot, but if I use the other slots, no problem. Perhaps try changing out the SIMM socket?