They have the package number 25544, right?
I observed that all chips with this number have 3x/4x multipliers.
There are some later Am486DX4-100SV8B with the 3x/4x multipliers. These chips also have the 25544 number. Contrary to the labeling, they have 16 kB Cache.

There is a DX2 66 16BGC of some sort. I recommend that one if you want to try 2x66. It should be the 2X 3X multiplier version you are looking for. Basically the very last DX2, DX4 and DX5 chips were all the same chip.

Anonymous Coward wrote:

There is a DX2 66 16BGC of some sort. I recommend that one if you want to try 2x66. It should be the 2X 3X multiplier version you are looking for. Basically the very last DX2, DX4 and DX5 chips were all the same chip.

How can this be if they only have 2 multipliers at the most?

BastlerMike wrote:

They have the package number 25544, right?
I observed that all chips with this number have 3x/4x multipliers.
There are some later Am486DX4-100SV8B with the 3x/4x multipliers. These chips also have the 25544 number. Contrary to the labeling, they have 16 kB Cache.

I seem to recall this once. I once tested an 8kb chip and it seemed to have 16kb. But iirc this was a DX2-80. I just assumed it was just a strange remarked chip

This AM486 DX4-100V16B doesn't have 25544, it has 25372
http://www.chipdb.org/img-amd-a80486dx4-100sv16b-247.htm

And how can this 25544-marked DX2-66GBC possibly have a 3x/4x? Was anyone running with a 22 MHz bus?
http://www.chipdb.org/img-amd-am486dx2-66v16b … graved-5391.htm

Can anyone test these chips? I don't think a marking of 25544 tells the whole story.

On another thought, I wonder if any tricks can be done to get an AMD X5 3/4x to run at 2x?

I can only find mention of a utility to control WT/WB cache.

As far as I can remember all 3V 486DX2 chips (except overdrive) could be configured as 2X/3X. Probably because they were just badge engineered DX4s. Though I am curious about the am486DX2 16BGC, because in theory it should have been a remarked DX5 since no true DX4 was being produced.

I got a V16BGC DX2 CPU 😀

No need for heatsink, stays cool

Beautiful condition. NOS?

Yes, had no time to try it 😀

It should run at 2x66 or 2x83 MHz if you're into that sort of thing.

I was just testing today one of those amd 486dx4 100Mhz cpus with 4x multiplier. One of those that have been marked to have 8Kb of L1 cache but actually have 16Kb.
Anyone knows if I can damage the cpu if I keep it at 4x multiplier instead of 3x long term?

No i had DX4-100 @ 5x86 160Mhz running for at least 2 years until i sold the system.

feiopa, i have currently no board which supports 66/83 which works. i began to rework the Biostar UB8433 but still had no time to investigate what BIOS Chip is used 😀 also missing the content..
on the board it says 271024, but other sources tell me 010 ?!

I've tested these and they work on the MB-8433UUD: PH29EE010, W29EE011, AT29C010A, and P28F001. The P28F001 is 12 V programming, while the others are 5 V.

matze79 wrote on 2020-02-11, 09:43:

No i had DX4-100 @ 5x86 160Mhz running for at least 2 years until i sold the system.

feiopa, i have currently no board which supports 66/83 which works. i began to rework the Biostar UB8433 but still had no time to investigate what BIOS Chip is used 😀 also missing the content..
on the board it says 271024, but other sources tell me 010 ?!

Do you need to increase voltage to run it at 160Mhz? There is an option for 4V on the motherboard.

No, it runs Stock 3.45V without Trouble.
Its same Core as AM5x86, for satisfying the market. Why produce several different versions if one is enough 😀

Thanks i will try it 😀

Baoran wrote on 2020-02-11, 09:39:

I was just testing today one of those amd 486dx4 100Mhz cpus with 4x multiplier. One of those that have been marked to have 8Kb of L1 cache but actually have 16Kb.
Anyone knows if I can damage the cpu if I keep it at 4x multiplier instead of 3x long term?

Same thing here. I tested my DX4s in speedsys when I noticed one of them had twice the cache advertised. So I configured it as a 5x86 and I saw "am5x86 133MHz" 😁.

AMD AM486DX4-100SV8B and AM486DX2-66: Interesting observation / CPU relabeling

Deksor wrote on 2020-02-11, 23:32:

Baoran wrote on 2020-02-11, 09:39:

I was just testing today one of those amd 486dx4 100Mhz cpus with 4x multiplier. One of those that have been marked to have 8Kb of L1 cache but actually have 16Kb.
Anyone knows if I can damage the cpu if I keep it at 4x multiplier instead of 3x long term?

Same thing here. I tested my DX4s in speedsys when I noticed one of them had twice the cache advertised. So I configured it as a 5x86 and I saw "am5x86 133MHz" 😁.

I just noticed today that when it is configured to am5x86 133Mhz with motherboard jumpers and I turn turbo off the system freezes and when booting turbo off it gets stuck at himem.sys testing memory. When turbo is on, everything works fine.

In other words, it's not stable at 133Mhz.

The Serpent Rider wrote on 2020-02-12, 09:44:

In other words, it's not stable at 133Mhz.

It is stable at 133Mhz. It is when you slow it down using by turning off turbo when there seems to be a problem. I haven't had a single freeze or other kind of problem when it is at full speed.
Not that I would use this kind of computer for speed sensitive games, so I don't need to slow it down. I guess I could try to set the cpu selection jumpers to dx4 and just change the multiplier jumper from that to make it 133Mhz instead of setting the jumpers to 5x86.