You will need more than 10 chips, then keep swapping them until you find a set that does not hung the system.
But first of all - did you confirm that disabling fhe l2 cache gets you in dos properly ?

pshipkov wrote on 2022-12-02, 16:09:

You will need more than 10 chips, then keep swapping them until you find a set that does not hung the system.
But first of all - did you confirm that disabling fhe l2 cache gets you in dos properly ?

Yes without l2 cache it works fine

Buy a bunch of 15ns modules and start binning.

@cachoalpuy
Your question here prompted me to share some data i had about this board.
Take a look. Some of the notes there may be useful to you.
DTK PKM-0033S E-0

pshipkov wrote on 2022-12-03, 03:27:

@cachoalpuy Your question here prompted me to share some data i had about this board. Take a look. Some of the notes there may b […]
Show full quote

@cachoalpuy
Your question here prompted me to share some data i had about this board.
Take a look. Some of the notes there may be useful to you.
DTK PKM-0033S E-0

Thanks, I was looking for this kind of information.
Luckily your board was saved before battery corrosion. Mine broke -+ 12v rails as well as power good and serial ports. Fixed all except serial ports. Could you provide a high resolution photo of the area around the battery?

If you click on the low-res image that will lead you to a high-res one.
If the high-res one is still not enough i can snap a battery only close-up.

CachoAlpuy wrote on 2022-12-01, 23:37:

I have a pkm033s and I can't get it to work with 60ns ram with fastest setting in the bios. It always hangs or gives errors when entering dos. What could be the problem?

The ns delay is just a rough measure.

A 486 needs ok sticks to work above 40mhz and really good ones beyond 50. One also has to take into consideration the chipset itself and if it, given perfect s/rams, would work at certain frequencies. One of the LuckyStar boards(SiS chipset) I have only works beyond 50mhz if it is cooled - just a few degrees made all the difference. On the other hand the 8810(UMC chipset) I have does not work with 50mhz fsb and the best ram timings no matter what module I use.

50mhz fsb with best settings is the sweet spot for 486 performance that can be achieved without enormous hassle. A HOT433 with these settings and 1024KiB cache can complete Doom timedemo3 in a hair over 1000 realtics which roughly translates to 74fps(finding larger srams which would work @50mhz and best timings was a tedious process).
For comparison - around 77fps(970 rtics) is the best I got with a Luckystar board @60mhz with selected SOJ packaged 128KiB of cache(thx to Tiido!). Just a few tics gained but a LOT of hassle.

Got 14,77 3DMarks in 3DMark 99 Max under Win2000.
(Am5x86 @ 160 with Matrix G450 @ 40 MHz PCI + HOT433 with 256 MB RAM and 1024 K Cache)

Re: 3dmark99 MegaThread

Thought this belonged here-

I got a block of dry ice in my latest shipment of frozen coffee. Decided to set up my 200mhz Lucky Star setup and see if 240mhz was possible.

Using the dry ice, once I got the processor to ~-40C (haha) and the system turned on and almost started booting. I even tried turning down the memory timing, turning off the cache, removing extra ram sticks. Normally, this chip needs 10C or less to run reliably at 200mhz.

I imagine with a truly unicorn AM5x86 240mhz benchmarking could be possible!

You gonna hold it there, or stick the lump back in the freezer, run down to the nearest thrift and buy a copper milk jug and a foam beer coozie and grate a bit off at a time into one of those special slushie mixes in that.

Now that's cool.

The Serpent Rider wrote on 2023-03-21, 05:04:

Now that's cool.

Literally! 😂

Very cool JonF.
Dry ice. Now that is a tangential approach.

Don't want to steal the thunder of your post, just adding a relevant note since you brought the point.
I can show the same POST screen with very-selected-cpu + 12V Peltier. POST never completed, didn't explore that properly.
Same for 125MHz P24T. That one i checked a bit more - it is extremely unstable - never succeeds to boot into DOS.

What was the peltier temperature?

~ -10 C

pshipkov wrote on 2023-03-21, 06:24:

Very cool JonF. Dry ice. Now that is a tangential approach. […]
Show full quote

Very cool JonF.
Dry ice. Now that is a tangential approach.

Don't want to steal the thunder of your post, just adding a relevant note since you brought the point.
I can show the same POST screen with very-selected-cpu + 12V Peltier. POST never completed, didn't explore that properly.
Same for 125MHz P24T. That one i checked a bit more - it is extremely unstable - never succeeds to boot into DOS.

I think your AM5x86 probably has a better die than mine if you only needed -10C. I ran mine at 5V too which isn’t ideal. Just shows it is possible though. Next ice shipment maybe with a 3D printed ice tray more testing could be possible.

-40 is achievable with phase change. 1/2 hp compressor with some r1270 would probably do it in single stage on a chip like this. The actual evaporation temp can approach -100f if you pull the evap far enough into a vacuum. (Yes it will still exchange heat at such a low vapor pressure))

I was experimenting with one of my AMD-X5-133ADZ's and figured out that in Doom, Quake and some other benchmarks 150MHz (50 x 3) config works approximately 4-5% faster than 160MHz (40 x 4). Even with halved PCI bus clock.
Some of installed PCI cards can't stand 50MHz bus and crash while booting in Win'98. I'm gonna figure out which one exactly later this week.
Is there any explanation for this?

Test setup: AMD-X5-133ADZ, LS-486e Rev.D, 256kB L2 15ns, 2x32MB EDO 60ns, CL-GD5446. 0WS, cache timings turned down to "Faster" (2-2-2-2).

froller wrote on 2024-03-04, 09:18:

Test setup: AMD-X5-133ADZ, LS-486e Rev.D, 256kB L2 15ns, 2x32MB EDO 60ns, CL-GD5446. 0WS, cache timings turned down to "Faster" (2-2-2-2).

Are you using identical BIOS settings for both configurations? If not, could you specify?

feipoa wrote on 2024-03-04, 11:29:

Are you using identical BIOS settings for both configurations? If not, could you specify?

Of course I use same settings.
All memory timings set to minimal values except "RAS to MA delay" set to 3 CCLK instead of 2 and "DRAM speed" set to "Faster" instead of "Fastest".