How do you guys manage these insane clock speeds? I've never even heard of a 486 board running at 60 MHz before I started poking around here. Is that a jumper config or did you change the clock? Was there ever a 486 chip that was factory rated at 60FSB?

Some motherboards have "generic" clock generators that are capable or higher clocks then officially needed. Of course these settings aren't in the manual and have to be found by trial and error. This particular board supports officially up to 50MHz and then also 60 and 66MHz unofficially. Then you need a chip that can "use" this higher clocks. In my case the AMD-133ADZ supports multipliers 3X and 4X. Using lower multi with higher FSB speeds and with good chip you can get really high speeds. This ADZ posts at 200MHz but freezes soon after. At the moment it is stable at 180MHz. But this comes at a cost... cache and memory timings have to be relaxed and are at slowest possible. Still it's a % faster then 4x 40=160MHz with fastest timings.

It's the same story with Intel 486DX4-100WB. You need to find a good chip to get reliable system. This one posts and boots to DOS at 3x 50=150MHz, but freezes soon after - can't run any utility as some garbage comes up 🤣

All this is at 4V - at least the jumpers are set to 4V.

486 officially went to 50MHz FSB. All the 2/3/4X clockers where 33 or 40MHz.

I hope I'll have one 2x66 capable too 😁
Are UMC VLB motherboard 66MHz capable like the PCI versions ?

Impressive results, but your screenshots are impossible to read.

Could you please hold the camera directly in front of the monitor next time?

No idea about VLB boards. Although I doubt it's possible as VLB boards usually don't have the dividers for VL-Bus and I'm sure no card is capable of running at more than 50MHz. Maybe someone has more information.

oerk wrote:

Impressive results, but your screenshots are impossible to read.

Could you please hold the camera directly in front of the monitor next time?

I hold it like this as otherwise you can only see flash 🙁

kixs wrote:

oerk wrote:

Impressive results, but your screenshots are impossible to read.

Could you please hold the camera directly in front of the monitor next time?

I hold it like this as otherwise you can only see flash 🙁

I have no issues with reading the text in any of the "screenshots"! 😀

I think the key is to not use a thingy when browsing Vogons but a computer.

Yeah, I'm reading everything from the thumbnails.

kixs wrote:

oerk wrote:

Impressive results, but your screenshots are impossible to read.

Could you please hold the camera directly in front of the monitor next time?

I hold it like this as otherwise you can only see flash 🙁

In that case, can you turn off the flash? Using a flash light when taking a photo of something self-illuminated doesn't make much sense.

Even better would be screenshots generated with a tool. In the past I have seen a lot of screenshots for DOS-based benchmarks. Recently a tool for screenshots of text screens was mentioned in some discussion.

idspispopd wrote:

kixs wrote:

oerk wrote:

Impressive results, but your screenshots are impossible to read.

Could you please hold the camera directly in front of the monitor next time?

I hold it like this as otherwise you can only see flash 🙁

In that case, can you turn off the flash? Using a flash light when taking a photo of something self-illuminated doesn't make much sense.

Even better would be screenshots generated with a tool. In the past I have seen a lot of screenshots for DOS-based benchmarks. Recently a tool for screenshots of text screens was mentioned in some discussion.

From experience I can say that taking photos of a black screen without a tripod or using the flash results in non usable images 99% of the time, at least when using auto foucus, auto ISO mode and auto shutter time. When there are bright colors on the screen it usally works to take photos without using the flash.

I have a hard time understanding how this even could be a problem as the images posted are fully readable.

To clarify, you found an Am5x86 chip stable at 180 MHz using 4.0 V? What type of cooling are you using? Have you run Windows 9x? Running Windows tends to be a game changer for stability. Have you run some Win9x benchmarks? Adding more L2 cache or RAM will also alter your point of stability. What are your RAM/cache intentions for this machine?

My final optimised values on my IBM 5x86c-133/2x system for Quake 1.06 are 18.7 fps. For Quake 1.08, it is 19.6 fps.

Based on your Speedsys screenshot, your L2 and RAM speeds are about that of an Am5x86-160. At least they are not less. So it seems the main benefit of the 60 MHz FSB is to allow you to run the Am5x86 at 180 MHz. You also will loose out a little on PCI bandwidth though (30 MHz vs. 33.3 MHz).

OK, I'm sorry! Seems like I'm the only one having issues with the screenshots 😀 Fair enough. I can read them though it's kind of annoying, and I AM using a computer. What kind of VOGONS member wouldn't? 😉

Can we please get back on topic now?

idspispopd:
I hold camera in my hand and without flash image is not sharp. I certainly wouldn't use any background programs when running benchmarks.

Tripod would help of course... maybe some other time.

feiopa:
Cooling is just a simple 486 cooler with a fan.

At the moment it's Quake stable. Doom runs... but in demo it just goes the wrong way. I haven't tried Windows on it yet.

Cache is already at the max... 1024KB. Memory is 16MB, quite enough for usual 486 needs. I don't have any plans for it. Just like to play with the hardware 😁

As said in my other comment. Cache/memory timings are at the slowest possible. 4x40MHz combination has everything at the fastest possible, but overall the 180Mhz still edges 160MHz by a few %.

I have a few other 486 boards that I can compare which combo is the fastest.

Cyrix 5x86 is a beast in FPU. My 100MHz would do stable 120, but with 133 is very unstable.

That's funny, in my tests, whenever doom demo doe'snt finish the right way, quake crashes 😁
I've never been able to see Quake more stable than doom.

Me too. I run Quake for system stability. If it finishes a couple of times then it's 99% chance everything will work just fine.

kixs wrote:

No idea about VLB boards. Although I doubt it's possible as VLB boards usually don't have the dividers for VL-Bus and I'm sure no card is capable of running at more than 50MHz. Maybe someone has more information.

Isn't a 66MHz VLB with 1 wait state the same as a 33MHz 0-ws from the point of view of the card ?

I will see what I can do later this week 😀.

I do not own that many interesting CPUs but I do have two AMD 5x86p75 and with luck two more incoming in a Trade. I checked my other Intel DX4-100 but its a write through chip.

My Lucky Star LS486E SIS chipset boards can do 60 and 66 MHz FSB and I have some hopes my UMC chipset board also can do high FSB.

While many SiS 496/497 boards have a PLL which can run at 60 and 66 MHz, I have not come across any SiS 496/497 boards which can do 66 MHz reliably. Running Windows games and benchmarks is a good way to establish stability, much more so than DOS.

feipoa wrote:

While many SiS 496/497 boards have a PLL which can run at 60 and 66 MHz, I have not come across any SiS 496/497 boards which can do 66 MHz reliably. Running Windows games and benchmarks is a good way to establish stability, much more so than DOS.

I do now know just how stable but member rad at least got his IBM 5x86C 100HF running at 2x66 MHz stable enough for some DOS benchmarks using the Lucky Star LS486E REV C2.

If some SIS 496/497 chipset boards are somewhat capable of doing 66 MHz FSB in DOS then perhaps 60 MHz FSB could be fully stable which would make a 5x86@3x60 possible. I have not tried anything else than seeing if my extremely crappy 5x86 which isnt stable at 133 MHz with 3.3V would post at 180 MHz at 4V and it did. This CPU is fully stable at 4x40 MHz @4V although it isnt fully stable at 4x33 MHz with 3.3V.