I'm assuming that besides you turning off CPU caches, you're probably using undocumented FSB settings from your motherboard. Usually Socket 7 motherboards used clock generators in conjuction with a 14.318MHz clock crystal reference. Some of these boards have undocumented settings that allow them to boot at REFERENCE / 2 which would give a 7MHz FSB speed, which combined with the K6-3+ lowest multiplier which is 2x, it would give 14MHz clock, which could be why your system identifies as 15MHz. Is this what you're doing?

I suggest you test some speed sensitive games that would require a 8088 4.77Mhz.

In my case the lowest i can with my current Socket 7 is around a 386 DX-25, because the clock generator on my board can't go lower than 50MHz.

carlostex wrote:

I'm assuming that besides you turning off CPU caches, you're probably using undocumented FSB settings from your motherboard. Usually Socket 7 motherboards used clock generators in conjuction with a 14.318MHz clock crystal reference. Some of these boards have undocumented settings that allow them to boot at REFERENCE / 2 which would give a 7MHz FSB speed, which combined with the K6-3+ lowest multiplier which is 2x, it would give 14MHz clock, which could be why your system identifies as 15MHz. Is this what you're doing?

Yes, I have all caches disabled, L1, L2 and L3, and also using the slowest FSB available with a multi of 2x. I get that index score of 1.4 when I have the system in Deturbo mode using the working Turbo Switch. If I keep everything the same, and enable the Turbo Switch, I get an index score of 13.5, and a 486 @ 140MHz, as shown below:

Here's the same system in Deturbo mode with the exact same settings, but this time you can see me using a PS2 mouse, which I am moving rapidly:

... And you can see this affecting the speed - it drops to an index of 0.1!

carlostex wrote:

I suggest you test some speed sensitive games that would require a 8088 4.77Mhz.

Hmm, I don't even know of any games that are speed sensitive and require an 8088. Can you recommend any or even demos?

carlostex wrote:

In my case the lowest i can with my current Socket 7 is around a 386 DX-25, because the clock generator on my board can't go lower than 50MHz.

That's a fairly common low-end when it comes to IA-32 class CPUs. Though, I'm not convinced it's all totally clock-driven, but may also be extreme bus-width restrictions.

Striker from 1985
https://www.youtube.com/watch?v=yTSKuRU1Y0I

I set my K6-III+ to 60 MHz FSB, Multi 2x, disable all caches & use throttle.
But this game still runs slightly too fast.
It requires 4.77 MHz to run properly.

falloutboy wrote:

Striker from 1985 https://www.youtube.com/watch?v=yTSKuRU1Y0I […]
Show full quote

Striker from 1985
https://www.youtube.com/watch?v=yTSKuRU1Y0I

I set my K6-III+ to 60 MHz FSB, Multi 2x, disable all caches & use throttle.
But this game still runs slightly too fast.
It requires 4.77 MHz to run properly.

Nice one - thanks for the link. I researched the game and it's apprently Shareware (one of the first ever?). It runs fine without any lag - instant fire and movement, and I didn't think I'd say this, but seems to be running too slow at around 15FPS. That was probably high-end back then! Looks like a nice little clone of the arcade game Scramble from the early '80s. Any other suggestions for testing would be great.

j^aws can you post you full system specs? I'm specially interested in your motherboard model?

carlostex wrote:

j^aws can you post you full system specs? I'm specially interested in your motherboard model?

Yeah, sure - I'd love to. But it will have to wait until I've sourced at least spare parts. I'm sure you can understand that these parts aren't easy to come by, and once I dislcose everything, I'll probably be looking at some ultra-high BIN or some bidding war, and can kiss goodbye to any reasonable prices.

I may start a System Spec thread with all the details or do a mini roundup with all my findings. But for the time being, the board is a Socket 7 (not Super) with a K6-III+ adapter, and has a working Turbo Header. I know this is kinda cryptic, but if/ when I get to do a System Spec thread, all will be revealed!

Some more tests - same setup as before, but a different board:

... So around half the speed of the last record at an Index of 0.7 compared to 1.0 of an IBM XT 8088 @ 4.77MHz. Normalised, we're looking at something around an 8088 @ 3.33 MHz! Quite possibly one of the slowest IBM compatibles around...

It maybe even possible to go slower by modding this board, but not sure why I would need to yet. The aforementioned game, Striker, runs too slow at this speed.

I might have taken this too far; so, for posterity but mainly for shits 'n' giiggles, here are some more results:

... Still using the previous setup using a K6-III+ and a Socket 7 board, managed to find a slower FSB, so that an Index of 0.3 was achieved compared to 1.0 for an 8088 @ 4.77MHz. Normalised, we're looking at an 8088 @ 1.43MHz...

Tested the game Striker again, and it still runs but loading it takes an eternity, and once loaded, it runs around 3 FPS. Unsurprisingly, completely unplayable, except perhaps using this slowdown as a cheating tool. And explosions in slow motion always look cool.

Gotta admit though, when running DOS, the scanlines updating the screen with text is so slow that looking at the monitor becomes quite mesmerising.

Some more test results with different benchmarks:

Landmark 2

Topbench

I don't know how accurate these benchmarks are at these extreme slow speeds, but the Landmark 2 results are reporting this CPU speed in the KHz range, i.e. below 1MHz.

And I couldn't get a lower score using Topbench, where it gave me the lowest score of 1, meaning it's the slowest PC in that database - even slower than the IBM PC jr. And quite probably the slowest IBM PC benched...

Could any of these banches work w/o use input? 😊

^^ Try the following entry in Autoexec.bat, at the end of the file, type:

1CD "PATH to the directory where your sysinfo file is"
2SYSINFO /DEMO
3
4e.g.
5
6CD C:\SYSINFO6
7SYSINFO /DEMO

... The demo should run automatically.

It does not, it require the NORTON.INI file. What about the others? (Landmark 2 and Topbench) Will they bench autimatically when started, or started with some option...?

trodas wrote:

It does not, it require the NORTON.INI file. What about the others? (Landmark 2 and Topbench) Will they bench autimatically when started, or started with some option...?

Here's Norton SysInfo 6 with the .ini file:

I tested it with Autoexec.bat and it works.

And Topbench:

For Topbench, use "TOPBENCH -i"
You can check syntax by typing "TOPBENCH /?"

I couldn't find an option for LandMark 2.0 though... Maybe the later releases have them.

I made a boot floppy with a few 386 class benchmarks on it: http://www.philscomputerlab.com/386-benchmark-suite.html

Just press ESC after each benchmark, it's all automatic.

Have a look at MIPS bench.

Also you may find useful:
Speed Test by Agababyan (Run: speedtst >res.txt)
System Speed Test by Afanasiev

For games to test look at XT era titles in "Total DOS Collection".

Tertz -

Just press ESC after each benchmark, it's all automatic.

Funny. My keyboard is not working, so... cannot do it. I can, however, create boot CD ( http://www.mediafire.com/?2ecdimfui9g9cu5 ) and test the Norton SysInfo 6 (w/o caches it give 1.6 speed relative tothe original IMB PC, yet I still might tweak the ram settings to slowest, possibly bringing it down a bit) ... and create another with the TOPBENCH... ( http://www.mediafire.com/?1a3dtgx87vctamc )

Lacking a FDD, so going for FDD images on the CD to boot (use UltraISO to work with the bootfile, extract them as *.bif files, change and then load back again to the seemingly empty iso file).

There are the promised results:

Pentium 90 at 10.7MHz caches off:

Pentium 90 at 10.7MHz:

Pentium 90 at 10.7MHz detection:

Sysinfo cannot test HDD:

AMD K5-75 at 90MHz:

AMD K5-75 at 90MHz detection:

AMD K5-75 at 90MHz caches off:

Topbench score 10 Pentium 90 10.7MHz:

(not sure now, if with caches or without them... I will have to re-run it...)

^^ Very nice - thanks for the results. BTW, what's the slowest test for your K5? I can see that you've ran it at 90MHz - can it approach 10MHz like the Pentium? IIRC, there maybe a way to run K5s at 1x multi...

@philscomputerlab: Thanks for the link. When I was looking for benches around XT speeds, these were taking an eternity to complete, hence why I tried some others.

@Tertz: Thanks for those. BTW, SST (Speedsys) was taking like an hour to complete at XT speeds, and all results were below the index of 1.0. An XT at 4.77MHz was around 0.5 on Speedsys. The other benches linked are nice and fast though, and provide relative speeds to an XT.

You are welcome. I was somewhat surprised, that with the caches on, even deadly slow 10.7MHz Pentium perform about 33MHz i386. IPC gains, I assume and the 512kBy of soldered L2 on the board. Off they go and 1.6 is there.
I was badly surprised that memory settings did not played any role. I man... I set the memory settings to slowest (forget 2-2-2 burst, try 4-4-4 ... forget CAS 2, try CAS3 ... etc.) and there was absolutely no difference. I hoped for 1.4 or 1.2... but no. Suxx.

The K5-75 is classic 50x1.5 CPU. It can be overclocked to 60x1.5 but that it is. 66x1.5 fail to post... What is even worser, at 7.14MHz FSB it also fail to post completely. And I already maxed out the Vcore (3.5V), so there is nothing I can do. No way K5 post at such low FSB settings 🙁

Maybe K6 can work at such low FSB clocks? Who knows... but I definitively will try it. You know, K5 is not even a x86 CPU, but rather experiemental superscalar AMD cpu, that get the x86 instruction decoder in front of it and that it is... Maybe that is where it fails?

Another possibility will the the AMD K6-2, witch is not mentioned in the supported CPU, but I quess that it can work, or the bios can be updated to make it work... It does work in Asus TXP4 mobo anyway: http://hwbot.org/submission/2373822_

So there IS a chance for AMD cpu to get into superlow clocks. But I would want CPU, that can be forced to low multipliers. I heard that Pentium Overdrive can be switched to x1 multiplier by disconnecting the fan. Then it switch to x1 in order to prevent overheating... Never tried it myself, but sounds reasonable and people used this there:
http://www.winhistory.de/more/386/xpmini_eng.htm

Another way is mainboard, that have good frequency control. The Asus TXP4-X use as clock generator the ICS 9169 PLL. That chip, however, cannot be controlled by software, because "do not have a System Management Bus connection or the FSB could not be set via System Management Bus. So there is and will be no program that supports these PLL's!!! There is absolutely no chance!"
http://www.cpufsb.de/FSB.HTM

But for some, 2MHz is possible on CPU too: http://www.youtube.com/watch?v=y3wkhDfzqlo (FIC 486-GAC-2 mainboard)

j^aws wrote:

An XT at 4.77MHz was around 0.5 on Speedsys. The other benches linked are nice and fast though, and provide relative speeds to an XT.

I'd call speed close in +-15% difference. More difference can be noticed clear, hence is not good for games linked to the timer.
I like the idea to find a single machine for retro games. The ideal result should support games from 1981-2001 times: full DOS/Win9x period, XT 4.77 - P3 600. I see philscomputerlab's researches in this direction. You do similar research. Some parts may be changed in the using of "retro machine", besides MB, drives, case.