Hold on.
It just occurred to me - i have POD100 in the VLI computer while you are running Am5x86 i think.
Can this be the difference ?
Worth checking ...

---

@WJG6260

When i got into the retro hobby i was quite period correct, but over time that was kind of forgotten.
It is good you guys brought it up and further emphasizes the progress leap at that time.

Yes. Sometimes the long shot hits the bullseye.

WJG6260 wrote on 2022-10-03, 01:12:

@andre_6 The PhilsComputerLab benchmark suite is excellent, as noted by pshipkov. It's very easy to use. I'd recommend running […]
Show full quote

@andre_6
The PhilsComputerLab benchmark suite is excellent, as noted by pshipkov. It's very easy to use.
I'd recommend running options 2), 5), 6), a), b), c), and whatever else suits your interests.
Those are what we tend to use the most in this thread.

I'm curious to see your results! Please do share them, if you don't mind.
These past few posts have piqued my interest a bit in Trident hardware.

No problem, might take a while but I'll get to it. I knew that the ISA only aspect of my 486 would always limit it greatly, but in truth it's more than capable for what I need it for. Everything else I'll just use my Pentium Pro PC. I will always be curious to squeeze that little bit more of performance out of it though. It's a testament to Vogons that a learning user like me can briefly hijack a thread because he spotted something he never knew, and then the thread goes on naturally. That and the unsurpassed knowledge of everyone around here. Best forum around

Caching Tech Corporation C386MX revision 1.0 based on Macronix MX83C306FC/MX83C395FC chipset.
... with a bit of contradiction between the company name and the L2 cache-less nature of the motherboard.

Always liked the Octek Jaguar V motherboard (that is based on the same chipset).
It is not the fastest 386 implementation but pretty up-up there. Also, overclocks pretty well.
While good performance is certainly a factor, what really impressed me was how snappy the assembly is.
Alaris Cougar is snappy, DTK-PEM-4036Y is snappy, but nothing compares to the Jaguar V.
A second after restart and you are in DOS prompt.
Unfortunately it has a surface mounted Am386DX-40 CPU which is a limiting factor considering DLC, SXL2, DRx2, BL3 processor upgrades.

For long time i wanted to find another motherboard with that chipset but with CPU socket and see where it can go.
Turns-out these are very rare. The only one i was able to find was this C386MX guy.
It was clear that it won't be a top performer, but the point was to find-out what CPU upgrades are supported by the chipset ... while keep looking for the unicorn.

Everything started well. Board lit-up right away with the 386DX CPU that came with it. Then:
Tried DLC - worked.
Tried SXL2 - all good.
Tried BL3 - no lights.

Good start.
Decided to put it on the side (all this happened at least a month or two ago) in favor of some other retro activities.
At some point later came back to it - no lights. Noticed that the clock generator gets burning hot. Ordered a replacement. More waiting.
Replaced it. Lights are back on. Was about to run a brief set of tests for the record but the keyboard controller stopped working.
Tried several things - no go. Soldered a keyboard controller socket. Tried bunch of different chips - no and no.
Decided to run it through the oven for a quick bake to reflow cold joints.
Had great success bringing back to life few motherboards like that and/or fixing issues on them.
Aaaand that was the end of the poor guy - permanently dead after that. Too bad.
From my initial tests the board was very slow, but don't have captured metrics. So you have to take my word on it.
Still, the mission is kind of achieved - now i know what CPU upgrades work on that chipset.

In my experience the chipset is important, but performance can also be hindered by a crappy motherboard and/or BIOS...so if you think the MX chipset has potential, I wouldn't give up on it just yet. I think I saw a few different boards based on these for sale in China. If life here ever goes back to something close to normal, maybe I'll give one a try. Personally, I'd much rather have this chipset on an SX board though.

Didn't give-up on that chipset. The contrary.
Hope to find one day a motherboard with the qualities of Octek Jaguar V, but with CPU socket.

With that said, i am going to pencil down for some time on 286/386.
Some cool late XT boards are staring at me more and more intensely.

@feipoa
My pleasure! I am glad to have been able to confirm the oddness on my end. I used the NICE/MCCI SuperEISA that's shown earlier. My board was equipped with an AM5x86-133 running at 160MHz at the time, and the board was set so that all BIOS timings were at their tightest/most optimized. I did test the board multiple times and found that, no matter the timings, bus speed, setup, or anything else, I could not get soft resets working with the Trio64 -P and Diamond v2.09 BIOS.

That being said, I can't try a POD on my board, unfortunately, as my SuperEISA lacks a ZIF socket. I suppose I could run it without the extra pins, but I've only got one POD and don't want to fry it, especially since it's seemingly stable at 100MHz without modification.

I will add that I did not have any issues with VESA modes with the Number 9 BIOS, but I do not recall which modes the v2.09 BIOS added. I can retest and see if all VESA modes work there with the Number 9 BIOS.

----
@andre_6
I wouldn't call it derailing in the least!
It's an interesting thought and contribution and always good to have this information available!
ISA-only 486s might be "limited" in some regards, but the truth is that if it makes you happy and does what you want, it doesn't matter!
I can personally confirm that later/faster ISA cards like the ones @pshipkov has numbers for earlier in this thread are excellent and truly do shine in Windows/DOS/all-around.
But ISA is limited and it is indeed the limiting factor. You'd be hard-pressed to really find a noticeable difference in performance, unless you were seeking higher color modes, and then the PPro is perfect for that anyway.

----
@pshipkov
It's always nice to see how things changed and look back.
Definitely agree that it's easy to forget.
Mind-boggling to think, for example, that the Cyrix MII and PII/III coexisted at one point.
Pretty amazing to see how much of a leap the Pentium was/is.
486s are "where the fun is at," but Pentiums have all the "go," albeit with less "show."

Absolute shame about that MX board. I think it's a really nice-looking assembly, but it's such a shame that its time came.
The Jaguar V is a great board. Good to know the upgrade CPUs work on the MX assemblies.
There are a few with sockets, but they are not that common. This one is one of the few I've seen.I have the SX version, which is nearly-identical.

I think @AnonymousCoward is right about this chipset being interesting on SX platforms.
I have one such board and it's noticeably faster than a non-cache, 0ws 386SX. Nice stuff.

Looking forward to seeing some more XT/386 hardware around here! 😀

Hello,
I tried to join your "486 party". I only have one board that allows higher speed, it's the asus sv2gx4 rev.2.0.
I have 3 different pieces of Am5x86-133, so I tried: Set the bus to 50MHz (what else?). All 3 worked fine - no surprise - at 150MHz, but none of it showed a sign of life at 200MHz. Damned. 😁
Will try more of it, maybe I find one that runs at 200.

Hey CoffeeOne,

Bring your hardware bag and welcome to the "486 party".

A good motherboard you have. That's for sure.
I don't know why, but more CPUs lit up on VLI at 200MHz (4x50) than any other board.
But so far no one succeeded to run any of them stably. To my knowledge at least.
You will most likely need a Peltier element to freeze the processor and a big heatsink+fan to disperse the excessive heat produced.
That is if the CPUs you got there have the chops for 180/200MHz.

pshipkov wrote on 2022-10-03, 21:20:

Hey CoffeeOne, […]
Show full quote

Hey CoffeeOne,

Bring your hardware bag and welcome to the "486 party".
1.jpg

A good motherboard you have. That's for sure.
I don't know why, but more CPUs lit up on VLI at 200MHz (4x50) than any other board.
But so far no one succeeded to run any of them stably. To my knowledge at least.
You will most likely need a Peltier element to freeze the processor and a big heatsink+fan to disperse the excessive heat produced.
That is if the CPUs you got there have the chops for 180/200MHz.

Hi,

At the moment I can't really do useful tests, because I don't have any IDE disks available. So I will get also such a compact flash adapter, used by a lot of people and then do meaningful tests and provide some pictures (but nothing special).

I have a Trident 9440 VL card, which seems to be really damned fast at 50MHz, so maybe I can have good values (at least for graphic throughput) with the "low" 150MHz.
Not yet sure, what timing parameters for cache and ram I can use at 50MHz.
But as I said, need the flash thing, because right now, I can only go to the bios and boot from floppy disk.

I don't think I will go for the Peltier element route, I don't like that. So I only will test with normal air cooling, but I test in the basement and winter is coming, haha.

EDIT:
And I forgot: I now have an old BIOS 3.04 I think. So I need also a programmer to program the latest version with the famous write-back modification 😀
So a lot of things needed, before I can compete ....

@CoffeeOne
Nice board you've got there! Excited to see some results.

Trident 9440 VL is a great card. Very good all-arounder.
@pshipkov and @AnonymousCoward have some very helpful discussion on the previous page or two.
Is your 9440 a 9440, 9440AGi, or 9440-1, by chance? I am curious, as I still cannot get my 9440-1 going in Windows with AGi drivers.

Regarding compact flash: You can't go wrong with VL IDE and CF.
@pshipkov has excellent tests linked earlier. Very helpful and enlightening.
IDE to SD is another route. I've used it myself and it's just as easy. CF is a little more "hassle free," but that's about the only difference.

WJG6260 wrote on 2022-10-03, 22:59:

@CoffeeOne Nice board you've got there! Excited to see some results. […]
Show full quote

@CoffeeOne
Nice board you've got there! Excited to see some results.

Trident 9440 VL is a great card. Very good all-arounder.
@pshipkov and @AnonymousCoward have some very helpful discussion on the previous page or two.
Is your 9440 a 9440, 9440AGi, or 9440-1, by chance? I am curious, as I still cannot get my 9440-1 going in Windows with AGi drivers.

Regarding compact flash: You can't go wrong with VL IDE and CF.
@pshipkov has excellent tests linked earlier. Very helpful and enlightening.
IDE to SD is another route. I've used it myself and it's just as easy. CF is a little more "hassle free," but that's about the only difference.

Hello,
About the 9440, see the picture.

OK, I did not even really start, but I have already a severe problem....

As I want to use the external 50MHz, I tried only 150MHz and 200MHz, where there is nothing with the 200MHz setting (Am5x86 setting for L1 write back, clock 50).
OK, but there is another problem, the jumper setting which I think is correct, does not work at all for multiplier 4.
I cannot set 133 (33MHz x 4) neither.

I use that setting:
JP16 1-2, 5-6 short
JP17 1-2, 5-6 short
JP18 1-2 short
JP19 1-2, 3-4, 5-6 short
JP20 All Open for 3x clk, (1-2 short for 4x clk)
JP21 3-4 short (WB)
JP22 1-2, 4-5 short

and JP5 and JP6 are both on 1-2
for: "i486DX4-&EW, P24D, P24T, Am486DX4-SV8B"

so when I have JP20 "all open" everything is fine, I get 100, 120 or 150MHz, with JP20 1-2 short: no picture.

Are my settings wrong???

I assume it is the BIOS, I have 3.05 (above I wrote 3.04 which is not fully correct)
So I need the new BIOS as a first thing, right?
Maybe I have already a 200MHz and don't know it?

Try this instead

JP16 1-2, 5-6 short
JP17 1-2, 5-6 short
JP18 1-2, 5-6 short
JP19 1-2, 3-4, 5-6 short
JP20 1-2 short
JP21 3-4 short (WB)
JP22 1-2, 4-5 short

pshipkov wrote on 2022-10-04, 21:02:

Try this instead […]
Show full quote

Try this instead

JP16 1-2, 5-6 short
JP17 1-2, 5-6 short
JP18 1-2, 5-6 short
JP19 1-2, 3-4, 5-6 short
JP20 1-2 short
JP21 3-4 short (WB)
JP22 1-2, 4-5 short

Thx for helping out.
I have to say, I feel like a complete newbie now, but actually I owned used that board with the same revision 2.0 (as base for my main workstation) in the late nintees with an Amd 5x86-133, ooops.

Your posting was similar to what posted that I have set, except I had JP 18 only 1-2, you have listed 1-2, 5-6. In fact what I wrote here is the setting from
the zip file:
Asus_486SV2G_and_486SV2GX4.zip
under jumpers which can be found here on Vogons.

BUT - SORRY
I did not set it 100% like that, instead I mixed up with the settings from the pdf file
vl_i_486sv2g.pdf => which can also be found here on Vogons. So for example JP 16 was 1-2, 4-5 not 1-2, 5-6

Then I set the jumpers exactly as you wrote, but I could not boot at all (independent of multiplier setting 3x and 4x) => WTF.

After this total confusion, I started over. And used the setting for a normal Intel DX (SL Enhanced), which should be fine for the Intel DX4 in write through mode, too. And for Amd 5x86 write through, too.

so:
JP16 1-2, 5-6 short
JP17 1-2, 5-6 short
JP18 1-2, 5-6 short
JP19 1-2 short
JP20 All Open
JP21 All Open
JP22 1-2 short

and hardware trap: JP5 1-2, JP6 2-3 (instead of 1-2, 1-2 from before).
Voila, the machine boots again! With 33MHz showing DX4-S and 100MHz.
But behaviour is similar as before: When I now set JP20 to 1-2 (which is 2x for the DX4 and 4x for the Am5x86) => No boot! So still no 4 x multiplier, but system now in L1 WT mode. 🙁
I will continue with that on the weekend ...

But I now really believe I cannot get the 4x multiplier because of the Bios 3.05. But not sure.
Those different manuals (both say that it is for Rev.2.0 and 2.1) are a bit annoying, too, honestly spoken.
Anyway I want to have the new bios with the additional "write back fix" .....

No worries.
We are all in the same boat really.

It is best to go with an image reference for these things.
Take a look at this one - https://www.petershipkov.com/temp/retro_pc_im … gx4_rev_2.1.jpg
Jumpers are configured for Am5x86 and 4x40MHz.
Is that how you have it ?

pshipkov wrote on 2022-10-04, 22:34:

No worries. We are all in the same boat really. […]
Show full quote

No worries.
We are all in the same boat really.

It is best to go with an image reference for these things.
Take a look at this one - https://www.petershipkov.com/temp/retro_pc_im … gx4_rev_2.1.jpg
Jumpers are configured for Am5x86 and 4x40MHz.
Is that how you have it ?

Hi, sorry I did not answer.
Yes, I had the jumpers like that.
But as I wrote above my last try was this:

I know that is not the correct setting it is for an Intel DX or DX2 or DX4.
So that works and gives me 33 x 3 = 100MHz. But putting the green jumper to 1-2 => which should be 133MHz no boot.
So as I wrote I stopped playing and will continue with a new BIOS. I am sure I will get then the 4 multiplier ....

Not sure if new bios will help, but give that a try.
Something else must be going on. To clarify - did you confirm that each and every jumper on your motherboard was setup like the photo-reference i provided early ?

---

My interest in Turbo-XT machines culminated in putting together an 8086 based Tandy 1000 SL2 beast running at over 17MHz.
(covered here, here and the finale here)
At the same time 8088 never stopped being of interest to me - especially late, highly integrated, standard Turbo XT implementations.
They are not that many, which makes them a nice area to explore without breaking the bank account.
Will share findings as i go.

First in line is DTK PIM-TB10-Z revision 9 (most recent model of the series).
Utilizes smaller number of higher density RAM chips compared to early variants.
This is important factor since overclocking of XT and 286 hardware is predominantly limited by the quality of the installed RAM.
More chips = bigger chance for some of them to be flaky. Memories from that time were much more fragile than later generations.
Ok, back to the mobo - not exactly a very late XT model given the relatively big amount of logic components on it, but close enough.
PCB is still in great condition. Very sturdy and well built. Also, it just works, after all these years. No effort whatsoever.
Default frequency is 10MHz.
Came with 640Kb 150ns rated RAM chips and 10MHz rated 8088 Jujitsu CPU.

After the initial round of tests, where everything checked right away, it became clear that this is a fast assembly.
Decided to see how well it overclocks and how far it can go on the performance scale.
Inserted a NEC V20 CPU and 10MHz rated 8087 FPU.
Swapped the 150ns RAM chips with 80ns ones.
Soldered a crystal oscillator socket.

Started with 36MHz oscillator for 12MHz base frequency.
The very short POST process never reaches a point to start testing the installed memory.
Stepped down to 33 and then 32 MHz oscillators - 10.3MHz.
Again - POST does not start testing RAM.
Ok, this was the fastest overclock failure i had so far.
Tried hard to get it to do anything above the base 10MHz but failed.
There is something in the motherboard that is really not having it with "higher" frequencies.
Still, it was worth giving it a try. Not disappointed really - the board is great as it is at its 10MHz.

There are not a lot of good benchmarks that cover XT hardware well enough.
So going with AT Perf, CheckIt, SuperScape, LandMark to measure and compare performance.
Here is what they say for this motherboard. Used XT-IDE adapter and ET4000AX video card.

An interesting observation about video cards.
On the Tandy 1000SL ET4000AX was slightly behind compared to some earlier video adapters, but no such difference was observed with the DTK board.

benchmark results

Performance is not bad at all, but the system does not overclock very well.

Superwave Super-10b.
Late 8088 Turbo XT motherboard.
Very high level of integration. Basically a single M1101 chip which i believe is Acer Labs (ALI) silicon.
PCB is in great condition. Came with standard Intel CPU and no FPU. 30MHz oscillator for 10MHz system.
Mix of 120 and 150 ns chips for 640Kb RAM.

Everything worked right away as advertised.

Upgrades:
Replaced the CPU with NEC V20.
Added Intel FPU.
Swapped the RAM chips with 80ns ones.
Made a 2-pin crystal oscillator socket using legs from 4-pin crystal oscillator socket. Soldered them directly in the through-holes. (upper right corner with inserted oscillator in them).

This allowed smooth overclock to 12MHz.
Board keeps lighting up all the way to 16MHz but after the 12MHz mark POST never completes.
Tried hard with trusted components - different CPUs, minimal RAM, etc. Didn't help. The limitation seems to be in the motherboard itself.

As mentioned in the previous post - there are no great benchmarking programs for XT class hardware so going with some synthetic tests.

Default 10MHz:

Maximum possible overclock 12MHz:

Some graphics corruption in the Superscape test. No such issues in games and other VGA graphics. Just this one.

benchmark results

Few things to notice:
1. DTK PIM is slightly clock-to-clock faster than this motherboard, but with +2MHz overclock Superwave comes on top.
2. The 12MHz 8088 based Superwave is slower than the 8086 based Tandy 1000 running at 10MHz.

Didn't know what to expect from this board.
Turned-out great. One of the best i have tried.
Enjoyed many hours messing around with it.

---

During the testing phase i accidentally shortened two pins of the third ISA slot. This resulted in the corresponding lines to burn out.
Didn't notice it initially, but the no lights on power was very noticeable. Close inspection revealed the damage.
Was so bummed by "running" the pristine PCB condition.
My solder work is visible on the right hand side under that ISA slot.

Been a while, but hope everyone here has been well.

@pshipkov
Really loving the XT adventures.
Interesting to see the disparities between the 8086 and 8088.
On paper, it's easy to understand the differences, but in reality, it seems the picture is even less clear without some numbers.
And wow, oh wow, what numbers!

Pretty neat to see a NEC V20 going against your awesome Tandy setup.
I've always wondered how different things would be if the IBM 5150 was an 8086-based machine.
Seems that there's performance to be had with the 8086, but the V20 is no slouch. It certainly bests the expectations I had for having half the external bus width.

That Superwave board is very nice; seems to be a great quality assembly.
Had high hopes for the ALi chipset; good to see that it isn't too bad and can take 12MHz.
Bad news is that it's clock-for-clock behind the DTK board.
That DTK board is a great one-a shame it struggles past 10MHz.
I can't believe the 32MHz oscillator was still too much for it!
I admit that XTs are not my forte, but I find the hard struggles at 10MHz interesting, nonetheless.

Got me thinking a bit about some of our 486 experiments; I am curious if just, at a certain point, we're seeing noise issues or something else.
Could be quality of the PCB, could be trace layouts, etc.
VLI, for example, has proven to be the best of the best so far.

-----
Finally laid hands on an S3 Vision964, and have some numbers now for the Vision family sans 968 and Trio32/64 sans Trio64V+.

First, however, the main event:

What we have here is a Miro 20SV.
You might notice the BIOS has been replaced. I am currently running this board with the BIOS from the Diamond Stealth64 VRAM v1.07H.
There are two BIOSes. Diamond specified that the H BIOS was appropriate for Korean-produced chips and the U version applied to Taiwanese chips.
I have the original BIOS, and I think it provides some insight into the S3 Vision868 performance quirks exhibited by the Miro 20SD.

Long story short, BIOS performance with the Miro 20SV BIOS is abysmal in comparison.
It appears that the Miro BIOS employs relaxed timings, as SuperScape, for example, produced a score of ~84-85 stock.
The S3 Vision964 itself uses the same 2D core as the rest of these cards (as far as I am aware, feel free to correct me).
Performance therefore should be approximately equal.
Some differences do exist, and it is notable that the Vision868 and Vision964 follow the pack in Wolf3D with "subpar" scores.

The Miro 20SV uses the same clockgen as the Diamond Stealth64 VRAM (964 version) and uses the same clockgen.
The RAMDAC is the only difference, but the AT&T20C505-13 is a 135MHz RAMDAC compatible with the Brooktree 135MHz unit on the Diamond.
In fact, some revisions of this PCB contain said Brooktree RAMDAC.

There are Miro 40SV cards, which have 4MB of VRAM.
Otherwise, they are pretty much the same, as far as I can tell.

In Windows, the VRAM advantage is clear.
The 2MB of 70ns VRAM on this card produces an excellent Windows experience, no doubt.
Acceleration is fast, and it is clear that there are fewer memory bandwidth limitations on this unit, especially at higher colors.

The Vision868 should be faster than the 864, and the 964 should be slightly outdone by the 968.
The reasons for this are not exactly clear, but this is what I have read and heard.
It seems that the x64 series of cards lacks YUV video scaling but retains CGA/EGA backwards compatibility, unlike the x68.
From what I understand, the Vision868 can utilize single-cycle EDO timings for increased performance. Perhaps this is where the difference lies?
Ideally, I would replace the BIOS on my 868 and re-run some numbers. I am still in search of a suitable replacement BIOS for the Miro20SD Vision868.
I have tried the #9 and Stealth64 DRAM BIOSes, to no avail.
I would like to try an ELSA Winner 1000Pro's BIOS, but those are seemingly hiding from me at the moment.

A Vision968 like the Diamond Stealth64 VIDEO would be equipped with a 64-bit bus to a RAMDAC with an internal PLL like the TVP3026.
The Vision968 model of the Orchid Fahrenheit Pro Video 64 VLB, however, uses the same Brooktree RAMDAC as this card/the Diamond Stealth64 VRAM.
Unfortunately, this situation is not quite as nice as the Stealth64 VIDEO's, as the RAMDAC's datapath is halved.
The 964 is certainly competent and quite potent, but a 968 configured with a 64-bit bus to the RAMDAC would be superior, undoubtedly.

I'd love to add some results for a Trio64V+ and Vision968, but both are rarer than hen's teeth.

For now, here's the gambit of S3 VLB accelerators from the Vision/Trio line.

The Verdict: All of these cards are incredible for DOS interactive graphics and Windows acceleration (except for the Trio32, with respect to that latter point, by design). You really cannot go wrong, but the VRAM-based 96x cards are bound to be the ultimate in Windows and penultimate in DOS.

Using the same SuperEISA configuration from earlier, here's a few details:

Notes: One important thing to note is that the S3 Vision964 was run with only the Diamond BIOS. Performance with the Miro20SV stock BIOS was subpar, to say the least.

These Turbo XT mobos are fun of its own type.
Best part is they are pretty much no hustle and there are not that many of them which makes it easy to cover the area without spending the national treasure along the way.
Been itching me for a while to run some facts finding around 8088 / 8086 class hardware and related NEC V20/V30 processors. Doing it now.

To your second point about overall quality and reliability of 486 and other vintage hardware classes:
It is extremely difficult to build top performing, error free, fully featured computers from this old "junk".
It takes a lot of effort and lets be honest with ourselves - money, to achieve that. Especially if error-free has to be fully respected.
Agree with you that in the world of classic 486, Asus VLI (and PVI) are some of the best motherboards to build such computers around.

Great info about the S3 VLB cards.
Please remind me, i remember you saying that you have 968 in a box, is that correct ?

I don't have any of these Vision VLB cards, so the info is very interesting to me.
It looks like they are all equal in DOS interactive graphics. The reason for that is most likely a CPU bound performance.
As you said Diamond Stealth DRAM being the best overall there, by a hair.

The Windows accelerated GUI is also telling a good story 964 is about 7% faster than Trio64.
We assume that 968 is probably 10% faster.
So in reality these late Vision and early Trio64 are more of the same really.
An advantage some of the high-end Vision models is the presence of 4Mb RAM - that's higher resolution GUI, which is nice.

Finally, from your notes - it really shows how important is the software integration.
Same hardware + different BIOSes = different outcome.