will respond in your dedicated thread

MiTAC/Trigon IH4077D rev D1 based on EFAR EC100G, EC802G (OPTi802GP in disguise).

Bought it for one reason - the label of the main IC, despite knowing it is the not good OPTi 802 silicon.
Didn't expect much from the assembly ... and it delivered on that.

Came in great overall condition but covered with a thin layer of dist. Apparently it sat in somebody's warehouse for long time.

5V CPUs only.
No VRM on board.

Built-in IDE, FDD, COM, LPT controllers.
Very convenient, but CF cards get recognized incorrectly.
Requires manual adjustment of sectors.

Curious connector in the lower right corner.
There is a "terminator" plugged in.
If removed - no lights.
Hints at industrial assembly.

20-pin DIP socket under the array of 28-pin ones for the level 2 cache.
Couldn't figure out what chip goes in because was unable to get the level 2 cache working.
Tried extra hard - no luck. Can be age, can be some subtle damage, can be me tripping somehow.

No support for EDO RAM, obviously.

No support for Am 5x86-133 processors. Hard no lights.
Tried with 5V tolerant chips and Evergreen (and other) adapters that have VRMs. Still no.
Intel/AMD/Cyrix 486 DX, DX2, DX4 processors work just fine.

Intel Pentium Override P24T work well, but the reference manuals available online are incorrect about the jumper configuration.
JP16 must be closed or the memory controller does not see any memory. The memory counter stops at few Kbs only.
JP38 must be closed on no lights at all.

The board came with an impressive ATT BIOS - lots of options. Check it out:

No, i didn't forget to provide the rest of the screenshots.
That's it.

Luckily there is an AWARD BIOS available.
Provides decent amount of settings.

So that's kind of it.
Didn't spend time testing the motherboard.
All 486 tests are performed with Am5x86 processor starting at 4x40.
At some point was thinking to run the POD100 tests, but that's impossible now since the board is not with me anymore.

pshipkov wrote on 2024-04-01, 06:05:

...but that's impossible now since the board is not with me anymore.

Did someone actually buy it, or did you throw the lemon away?

Somebody bought it almost immediately.
I guess the person is either a collector, or intends to build an average DX2 or DX4 system.

pshipkov wrote on 2024-04-01, 06:05:

5V CPUs only. No VRM on board. […]
Show full quote

5V CPUs only.
No VRM on board.

Curious connector in the lower right corner.
There is a "terminator" plugged in.
If removed - no lights.

"terminator" loops 5V back to cpu, its a socket for VRM. I think silkscreen even says VCC GND and VCC3

You are right.
I admit I didn’t put any thought into this contraption despite markings were right there.
I had one or two more 486 mobos with “external” voltage regulators.
Not sure how that didn't trigger a chain of thought at least.
I guess the fact that am5x86 cpus didn't work at all made me shutdown further thinking about vrms.
Thanks for the note.

WJG6260 wrote on 2023-04-27, 00:18:

This board has driven me positively mad; the otherwise standard AMI WinBIOS is obscenely complicated, with numerous features not exposed on most other "normal" boards facing the end user. There's listings for numerous configurable options, some of which prove interesting, and others of which prove absolutely confusion-inducing. I have provided a few pictures below to elucidate the point at which I am trying to reach; that this board, at its very core, is absurdist. The manual explicitly notes, "For optimization purpose[sic], most of the BIOS parameters have been pre-determined[sic] by the manufacturer." They are not even close to joking. Changing even more than a few of the BIOS settings from the "Optimal" configuration setting can hang the board in BIOS. I found the most stable changes that improved performance to be as follows:

Fascinating stuff... Are you able to share the BIOS dump?

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

The GLQuake/MGA numbers come from @Chadti99 - from his MS-4144 and Biostar UUD boards runing properly. He built a fully stable 1 […]
Show full quote

The GLQuake/MGA numbers come from @Chadti99 - from his MS-4144 and Biostar UUD boards runing properly.
He built a fully stable 180/200mhz PC around the UUD.
I will let him clarify here.

---

as for phil's benchmarks:
they provide minimally automated minimal set of DOS benchmarks. nothing interesting on its own, but were the first consistent thing around.
i guess we collectively pushed them as the standard over the last few years, out of lazyness.
i actually have a disk image that has few more things in it but never bothered to promote it. few members here use it. some of the tests there are few more clicks than the single hotkey in phils tests, which is already too much i suspect - C compiler tests, complex offline graphics tests, couple of games to test audio and networking, etc.

---

wintune2 in win31 and win95 - this test always fluctuates between runs, but results are fairly consistent between win31 and 95, so i use it on both OSes with high confidency.

Hey guys, it’s been a while. I’m pretty sure the 35.1 GLQuake score was with a Voodoo3.

Referenced here: Re: The World's Fastest 486

Was thinking it would be cool to decide on a couple dos/glide based 3d tests to standardize on for measuring socket 3 performance. With Voodoo hardware at least. Lots of variables once you get into Windows.

Look who is back. : )

Thanks for the clarification. Makes sense.

Can you suggest such titles ?

There's a list of glide capable games with a frame rate display that work on socket 3's in this Voodoo1 vs. Voodoo2 thread I ran years ago: Voodoo 1 vs. Voodoo 2 on a 486 The challenge is to get the frame rates into some sort of timedemo.

Chadti99 wrote on 2024-04-06, 13:41:

Hey guys, it’s been a while. I’m pretty sure the 35.1 GLQuake score was with a Voodoo3.
Referenced here: Re: The World's Fastest 486

1<Chart>
2GLQuake (fps)
3
4Microstar MS-4144 (4x50), MGA = 35.1
5Luckystar LS-486E rev:D (3x66), Voodoo3 = 35
6Biostar MB-8433UUD-A (4x50), MGA G200 = 30.5
7<Chart>

OK, so the MGA on the MS-4144 heading should be switched to Voodoo3. But what about the 8433UUD listing, claiming a G200 can do 30.5 fps in GLQuake. That also seems erroneous. Should that read Voodoo2?

Oh yah, I’m quite certain that was a Voodoo2 on the 8433. Perhaps I had it paired with a G200, would need to check notes.

Updated the charts according to the clarification from ChadTi99 and Feipoa.

Feipoa, i remember the post you linked very well.
Was going through it multiple times before became an active poster here.
It was one of the few info sources at the time with enough sample points for credibility.
Thanks for that.

---

GEM Computer products GEM-10 based on FARADAY FE2010A-ES

Cute, very compact 8088 assembly.
Been with me for quite some time but sat in the pile of dead components because it came DOA.
I remember trying to resurrect it multiple times, but without success.
Multimeter and oscilloscope probing all checked out, rotated multiple sets of trusted components, but no signs of life.
Yet my sixth sense insisted the board is not dead.

Last week i was exchanging components with somebody and needed a specific type of DIP-16 RAM chips to accompany another 8088 board i was sending to him.
Found them on this GEM-10 guy. After the extraction i was about to throw it back in the junkyard when my sixth sense kicked in.
Scraped few more RAM chips for 256Kb memory buffer. Power connector in. Video card in. Power-on - monitor lit up.
For my great surprise and pleasure.
No idea why it came back to life and how, but hey, i take it.

I was so interested in what the Faraday silicon can do, finally had a chance to find out.

The system runs at 9.54MHz.
According to this page

1Faraday FE2010A can provide 4.77 MHz and 7.15 MHz CPU clock frequencies using 14.31818 MHz crystal oscillator, or 4.77 MHz, 7.15 MHz, and 9.54 MHz CPU clock frequencies using 28.63636 MHz crystal oscillator.

Interestingly enough the bard starts without the presence of the 28.636MHz crystal oscillator in the upper right corner, but the XT-IDE controller never starts initialization.
Also, using crystals with higher frequency than the default one still results in 9.54MHz to CPU.
This was the part i wanted to validate. Did now - 9.54MHz is the max.

Used the next components:
16MHz rated NEC V20 CPU
10MHz rated Intel 8087 FPU
Western Digital WD90C00-JK video card
Lo-tech XT-CF-lite rev.2 IDE controller
full set of 80ns rated chips for full 640Kb of system memory

The system is completely stable . Very satisfying.

Couple of screenshots from some of the standard tests used for evaluation:

Updated the benchmark results accordingly.
At 9.54MHz this motherboard is below average performer, but still faster than 10MHz Juko ST (CPU-12 and some of its derivatives).

All in all - really nice compact assembly.

pshipkov wrote on 2024-04-07, 22:38:

Feipoa, i remember the post you linked very well. Was going through it multiple times before became an active poster here. It wa […]
Show full quote

Feipoa, i remember the post you linked very well.
Was going through it multiple times before became an active poster here.
It was one of the few info sources at the time with enough sample points for credibility.
Thanks for that.

And you still don't own a Voodoo2?

I agree with ChadTi99 that adding some standard glide tests to socket 3's would be of interest. Between the Voodoo1, Voodoo2, and Voodoo3, it would be best to select the card which works well on most PCI based socket 3's - so either the Voodoo1 or Voodoo2, with Voodoo2 yielding better performance.

pshipkov wrote on 2024-04-07, 22:38:

The system runs at 9.54MHz.
Interestingly enough the bard starts without the presence of the 28.636MHz crystal oscillator in the upper right corner

wait, what? I dont see any other clock sources on board

pshipkov wrote on 2024-04-07, 22:38:

Also, using crystals with higher frequency than the default one still results in 9.54MHz to CPU.

as shown by utilities, right? thats because main clock influences all the timers 😀 so CPU runs faster but so does integrated 8253. All benchmarks run will give results tainted by this effect. Only way to measure performance is with a stopwatch 😀

pshipkov wrote on 2024-04-07, 22:38:

At 9.54MHz this motherboard is below average performer, but still faster than 10MHz Juko ST (CPU-12 and some of its derivatives).

you need to calculate how much faster hardware timer works, then multiply all the results by this ratio.

rasz_pl wrote on 2024-04-08, 12:50:

wait, what? I dont see any other clock sources on board

I know. Quite surprising, right ?
Wonder if there is a built-in crystal in the big Faraday plate that produces 4.77MHz frequency to CPU in this circumstance.
Thinking about this now, i didn't measure that. Will check soon.

rasz_pl wrote on 2024-04-08, 12:50:

Only way to measure performance is with a stopwatch 😀

Exactly.
The silly synthetic tests like CheckIt, Landmark, etc. are hard to validate, but Superscape completes for roughly the same wall clock time.
I need to flip through previous posts because there was at least one other motherboard like that ...

rasz_pl wrote on 2024-04-08, 12:50:

you need to calculate how much faster hardware timer works, then multiply all the results by this ratio.

Agreed.
For now i reported the perf at 9.54MHz in the combined charts.
My point was that at 0.05MHz disadvantage the board is faster than the 10MHz Juko ST.
Obviously that is a clock-to-clock advantage.

feipoa wrote on 2024-04-07, 23:07:

And you still don't own a Voodoo2?

I have Voodoo 1 card, not used, and Voodoo 3 in the LSD 200MHz hotrod - fits very nicely there.
Tried few times to get energized about Voodoos but doesn't click.
I guess my attention is mostly on professional 3D graphics hardware, or consumer products that can function in that role.

feipoa wrote on 2024-04-07, 23:07:

I agree with ChadTi99 that adding some standard glide tests to socket 3's would be of interest. Between the Voodoo1, Voodoo2, and Voodoo3, it would be best to select the card which works well on most PCI based socket 3's - so either the Voodoo1 or Voodoo2, with Voodoo2 yielding better performance.

We can certainly use a set of additional Glide tests to further validate stability, but from what i see here and what is evident from your post on the subject - the 486 class hardware is too slow to keep these cards busy, so the difference in performance between Voodoo 1/2/3 is negligible.
What you think ?

@rasz_pl

So, without a crystal the Faraday chipset produces frequency of ~4.11MHz on its own.
Unfortunately IDE/FDD extension cards don't work without a crystal, so POST stops after RAM count and keyboard controller check.
I don't have crystals that cover the space after 28.636MHz space with enough granularity.
So, i can check 30MHz and 36MHz only. At 36 POST does not complete. At 30 there is virtually no perf difference compared to the standard for this board 28.636MHz, especially if we consider that there is a human error on the stopwatch. But also, the system gets flaky even at that moderate overclocking.

pshipkov wrote on 2024-04-08, 21:01:

So, without a crystal the Faraday chipset produces frequency of ~4.11MHz on its own.

Why would Faraday/WD not document this? This is fascinating, how would they derive such a fast clock with just silicon, internal RC oscillator (not easy in the eighties)? How would it know when to oscillate versus switching to external? One thing that comes to mind is PCB routes RC components to clock pins to begin with, and sticking additional crystal overrides it. Would be interesting seeing stability of this 4.11MHz. Accidental or RC derived clock would be all over the place, this would explain IDE not cooperating with such dirty clock.

Are these two chips maybe 386-CPUs from IBM?
Should I risk to open/take-down the aluminium covers of them without damage of the cores?

Sadly I can not give any additional infos about them, as they are from "a friend of a friend of a friend"...