First post, by Mithloraite
- Rank
- Member
Recently I stopped fiddling with the easy and straightforward Pentiums and ventured to explore more of the world of 486 quirks and horrors.
It lives up to the reputation! 😄 Amazingly, the matter of performance was of grave surprise here.
1.
The best I managed to get from A-Trend ATC-1415 UMC PCI board
http://img189.imageshack.us/img189/6738/atc1415a.jpg
was the meagre score of 43 on Speedsys 4.78 processor benchmark.
The video PCPBENCH benchmark resulted in a score of 6.0 on a Matrox Millenium-II PCI 4mb card.
Yeah, this board failed to recognize AMD X5-133 so I had to give it a Bus Kick to 40mhz bus... just to get 120mhz speed instead of 100, seemingly fixed by BIOS.
This UMC selects the CPU multiplier by its own bios autodetecion, not the jumpers.
Even if set as Intel Overdrive, or other Intel, or any AMD it boots up and it sees only an AMD DX4-100. Such a dedication on its part.
a PCI divider is also present here as "auto" by bios or one can enter it manually.
Well, I suppose that's merely the lack of the last bios.
Otherwise it seems to have an OK speed.
but I really dislike the total lack of mouse (COM) and floppy ports on this board @__@
It just looks like malice and misery on designers' part. Crooked amputees are not my favourite sight. This forgotten horror does have great IDE ports, though... as if as a cruel joke.
2.
A Sis 496/497 PCI Lucky Star board
http://motherboards.mbarron.net/models/486pci/ls486e.htm
gave me an abysmal score of 24.8 on Speedsys 4.78 processor benchmark.
And yes, it recognized the AMD X5-133 with no problem.
This is attested and mirrored by the vastly inferior Video Pcpbench score of 3.3 (or 3.8 with better bios settings.)
3.
A Sis 496/497 PCI A-Trend ATC-1425B board
http://motherboards.mbarron.net/models/486pci/atc1425b.htm
has given the same so horrid 24.8 on processor benchmark.
Again no problem to recognize the AMD X5-133 as it is.
For some reason the Video score was somewhat better, 4.4 on Pcpbench.
(same video card, same RAM, same CPU "performance"!)
Shall we conclude Lucky Star is ~crappier~ than A-Trend in its 3D video ability?
4.
Well I started to dislike this kind of performance and tried another ATC-1425B board.
(this one, with 5 leads on voltage stabiliser instead of 3. was it a more 'modern' version I thought?)
The results were exactly the same 24.8 for CPU and 4.4 for video.
No problem with recognizing the processor. But such a persistent misery.
In an AD&D Gothic Horror game this would be the time to roll the dice for Fear and Horror checks.
Like characters screaming in fear or feeling revulsion 😄
These reactions though don't help anyone so I decided to check my processor instead.
Having used an ADZ-133 version instead of ADW I saw, well, the same worthless score.
* * *
For now I still have some more horrors to see.
Like the 486SPM SiS board (it has lots of cache mountings for 32 dip chips as SiS tends to need!) and an EXP8449.
This last one has a funny habit of requiring 50ns EDO memory to function properly.
60ns chips need setting some BIOS wait states and other slowing measures to work.
To sum this dreary account up, is there a way to the Brilliance of performance, out of this darkness of vintage terrors?
P.S.
What I can think of might be using better BIOS versions, for all boards.
Another note on ATC-1415. When cleaned, the weird small battery got alive, after virtually 5 minutes of charging. Normally these are made by Varta, and are quite dead. Or at least supposed to be? "That is not dead which can eternal lie." 😄