Is the UMC UM82C491F chipset a known terrible chipset?

[Moogle!]

I have 2 motherboards from different manufacturers. They both have 256K cache. I have tried the same AMI BIOS ( and the same Mr. BIOS (recently posted on Vogons). Both were tested with two different 16MB ram sticks and two different DX2s. I have combed over the BIOS settings and jumper settings many times across many days. These boards are just slow. They score worse than a DX-40 on Speedsys and I can't get more than 2.9 frames in the Quake Demo. (as opposed to a slightly beat up Sis board which gave me 7.3 frames, and a Bioteq board which game 7.2)

I realize I am not the most colorful bowling ball in the shed, but I cannot see what I am missing here. Any ideas?

 

[Chuck(G)]

Just a thought--are the cache chips real or are they dummies? There were a couple of manufacturers who used dummy chips in their "cache".

 

[Moogle!]

In both cases they were pulls from other dead boards. Cachechk says they're present and real. The numbers get slower outside their range.

 

[Moogle!]

Apparently, the turbo setting is 'header open' instead of 'header closed'. I cannot believe I went so long without checking that.

 

[Chuck(G)]

So this is a "never mind"?

 

[rvdbijl]

I can't help you directly with the slowness of your board. But anecdotally, back in the 90s when I built and upgraded my PCs regularly (and also helped others do so), I avoided any UMC boards like the plague. I think I had more stability than speed issues, but perhaps confirming some of your observations.

 

[Moogle!]

A lot of boards require the turbo jumper to be closed. These two boards require it to be open for normal 'turbo' mode. I spent way too much time before I figured it out. One gives me 7.8 frames in Quake, the highest I have seen any board do with a DX2 66 (standard version).