brassicGamer wrote:

JidaiGeki wrote:

Hear you on the drowning in parts, I found this board yesterday in a box I bought a while ago ...

Heard that VLB-based Pentium boards existed. Never seen one. Opti made some crazy shit for sure.

I didn't know they existed until recently as well. Found one on E-Bay a few months ago, but too much $$$ in my opinion.

http://www.ebay.com/itm/301728666510?_trksid= … K%3AMEBIDX%3AIT

Brickpad wrote:

I didn't know they existed until recently as well. Found one on E-Bay a few months ago, but too much $$$ in my opinion.

I find these types of board historically fascinating but, really, what's the point of a Pentium system without PCI and pipeline burst cache? Either it's bleeding edge for the time or it's a late cop-out to use up over-stocked parts or something? Don't get me wrong I have a lot of love for OPTi, mostly because of the 495SLC.

Extremely rare vintage graphics card...

It's a GTX 275!

brassicGamer wrote:

what's the point of a Pentium system without PCI and pipeline burst cache?

- pci is not better,than vlb, even vlb is can be faster, than pci, on 40 or 50 mhz. And pentium1 is not enough to load any of this bus to 100%.
- I've used different benchmarks to find a difference between cache and pipeline burs cache on same pentium board - maximum difference was less than 5%.

Thing is, the original VLB implementation was tied to the 486 architecture. Whatever the OPTi chipset does, it is doing some bridging and thus the VLB performance will greatly depend on the chipset - and if the VLB implementation is bad, the performance will suffer.

philscomputerlab wrote:

Extremely rare vintage graphics card

GTX 275 were middle range cards, not rare. If do not take into account concrete producer.

Tertz wrote:

philscomputerlab wrote:

Extremely rare vintage graphics card

GTX 275 were middle range cards, not rare. If do not take into account concrete producer.

I was being sarcastic 🤣

I wouldn't call it middle range though, it's just a step below the GTX 285, the top of the line GeForce at that time.

kithylin wrote:

I have no idea what all this opti hate is. I own two opti motherboards myself and they seem to run just great. Both of my 486 motherboards are opti chipset and I love them. Very stable.

OPTi's best days were probably back when they were making 386 chips (Anyone with a 381 or 391 that can confirm?). Their 486 stuff was *okay* until PCI came along. Everything after that was crud. OPTi made about a bazillion different chipsets, but I can't seem to recall any that were outstanding performance wise. I've had a number of OPTi based boards. Two based on 495SX, a 295SX, a 572, a 683 and a 693. Of these, only the 295SX really impressed me (and it's on an SLC board, which was a budget system back in the day).

Based on what I've used, I'd much rather have SiS, UMC, Forex or possibly even ALi.

easy_john wrote:

brassicGamer wrote:

what's the point of a Pentium system without PCI and pipeline burst cache?

- pci is not better,than vlb, even vlb is can be faster, than pci, on 40 or 50 mhz. And pentium1 is not enough to load any of this bus to 100%.
- I've used different benchmarks to find a difference between cache and pipeline burs cache on same pentium board - maximum difference was less than 5%.

CPU aside, I've been trying to do this very thing recently with three 486 systems to find out how much difference bus speed actually makes to the system's performance. I gave up in the end to revisit it when I've got more time but I really wanted to focus on real world productivity benchmarks rather than the usual synthetic ones that people use. Otherwise many performance gains / drawbacks remain hidden.

Anonymous Coward wrote:

kithylin wrote:

I have no idea what all this opti hate is. I own two opti motherboards myself and they seem to run just great. Both of my 486 motherboards are opti chipset and I love them. Very stable.

OPTi's best days were probably back when they were making 386 chips (Anyone with a 381 or 391 that can confirm?). Their 486 stuff was *okay* until PCI came along. Everything after that was crud. OPTi made about a bazillion different chipsets, but I can't seem to recall any that were outstanding performance wise. I've had a number of OPTi based boards. Two based on 495SX, a 295SX, a 572, a 683 and a 693. Of these, only the 295SX really impressed me (and it's on an SLC board, which was a budget system back in the day).

Based on what I've used, I'd much rather have SiS, UMC, Forex or possibly even ALi.

Iirc, you've had a couple Pentium Opti-chipsetted boards some years ago and used them somewhat extensively.
I once had the chance to snag such a UMC Pentium board (had some Viper chipset? Not sure it had VLB though but it was definitely Viper...something) and I didn't bother because of what I'd read about it's terrible performance. Oh well, can't have it your way all the time 🤣

Got this little beasty today 😀

Welcome to the club.

Have fun watching it toast Intel's offerings in various applications.

On the OPTi discussion. My 307H is 319 based, never had any problems with it... I do think it has a minimum wait of 80ns for RAM access though, but that doesn't really mean anything and I have no idea about the other vendors chipsets.

HighTreason wrote:

Welcome to the club.

Have fun watching it toast Intel's offerings in various applications.

On the OPTi discussion. My 307H is 319 based, never had any problems with it... I do think it has a minimum wait of 80ns for RAM access though, but that doesn't really mean anything and I have no idea about the other vendors chipsets.

Yep, I already was impressed by these UMC CPU's, a pal has a surface mounted one in a PC-CHIPs VLB board, it goes really well. Here is a pic of his CPU, also a 40MHz part.

Is there any difference in these parts apart from the packaging?

Best,
Chris

Have never used the PQFP version, but I believe it is the same exact core. The CPGA U5SX and U5S are identical anyway, so I would imagine the PQFPs sharing these model names would be the same as well. The only ones which are any different are the LV versions (U5SFLV for examply) - these were for low-voltage 3 Volt systems.

Cool, be nice to find a U5D one day with the FPU, not that it makes a fat lot of difference from this era. Is it just me or does the "40" on my ceramic chip look like it was tacked on later as an afterthought/re-mark?
Bit of interesting history on my paticular chip, the original owner purchased it from Golden Arcade, Hong Kong in 1994, used it for a year then has been in storage since here in the UK. Was a "best offer" bargain for me from ebay. 😀

Best,
Chris

I am curious - Why weren't / aren't these UMC CPUs allowed to be sold to the US? Something to do with patent infringement?

@BSA Starfire; Yeah, I want one of those though every time I've almost had one, it has been swiped from under my nose by the furious computer gods or something... I dunno, something always goes wrong, though I have at least been refunded every time up to now.

Some of them were relabeled. I do not know why, perhaps they changed their quality control or their silk screening process was broken somewhere.

@Brickpad; Supposedly so. I don't know which patent exactly, given the internal workings appear to be very different to those of an Intel processor. I guess they didn't have a license to the arc though.

Perhaps as they refined the manufacture they realised they were OK at 40 MHz as well? Looking forward to giving this CPU a test after it's 20 year sleep!

Brickpad wrote:

I am curious - Why weren't / aren't these UMC CPUs allowed to be sold to the US? Something to do with patent infringement?

Intel won a case because they proved the chip infringed on the 486 patents.

HighTreason wrote:

@BSA Starfire; Yeah, I want one of those though every time I've almost had one, it has been swiped from under my nose by the furious computer gods or something... I dunno, something always goes wrong, though I have at least been refunded every time up to now.

Some of them were relabeled. I do not know why, perhaps they changed their quality control or their silk screening process was broken somewhere.

@Brickpad; Supposedly so. I don't know which patent exactly, given the internal workings appear to be very different to those of an Intel processor. I guess they didn't have a license to the arc though.

Ah! the furious computer Gods, that why I can't get a 2.2v Cyrix MII, RiSE MP6 or a Renditon Verite card, mind you, if I did get those only the really impossible parts would be left like the NexGen 5x86 etc...... 🤣