analog_programmer wrote on 2024-10-11, 07:50:

Unnecessary fake cache chips as L2 cache is integrated into CPU:
1. PCChips/Hsin Tech M666 (there are some plastic fake cache "chips" on the board, but L2 cache is actually integrated in PII Xeon CPU)

I don't know which is more shocking to me.

1) PCChips actually sold a Slot-2 dual Xeon motherboard

2) Somebody who was deep enough into PCs to actually be in the market to buy a Slot-2 dual Xeon thought it should have external cache chips on it.

EDIT:
are those 2 chips between the CPUs actually fake cache chips? The picture isn't detailed enough to see what they say.
I know the TRW page says it has fake cache, but it looks like it's listing every possible "caveat" they have in their database, including references to Ni-Cd batteries, tantalum caps, and the i820 chipset which doesn't agree with the picture. Looks like a glitch.

You can see that the chips on the M666 are those I’ve just posted for the M919. They say “write back” on them.

shamino wrote on 2024-11-06, 08:16:

I don't know which is more shocking to me.

1) PCChips actually sold a Slot-2 dual Xeon motherboard

2) Somebody who was deep enough into PCs to actually be in the market to buy a Slot-2 dual Xeon thought it should have external cache chips on it.

I'm in favour of:

3) Every moron in ECS/PCChips marketing department thought, that somebody who was deep enough into PCs to actually be in the market to buy a Slot-2 dual Xeon thought it should have external (L3 ???) cache chips on it. (For a very long time engineers don't have the final word on what and how to be produced.)

shamino wrote on 2024-11-06, 08:16:

EDIT:
are those 2 chips between the CPUs actually fake cache chips? The picture isn't detailed enough to see what they say.
I know the TRW page says it has fake cache, but it looks like it's listing every possible "caveat" they have in their database, including references to Ni-Cd batteries, tantalum caps, and the i820 chipset which doesn't agree with the picture. Looks like a glitch.

Yeah, there's a ton of not so correct or verified info on the TRW site. But after some rotating and zooming, the picture is "clear" enough to see "WRITE BACK" molded inscription - check it out what it means 😉

Back in the days my best friend and I both had almost identical 486DX2-66 setup, with the same amount of cache or so we thought... But we could not understand why my computer was systematically faster in all benchmarks. It's only thanks to this site that years later I got the explanation: his computer had fake cache chips!

Retrospectively it is really sad because it's not only the manufacturer to be blamed, I am pretty sure the entire supply chain down to the local computer shop at the corner of the street who knew what was going on, but still preferred to sell the mobo and get the dirty money.

Turns out Octek wasnt alone! Found another vendor supporting EDRAM in a "History of the DRAM" document https://site.ieee.org/pikespeak/files/2020/08 … untain-DRAM.pdf

"Ramtron Developed a Custom Chipset Based Upon Intel FlexLogic FPGA", "Won First Taiwan Motherboard Design with Digicom"

Sure enough "QUICKCACHE EDRAM, MEMORY USING EDRAM ONLY" sticker over appropriately sized chip https://theretroweb.com/motherboards/s/digicom-digis-486el
Silcon-Mountain-DRAM.pdf has picture of another Digicom board with this additional Intel FPGA

Btw Iv never ever saw any gear with Intel FlexLogic FPGA, Intel FPGAs totally failed in the market in the nineties.

Shame on PCChips.
I had a M577 at the time with K6-II then upgraded to K6-III and the system always seemed to lack power. I wonder if the mobo cache was also fake. In case of fake chips, are they still detected by benchmarks and system test softwares?

You do realize that the M666 is an easter egg on TRW that doesn't actually exist, right?

Xanxi wrote on 2024-11-15, 21:35:

In case of fake chips, are they still detected by benchmarks and system test softwares?

Sure, you can use some system diagnostic software that can test/report the motherboard's cache.

watson wrote on 2024-11-15, 21:39:

You do realize that the M666 is an easter egg on TRW that doesn't actually exist, right?

Nope, but I thought someone overdid it with the board's defects list.

I don't know what kind of joke this is, as if the remaining inaccurate or misleading information on the website is not enough.

So, this fake board TRW article reminds me of another "phantom" motherboard. Has anyone heard of PCChips M569 mobo? I can't find a trace of such model all over the internet, except that rmay635703 mentions it a couple of times in other forum's topics and also in some other retro PC forums.

thanks