VOGONS


Is this Cache?

Topic actions

First post, by Iris030380

User metadata
Rank Member
Rank
Member

Hi guys,

I found some cache on a stick. Not sure how much cache is on there though.

Anyone know what it is?

Written on it are :

IDT 7MPV6203A REV00

And also : 64KX 64SRAM
IDT 702

11540H8146ZJ111772D2DB

It's double sided, and on each of the 4 chips, it says :

IDT
71V432
S7PF
Z9652P

I5-2500K @ 4.0Ghz + R9 290 + 8GB DDR3 1333 :: I3-540 @ 4.2 GHZ + 6870 4GB DDR3 2000 :: E6300 @ 2.7 GHZ + 1950XTX 2GB DDR2 800 :: A64 3700 + 1950PRO AGP 2GB DDR400 :: K63+ @ 550MHZ + V2 SLI 256 PC133:: P200 + MYSTIQUE / 3Dfx 128 PC66

Reply 1 of 20, by DerBaum

User metadata
Rank Oldbie
Rank
Oldbie

It should be a 512kb (Kilobyte) Module. 32x32kbit (per chip) x4 (chips) / 8 (a byte are 8 Bits) .
If my math is wrong please let me know how to calculate it right.
Datasheet for the chips: https://www.renesas.com/us/en/document/dst/71 … uage=en&r=13441

FCKGW-RHQQ2

Reply 2 of 20, by pancakepuppy

User metadata
Rank Newbie
Rank
Newbie
DerBaum wrote on 2022-07-22, 18:53:

It should be a 512kb (Kilobyte) Module. 32x32kbit (per chip) x4 (chips) / 8 (a byte are 8 Bits) .
If my math is wrong please let me know how to calculate it right.
Datasheet for the chips: https://www.renesas.com/us/en/document/dst/71 … uage=en&r=13441

Your math is right on. There's a datasheet for an older revision of this cache stick series from IDT and it specifies the IDT7MPV6203 to be a 512KB Pipelined Burst module for Intel 82430 Family Core Logic PCISets (probably FX/HX/VX? I still haven't been able to figure out the particulars of COAST versions.)

Reply 3 of 20, by Sphere478

User metadata
Rank l33t++
Rank
l33t++

Yeah, as said above, 64x64 divided by 8 is 512

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 4 of 20, by TrashPanda

User metadata
Rank l33t
Rank
l33t

Ok now this might be left field but it well has to do with the question asked.

s-l1600.jpg
Filename
s-l1600.jpg
File size
564.06 KiB
Views
1368 views
File comment
Identify Me PLS
File license
Fair use/fair dealing exception
s-l1600 (1).jpg
Filename
s-l1600 (1).jpg
File size
387.47 KiB
Views
1366 views
File comment
Better Cache shot
File license
Fair use/fair dealing exception

These cache chips on this tiny mini board . .am I right in thinking they are in fact the wonderful fake cache ICs ?

I also have no idea what this board actually is other than it was made by HTE in December 1995, if anyone has an idea and wants to share it that would be great.

I was considering getting this board due to its compact size and that lovely looking TI 486 DX 100, its smaller than a baby AT board and the lack of VLB doesnt bother me since it has PCI.

Last edited by TrashPanda on 2022-07-22, 22:06. Edited 1 time in total.

Reply 7 of 20, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

PCCHIPS were sold under different brands, but all can be easily identified by revision number ("Vx.x") near a keyboard connector.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 8 of 20, by TrashPanda

User metadata
Rank l33t
Rank
l33t
The Serpent Rider wrote on 2022-07-22, 22:11:

PCCHIPS were sold under different brands, but all can be easily identified by revision number ("Vx.x") near a keyboard connector.

Thanks for that, now I know what I need to look for !

Shame about that board its a compact little number.

Reply 9 of 20, by DerBaum

User metadata
Rank Oldbie
Rank
Oldbie
TrashPanda wrote on 2022-07-22, 22:04:

These cache chips on this tiny mini board . .am I right in thinking they are in fact the wonderful fake cache ICs ?

Google gives exactly 0 results when searching for "16c328ah" ... Thats not a good sign.
And the "WRITE BACK" makes it look even more fake. It feels like "Hey look at these chips... its totally Cache and no fake stuff! Trust the text!"

I would put them in my collection. I like the weird stuff.
Changing the chips for real ones shouldnt be impossible... (If the board even works with real cache)

FCKGW-RHQQ2

Reply 10 of 20, by TrashPanda

User metadata
Rank l33t
Rank
l33t
DerBaum wrote on 2022-07-22, 22:33:
Google gives exactly 0 results when searching for "16c328ah" ... Thats not a good sign. And the "WRITE BACK" makes it look even […]
Show full quote
TrashPanda wrote on 2022-07-22, 22:04:

These cache chips on this tiny mini board . .am I right in thinking they are in fact the wonderful fake cache ICs ?

Google gives exactly 0 results when searching for "16c328ah" ... Thats not a good sign.
And the "WRITE BACK" makes it look even more fake. It feels like "Hey look at these chips... its totally Cache and no fake stuff! Trust the text!"

I would put them in my collection. I like the weird stuff.
Changing the chips for real ones shouldnt be impossible... (If the board even works with real cache)

I couldn't identify a PCCHIPS board till this thread but even I know that no replacing the chips will do exactly zip, not only did they put fake chips on the board they are not even connected via traces to the CPU on teh vast majority of the boards I have heard stories about. But I now know to look for the VX XX marking and will in future avoid these boards, as for adding it to my collection .. if it was considerably cheaper than it is .. I might just because I too have a thing for odd hardware and it should be a right of passage for every collector to have at least one PCCHIPS board.

Instead I have a possible replacement in a ZIDA mini board based on a SIS chip set ..dunno yet as it honestly doesnt strike me as a good looking board.

-Christ my spelling sucks after a 12 hour shift.

Reply 11 of 20, by DerBaum

User metadata
Rank Oldbie
Rank
Oldbie
TrashPanda wrote on 2022-07-22, 22:42:

...but even I know that no replacing the chips will do exactly zip, not only did they put fake chips on the board they are not even connected via traces to the CPU on teh vast majority of the boards I have heard stories about.

I thought to see traces between the chips on the first picture. this is why i mentioned a swap.
It also looks like it was prepared for even larger cache chips (4 unconnected holes above the chips).
But i have exactly zero experience with pcchips boards.

FCKGW-RHQQ2

Reply 12 of 20, by Iris030380

User metadata
Rank Member
Rank
Member

SHOCKING I have zero notifications have to do the hard workmyself of returning to the this SHEER TUMROIL of off piste discussion! BUT AT LEAST I know now, that I have 512KB COAST. With a IDT brand on it. USELESS that it is to me, entirely. But it looks cool!

Maybe I'll keep it. But if someone needs it, they can have it. Cos I certainly got no use for it.

I5-2500K @ 4.0Ghz + R9 290 + 8GB DDR3 1333 :: I3-540 @ 4.2 GHZ + 6870 4GB DDR3 2000 :: E6300 @ 2.7 GHZ + 1950XTX 2GB DDR2 800 :: A64 3700 + 1950PRO AGP 2GB DDR400 :: K63+ @ 550MHZ + V2 SLI 256 PC133:: P200 + MYSTIQUE / 3Dfx 128 PC66

Reply 13 of 20, by TrashPanda

User metadata
Rank l33t
Rank
l33t

Off post .. a question was asked, you got an answer to yours and since the question also applies to my situation I simply used this thread rather than start a new one asking the exact same question.

This is how forums work and I believe the mods do prefer to keep similar questions to the same thread rather than making multiple ones.

Reply 14 of 20, by Iris030380

User metadata
Rank Member
Rank
Member

So does anyone require it? Be a shame to throw it back in the box. Might be useful to someone.

I5-2500K @ 4.0Ghz + R9 290 + 8GB DDR3 1333 :: I3-540 @ 4.2 GHZ + 6870 4GB DDR3 2000 :: E6300 @ 2.7 GHZ + 1950XTX 2GB DDR2 800 :: A64 3700 + 1950PRO AGP 2GB DDR400 :: K63+ @ 550MHZ + V2 SLI 256 PC133:: P200 + MYSTIQUE / 3Dfx 128 PC66

Reply 15 of 20, by TrashPanda

User metadata
Rank l33t
Rank
l33t
Iris030380 wrote on 2022-07-22, 23:17:

So does anyone require it? Be a shame to throw it back in the box. Might be useful to someone.

Just a thought but Coast modules can be hard to come across especially 512k ones and since you have that one it might be worth holding on to it should you ever acquire a motherboard that can use Coast modules.

I mean if you are never going to use it then im sure someone here may be interested.

Reply 16 of 20, by Errius

User metadata
Rank l33t
Rank
l33t

That's a PCChips M921. I have one as the basis of a DX2-66 build. I like it a lot even with its derpy fake cache. With 8 MB RAM, it's too slow for Windows 95C, but is a great little DOS machine. It's the only machine I have that runs Dune II without audio problems.

Is this too much voodoo?

Reply 17 of 20, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

in a ZIDA mini board based on a SIS chip set

Zida 4DPS is practically LuckyStar LS-486E clone, with minor slot difference, which has good documentation and one of the top boards for overclocking.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 18 of 20, by TrashPanda

User metadata
Rank l33t
Rank
l33t
The Serpent Rider wrote on 2022-07-23, 08:38:

in a ZIDA mini board based on a SIS chip set

Zida 4DPS is practically LuckyStar LS-486E clone, with minor slot difference, which has good documentation and one of the top boards for overclocking.

Might be worth taking a second look at, wonder how it would go with a Am5x86-133 in it, pretty much a souped up dx4.

Would be a fun build to oc.

Reply 19 of 20, by mkarcher

User metadata
Rank l33t
Rank
l33t
DerBaum wrote on 2022-07-22, 22:33:

And the "WRITE BACK" makes it look even more fake. It feels like "Hey look at these chips... its totally Cache and no fake stuff! Trust the text!"

Indeed. Cache chips labelled "WRITE BACK" are fake. No exceptions. The choice between write through and write back cache is performed by the cache controller. The cache memory is exactly the same type in both cases. Memory Chip companies just write their name, the model number, the production date and possibly a batch code on their chip. Only the marketing people that designed fake cache chips got the crazy idea to label them "WRITE BACK".

On the other hand, Cyrix 486 processors having a "Cyrix FasCache" logo, with or without the addition of "write back" are legit. And they all support L1 writeback internally (even the ones not mentioning write back), it's up to the BIOS to enable it and up to the chipset to deal with it, though. If the chipset doesn't support writeback L1 cache coherency protocols, the Cyrix processors run in write-through mode, which is good enough in most cases.