Super Socket Seven

[banjosteve]

Hi there

Do you think we would ever get to see some super socket 7 CPU's? In theory they are just up spec'd Pentium 1 with more cache and fast FSB.

Admittedly there are the 3DNow instructions from AMD, but lets be honest, how useful were they really.?

Anyway keep up the amazing work

[leilei]

In theory they are just up spec'd Pentium 1

if only.........


We haven't even got to the other non-super Socket 7 CPUs yet

Admittedly there are the 3DNow instructions from AMD, but lets be honest, how useful were they really.?

They're very useful with official 3dfx Voodoo2/3 drivers since they have an optimized 3Dnow! path

[banjosteve]

ok so a k5 is more a Pentium pro in its architecture and a k6 is really a re-badge of NexGen Nx686 after they sucked it up and is really a risc chip which converts it to x86 so maybe these are kinda off the radar. But the cyrix 6x86 is very close in design to the winchip stuff which pcem does support, it just needs some tweaking with its cache sets and other areas to get to.

Socket 7 is easy in one way as its the same as it supported all the same processors that socket 5 supported its more to do with cyrix and AMD stuff.

Super socket 7 ok lets look at its basic's is really all about the higher FSB and the AGP support. Admittedly the AGP side of it looks like the biggest pain to get sorted. Then the CPU's roll in the cyrix and win chip stuff are just higher clocked and higher FSB of the socket 7 versions with only minor tweaks.

The fun comes with the k6-2 and upwards CPU's and this might be were the real hard work kicks in and these are completely clean sheet based designs

The 3Dnow side of things ok sounds like it could be advantageous, i admit that during this time i was and still am an Intel fan, but i did have a k6-2 400 as a second machine during this period. So never really used it in gaming anger.

May be the next step after bottoming out the current crop of intel pentuim 1 cpu's is getting a pentium pro ticking? As the step on to a pentium 2 is not that much and the pentium 2/3 era boards still had versions with just PCI on them so we could skip the whole AGP thing or at least put it off to a later date...

Anway it might be worth having some form of long tail plan as this always helps shake these things out..!?

[leilei]

I've only worked on a tiny bit of groundwork for Cyrix 6x86 support. emphasis on tiny

as in it's just the identification stuff and some cycle counts. No cyrix CPUs doing anything and they can't make it to POST. I can't work any more on this because A. cpu code scares me 2. critical art priorities
and even then from there, it's a long road to Cyrix6x86. That's a Pentium II-class CPU if you feed it only integerhappy stuff, i.e. not Quake and it has a lot of special instructions that were never used (or even enabled on the real thing).

[banjosteve]

there is a known problem with the cpu id's on real 6x86 cpu's and this boiled down to it only working on some types of bios's. Also the 6x86 was really socket 7 only so do we have a bios that at least come to life running this CPU?

I hear you on art work, i run a UX team so i have to deal with designers and UX'ers so its all good fun..lol

[leilei]

The 430VX BIOS used does recognize the Cyrix6x86 CPU, but doesn't make it past the IDE detection. Probably needs some special Cyrix check that's not emulated.

[banjosteve]

found some useful little bits of info around the chip...

6x86
(1997-1998, 0.65µm-0.35µm process, 120-188MHz)

Cyrix does get a mention in our little history for implementing the 75MHz bus on its 6x86 and 6x86MX processors. This meant the PCI bus ran at 37.5MHz, which caused problems with some add-in cards. Cyrix caused even more problems in 1998 when it upped the speed of its MII chip to 83MHz, forcing a completely unreasonable 41.5MHz on PCI. As a result things were unstable, particularly IDE controllers.

[banjosteve]

could we try using a SiS 5571B or a TX based one? these are supposed to be more stable with 6x86 CPU's and support the freaky bus speeds better

[omarsis81]

could we try using a SiS 5571B or a TX based one? these are supposed to be more stable with 6x86 CPU's and support the freaky bus speeds better

Cyrix 6x86 were not Super Socket 7.
Most common Super Socket 7 CPUs were the K6-2 and III, and most common chipsets used with those were the Ali Aladdin IV and VIA VP3
But I agree it would be very nice to have any of those emulated in PCem

[banjosteve]

Hi there

I think you will find that they were, specifically the Cyrix 6x86 M2 version.. Remember Super socket 7 was an extension to the old standard socket 7 by AMD as intel were being idiots with their slot 1 design as this was not open to AMD, you could put pretty much add any old socket 7 CPU in to a super 7 board.

The discussion centered on the SIS an TX based chip sets specifically around the Cyrix CPU's as leilei was having issues with the current bios versions the PCem supports, so maybe going for the new ones that support the odd bus speeds would help solve some of the issues.

I agree the ALI Aladdin were the most common of chips sets for the true super socket 7 CPU's but as it was stated more work needs to be done to get the standard P1 stuff running better..and to be honest its not far off now..

[SarahWalker]

FWIW the issue leilei was having with the 430VX BIOS was a CPU bug (I got one of the 'MOV TRx' opcodes wrong) and not really a BIOS problem.

[banjosteve]

Thanks for the clarification.. so any change of some Cyrix action anytime soon?

[SarahWalker]

Hopefully later today...

[banjosteve]

ohhh colour me interested!!