SiS 730s chipset - can it change multiplier?

Recommended by Wim's BIOS...

I got a PC chips M810S v7.0a board as dead, managed to fix it, even it shoot caps at me back, recapped it and now I want let it fold. The problem - the bios show NO WAY to change multiplicator above the default 11 for my oldie unlocked Barton 2500+.
Also, the max FSB is 133Mhz.
This give only 133x11 - 1463Mhz. A way too low for the nice there phase voltage regulator and quality Samxon GC caps I used for the recap. Foding run very slow on this one... Also the Zalman CNPS 7000Cu is entierly useless there as well, as the cool Revoltec heatsink I added on the chipset.
So, the question is - can the SiS 730s chipset change multiplicators in AXP cpu's at all? I have oldie unlocked Barton - not mobile - and there is no multiplicator option in bios (as well, as no voltage option).

And, the cursed AmiBCP keep saying that the bios is invalid, tough I flashed it to the board already... (the bios file start is a bit weird, too much empty, and then code...)

Anyway, the question is, if I can hope that the chipset CAN change the multi at all. If not, then there is no point of bothering with bios modding, even the bios is VERY spartan and comparing to ESC P6STMT bios with million and one timing option and settings this one suxx...

Bios link there, in case anyone care:
http://www.pcchips.com.tw/PCCWeb/Downlo ... lDesc=M810

[cp]

uhhhhh...(!) with the M810 you have to be grateful that it actually runs stock speed! really, this board is the cheapest ever seen with socket 462 and you shouldn't expect anything from it. OEMs used those in very cheap systems. there are no such things as multiplier changing or other fancy overclocking options. you can however try to change the multiplier on-the-fly or just alter the startup multiplier. you might read more about the M810 on this great site. there are modified bios versions out there that allow overclocking the fsb but you have to know that the chipset wasn't build to stand more than 133MHz. i would be very cautious with bumping that fsb.

just for your information: changing the multiplier is not a feature of a chipset. it's a feature of the cpu.

Thanks for swift response. I recapped the board (new caps) so it working (will working, CPU/RAM/chipset is stable, folded for a week but not keyboard/mouse because of the fu*king small 22uF 16V Teapo caps witch I did not exchanged yet, all rest exchanged for HI QUALITY Samxon caps with even slight bump in capacitance), you can check first photos there:
http://trodas.wz.cz/index.php?act=ST&f=16&t=337

Therefore is not unrealistic expect the board will run stable at 150, maybe more Mhz FSB (depends on the two 512MB modules) and at around 2Ghz clock of the Barton.
Even w/o cooling on chipset it run cool and I added the cool Revoltec heatsink, so...



Other than that, I got the mobo nonworking and I fixed it and I expect a lot of folding from it And yes, there is absolutely nothing in the bios. I could only go for 100x11 or 133x11. Both suxx, but at least 1400Mhz of folding power, yep?
On the "on-the-fly" list there is not any SiS 730 chipset listed, so this hardly could be a option. But the moded bios, the last one, the UNIBIOS on the page (a big thanks!) looks promising. I wish my keaboard works, and te Teapos make it and I can flash it, but I can't ATM. Hope very very very soon
And I think - given the very very cold temp of the chipset, the quality heatsink, the CNPS 7000 Cu Zalman bloving air partly on it and the hi-quality caps used - overclock will be w/o a glitch done

And I know that multiplied is a CPU thing, but the board did not looked that it can change it, so I start remembering that these first AXP boards did not know that and there was sort of hype around nF2 chipset that can change the multiplicators, so I starting to dubt if I can change the multi on unlocked Barton on SiS chipset or not...

Maybe this could help anyone with locked Barton too?
http://www.lostcircuits.com/cpu/amd_unlock/

edit: fixed last link

[cp]

i'm perfectly aware of the caps problem. besides i've had a few of those M810 boards myself and i wouldn't recommend them to anyone who want's a board that doesn't make trouble ;) to push that board beyond the BIOS offered speeds you could reprogram the pll clock generator that sits on the mainboard (the one in your 5th picture, the one with the ICS marking..sorry, i forgot the model). reprogramming can be done via the I2C bus with some programs.
but as you want cpu core speed you could also mod your barton to mobile and change the multiplier as well. that would save you trouble regarding the memory and the chipset. the reason why you didn't find any entries of the sis730s in the list over that site is probably that noone ever tried on an old board like that. i think it's worth a try :)

[Denniss]

At www.rhcf.com you may find some nice M810 Overclocking Bios by HoneyX.
At www.ocinside.de you may some nice ways to change your CPU multiplier via wire-mod inside the CPU socket.

Then you should know, that after complete exchange of the caps (cursed G-Luxons and even more eneaky evil Teapos) the board should kick back and enjoy great stability and very high reliability, as my download server with SiS 630 does.

Now I quote what I write on BadCaps about getting the board to life:

***
]Some time ago my friend Morphy give me this board for experiments and components, maybe if I could repair it? It is a bit older one, SDRAMs and 100 or 133Mhz FSB only, but it should take AXP cpu's since SiS 730s chipset support AXP. I mainly love on the board the nicely done 3 phase Vcore
The board however did not kick in, no matter what my stepbro did or did not, so it was pronounced dead until I noticed that someone made scratch on the mobo that damaged the PSB traces, so it was immediatelly clear to me, why it did not psoting at all. The traces between chipset and ram are important and mobo will not post w/o two or more of them at all.
After reasonably hard repair work I got lucky and these traces looks like pretty well repaired to me. It is enought to patiently with the tip of blade scratch the paint from the trace big as thicker dog chair and when it shine with the cooper, then solder, solder, solder it
I was happy how the repair went well, so I pushed my stepbro to give her another try. And you will never quess what happend next. When I power up the mobo then the room was shocked (and mainly my stepbro, who run away from the board to the other end of the room) by lound crack sound and from the bottom of the there primary Vcore pre-filtering 5V rail line caps flew into air about 50 cm high column of internals of the poor G-Luxon cap.
It was so loud, that my hearing get partialy noticably lowered for some time, it was like a gunshot! :



Dear G-Luxon went open, as you can see. Luckily the testing unlocked oldie AXP Barton 2500+ survived this, as well as everything else, even it took a while till the in shock turned this thing off.

Conclusion from this is - if you see even a little bit bulging G-Luxon, expect a loud booom. I know that already, this is how my Chemi-con KZG on my DFI board went to hell...

Another point - in our todays overscarred world from "terrorists" - what do you think it happen, when cap with such loud boom explode into your notebook when you fly into commercional jet? Hell, you might even get shoot :

***

And the bios promise overclocking way past 133Mhz. There is mentioned even 166Mhz option, witch I see as possible only if the ram's holt up, witch I dubt.
And yes, for folding is only important CPU clock. I measured it by some testings years ago and we come to conclusion that FSB as well, as memory timings did not affect the speed. Well, of course it does, but it was just a seconds per WU 500 fractions and these 500 fractions took sometimes day and half on 1,4Ghz Barton we talking now, so, no deal there. Yet I still would like to see 150Mhz. At least 140Mhz must the rams give, damn them (I'm mad overclocker, hehe)
My Barton is oldie one, unlocked from the start. He does, tough, have a wall at around 2250Mhz witch can be broken only with 2.2Vcore (you can't say I did not tried hard-enought, can you?) so I settle with 2100Mhz that it need only 1.850Vcore.
(just for comparsion, my 2600+ Mobile never need to run bellow 2500Mhz - it need just 1.725Vcore for that - with this CPU my overclock STARTED at 2500Mhz )

So for this one I did not feel the need connect these pins, but I do it anyway as then any inserted barton become unlocked and therefore highly overclockable.
The problem is, that 2100 / 150 = 14. Purrrfect
The question is, if the mobo can do 5-bit multiplier change. The obvious answer is - nope. So I'm probably stuck here with lower multis, up to 12.5 only.
150 x 12.5 is only 1875 - that suxx, but it is possible that this the oldie Barton 2500+ can made w/o increasing the voltage at all and I have no damn idea if the moded bios allow me any voltage changes at all...
And you right - it is worth a try. At worst it can crash - big deal

I love to play with old hardware. The only catch is getting the god-damn caps to repair them. At least I have reliable source for higher capacity low ESR Samxons. But in repairing a PSUs and stuff I ending up at mercy of others living in USA or Canada and I think the latest "friend" in Canada just con me... (package is strangely delayed, no answer when I asked for scan of the receipe...)

[cp]

mod your barton to mobile and write the MSRs directly. no need to do any hardware mods on the board. use CrystalCPUID .