Page 1 of 2

Pentium 266 Tillamook on MVP3 board?

Posted: Sun Jul 04, 2010 6:48 pm
by DrSwizz
Does anybody know if it is possible to get a Pentium 266 Tillamook to run on a Epox MVP3C2 or MVP3G2 board?

I have tried booting with these CPUs, but I cannot seem to get it to work. The CPUs work just fine on i430TX boards, but the MVP3 doesn't seem to like them. Has anybody got any ideas of how I somehow could get it working??

RE: Pentium 266 Tillamook on MVP3 board?

Posted: Mon Jul 05, 2010 4:16 am
by Jim
I wish I could help, But I know nothing about either the processor or the boards mentioned.

Posted: Wed Jul 07, 2010 8:48 pm
by Uranium235
:shock: Never had my hands on the very rare Pentium 266 Tillamook CPU. My guess would be a BIOS issue is preventing this CPU from being recognized. There is only one person I can think of that could help you out, Jan Steunebrink.

Here is the link to Jan's website:
or email him:

Jan Steunebrink has been patching the BIOS's on socket 7 boards to work with the K6+ CPU's for years. Maybe he could create a patch for the Tillamook to work on the Epox boards.

Good luck.

Posted: Thu Jul 08, 2010 2:16 pm
by jsc1973
There's no reason any Pentium shouldn't boot on that board, unless it isn't being recognized. There's probably no support in the BIOS for Tillamook, presumably since no one at VIA would have figured someone would try to boot one on an MVP3 desktop board.

If anyone could make it work, it would be Jan Steunebrink. My guess is that someone just needs to put an entry in the CPU table for it.

Posted: Sat Jul 10, 2010 4:21 am
by stevenaaus
Yeah.. The above post pretty much sums it up

I don't know this mobo, but have you futz-ed around with mobo's cpu voltage ? Some boards have automatic and manual voltage detection, and setting it manually will overcome board failing to detect correct CPU. Even try .1 || .2 volts extra in case board isn't supplying enough volts... Unlikely , but possible.

Posted: Mon Feb 28, 2011 5:31 pm
by rmay635703
stevenaaus wrote:Yeah.. The above post pretty much sums it up

I don't know this mobo, but have you futz-ed around with mobo's cpu voltage ? Some boards have automatic and manual voltage detection, and setting it manually will overcome board failing to detect correct CPU. Even try .1 || .2 volts extra in case board isn't supplying enough volts... Unlikely , but possible.
This chip will not even allow any stage of post on a board with a chipset that doesn't support it.

Only chipset I know of that can use it is the Intel TX.

I tend to believe it was like the old amd 5x86 where one or 2 pins functioned slightly differently or the difference between a socket 5 and socket 7 chip.


Posted: Wed Mar 16, 2011 10:48 pm
by jsc1973
There came a point where Intel got tired of people using their motherboards and chipsets to run Cyrix and AMD processors, and they started making them refuse to boot with anything other than a CPU that was specifically listed in the BIOS processor table. Prior to that, an unrecognized CPU would usually be defaulted to "486DX" and booted with the BIOS support needed for such a CPU. This is why advanced features often didn't work without being enabled in software after booting, but in those days the hassle was a small price to pay to keep a costly motherboard in service. (Old third-party chipset motherboards often defaulted to "AMD-K5" rather than a 486.)

Intel never wanted people to run Tillamooks on the desktop and simply never supported it. The only reason they even produced them was because they had nothing in the low- to mid-range laptop market, and were losing market share to the perfectly capable mobile K6. Rather than run a Tillamook, they wanted you to buy an overpriced Pentium II on a slow 66 MHz bus, and an even more overpriced motherboard to run it on. When you were done, you had an expensive computer that more often than not was slower than a comparable (and cheaper) K6 or even Cyrix.

Note that the same scenario played out in 1999, when AMD had nothing to service that laptop segment except outdated mobile K6-2s. Enter the K6-2+ and K6-III+. The difference is that AMD let anyone use them on whatever system they wanted to.

But if you could hack the Tillamook into the BIOS CPU table, I suspect it would boot just like an ordinary Pentium MMX.

Posted: Thu Mar 17, 2011 7:23 am
by DrSwizz
Actually at first my two 430TX motherboards had old BIOSes that did not recognize the Tillmook (the CPU was treated as an ordinary Pentium MMX CPU). The Tillmook seemed to work just fine despite that.

I have also got the Tillmook to work on one 430HX based motherboard. For some reason this board automatically disables the L2 cache when a Tillmook is used though (the Tillmook does have a larger L1 cache then the ordinary Pentium MMX, so there might some sort of incompatibility there).

Anyway, I have more or less given up on this now as none of the motherboard the Tillamook works with can provide the CPU with voltage that is low enough.

Posted: Fri May 06, 2011 3:42 pm
by Tetrium
I think this is an interesting topic. Once I have done some tests with a Tillamook, I'll report back here or will report on Vogons ;)

Sorry for the thread rez with little extra info to spare, I've (hopefully) got 2 Tillamooks on the way here now and plenty boards to test them in.

Posted: Mon May 30, 2011 2:25 am
by moondog

this was a topic on several german forums before, too... they were trying to get a tm266 to run on a mvp3 board and it wasn´t running, the problem shouldn´t be only the missing microcode and bios support but also the 2,5v IO, too. Finally I wonder about the reason to get this cpu to run, as it offers only little advantages... I overclocked a normal 233mmx to 300 mhz on my mvp3g5 some years back and it was running well and stable at 3x100@2,9 volt... have to search for my old benchmarks (I kicked about 30 different cpus of my socket7 collection through a benchmark parcour some years back, incl. overclocking...) as it was outperforming nearly all other socket 7 cpu´s at this speed, even the k6-3 400 had a hard stand against the overclocked mmx...



Uuups... have to admit that I have to correct myself... I stumbled over several threads on the german forums hardwareluxx and planet3dnow where some guys had running the tillamook one some boards, even overclocked to 400 mhz... as far as I could read the tillamook should run on the ga5ax and ga5aa (someone claimed that he even reached 460 mhz on this one...) unfortunately the thread ends as someone wanted to check out the tillamook on his epox mvp3g5, I guess he failed...

Posted: Sat Jun 04, 2011 5:42 am
by Tetrium
On Vogons someone also managed to get one running, but was unable to enable the motherboard cache, and no point in running 400Mhz without any L2 cache, right?

From reading some of the threads on a couple other forums, I got a hunch that setting the voltage higher "might" cure the "No L2 enabled" problem

Posted: Thu Jun 09, 2011 2:55 pm
by jsc1973
I saw that on Vogons, too. I'd like to see what a properly cached Tillamook could have done at 400 MHz. That's probably about the spot where it would max out in performance without suffering the effects of an L2 bottleneck, like the K6-2 does.

Posted: Thu Jun 09, 2011 5:05 pm
by DrSwizz
I got my hands on a QDI MVP3-based board a few days ago and today I decided to test one of my Tillamook CPUs with it.
This board could actually boot with the Tillamook CPU installed, but to get past the memory test & BIOS initialization a had to disable the L2 cache. Running the CPU at 400Mhz seems to be no problem though.

Unfortunately the QDI "speed easy" BIOS offers very limited voltage settings with the Tillamook installed so I will not be using this CPU/motherboard combination for anything other than some testing & benchmarking.

Posted: Wed Jul 06, 2011 3:20 pm
by Tetrium far it appears theres no solution for the Tillamook issue.

We did find some info about Tillamook in it's datasheet. Apparently some of the pins were disabled compared to it's older desktop brother.
Perhaps this has something to do with it?
I find it hard to believe Intel put somekind of microcode into Tillamook preventing it from working with non-Intel board L2 cache so perhaps Tillamook won't work with boards that cache more then 64MB memory, or have more then 512KB cache??

Posted: Tue Jul 12, 2011 5:58 am
by DrSwizz
I was looking thru some old documentation I have downloaded some time ago when I happened to to come across a paper titled "Mobile Pentium® Processor with MMX Technology on 0.25 Micron System Design Considerations". It describes what changes that are needed to adapt systems designed for Pentium MMX to be able to use Tillamook CPUs. Intel states that they made some changes in how the CPU interfaces with the L2 cache. I suppose it is these changes (whatever they are, the document is not very technical) prevents us from using the L2 cache.