Archived from groups: alt.comp.periphs.mainboard.asus (
More info?)
Yes. SL726 sounds like a fun processor for the system-builder's version
of drag racing.
I acquired an Abit VT7 Via PT880 P4 800FSB board as a chassis. I had
read that Abit's BIOS features make overclocking pretty simple, and this
was the cheapest Abit socket 478 I ran across on ZipZoomFly. I like to
go cheap with things I plan to set on fire.
I also got some Ballistix PC4000 RAM. My overclocking-newbie instincts
told me it might come in handy. I'm not sure if this memory can get up
to 600MHz, but I'll see. At least it has some copper on it.
As you referenced, abxzone (ya gotta love bigtoe) seems to be all about
FSB as opposed to multiplier. Sounds like the multiplier is rather hard
to get at unless one is an electrical engineer with time on his hands.
But 12 x 300FSB, well, that would be nice to have.
I bought my chip by mistake too. I was looking for the P4 533FSB with
Hyperthreading, a la SL6S5. Got this instead.
But it certainly has tweaked my interest in OC.
Cheers,
Toney
Paul wrote:
> In article <j2I2e.610$Fh4.149@fe2.columbus.rr.com>, Toney
> <tjone12@indy.rr.com> wrote:
>
>
>>Thanks very much. As usual, your articles are all light and no heat.
>>
>>I wonder if "CPU clock ratio" is included in the BIOS of this low-end
>>board for the benefit of people who know how to get around the
>>multiplier locking--by insulating particular pins or whatever it is that
>>they do.
>>
>>My processor is an sl726, and, indeed, there is no mention of
>>Hyperthreading in Intel's spec finder page for this CPU.
>>
>>I've never overclocked anything before, but I thought it might be fun
>>with this Mobile Northwood 3.06GHz 533 FSB.
>>
>>This CPU has a very high thermal spec. It idles at roughly room
>>temperature, even with an FSB setting of 158. It's clock ratio on the
>>P4GE-MX is reported by Linux as 12, and it operates at about 1.6 GHz at
>>the default 133 FSB.
>>
>>
>>Toney
>
>
> Oh, sweet. That CPU is a nice purchase. Now you need a good
> motherboard to go with that processor. Some people, for example,
> run that processor at 300MHz clock (FSB1200), for a Fcore of
> 3.6GHz, and run the memory at DDR600. The P4C800-E is a good
> candidate for some fun like that.
>
> You should visit abxzone.com and read about the adventures with
> the Mobile over there.
>
>
http://www.abxzone.com/forums/showthread.php?t=59931&highlight=sl726
>
> Your current motherboard doesn't have enough FSB for that
> sweet processor. Time for an upgrade
>
> Yes, the clock multiplier is 12. No matter which speed grade
> you buy. That is the main advantage of that processor, as
> it allows very high FSB operation. And, reports are, that
> it runs cool.
>
> I wanted one of those processors, but ended up buying a B
> stepping version by mistake. Mine won't go anywhere near
> 3.6GHz. (Canadian suppliers are too conservative to carry
> processor like that in stock.)
>
> The Mobile has Speedstep, which means there are two
> multiplier values in the chip. x12 is the power saving
> multiplier setting. 3066/133=23 is your full speed multiplier
> setting. According to this info in the Mobile Pentium datasheet,
> the GHI# signal controls the multiplier, and a mobile chipset
> has a signal to drive that pin.
>
>
http://www.intel.com/design/mobile/datashts/25302804.pdf
>
> "Enhanced Intel SpeedStep Technology
>
> The mobile Intel Pentium 4 processor, when used in conjunction
> with the requisite Intel SpeedStep technology applet or its
> equivalent, supports Enhanced Intel SpeedStep technology.
> Enhanced Intel SpeedStep technology allows the processor to
> switch between two core frequencies automatically based on CPU
> demand, without having to reset the processor or change the FSB
> frequency. The processor has two bus ratios and voltages programmed
> into it instead of one and the GHI# signal controls which bus
> ratio and voltage is used. After reset, the processor will start
> in the lower of its two core frequencies, the Battery Optimized
> mode. An operating mode transition to the high core frequency can
> be made by setting GHI# low, putting the processor into the Deep
> Sleep state, regulating to the new VID output, and returning to
> the Normal state. This puts the processor into the high core
> frequency, or Maximum Performance mode. Going through these steps
> with GHI# set high, transitions the processor back to the low
> core frequency operating mode. The processor will drive the
> VID[4:0] pins with the VID of the current operating mode and
> the system logic is required to regulate the core voltage within
> specification for the driven VID."
>
> Page 280 of this doc, shows the GHI# signal being driven by CPUPERF,
> and that is how a mobile chipset controls whether the multiplier
> is x12 or x23.
>
>
http://developer.intel.com/design/mobile/desguide/253026.htm
>
> Now, the quoted text above is a bit weird. That implies just
> jamming the GHI# signal to ground, won't necessary trigger the
> x23 multiplier for you right away. It takes a transition into a
> sleep state, to pick up the status of the pin. The GHI# signal is
> AGTL, and is supposed to be driven by an open drain driver coming
> from the Southbridge (ICH4-M). The CPUPERF signal drives GHI. It
> looks to me like grounding the wire meets one of the pre-requisites,
> but I don't know what level of sleep is required to latch in the
> new state of the signal.
>
> You can experiment with CPUMSR. Make sure your FSB is set to
> the nominal value of 133Mhz, in case you actually manage to
> trigger the x23 multiplier. I think the odds are very low that
> this will work, because I think the tool depends on a mobile
> chipset. The tool also requires an ACPI install of the OS, which
> you likely have.
>
>
http://www.cpuheat.wz.cz/html/AXP_multiplier_FAQ/AXP_Multiplier_FAQ.htm
>
http://www.cpuheat.wz.cz/html/main.htm (CPIMSR download)
>
> In any case, it is known you'll have fun with a P4C800-E with that
> chip, without needing to do anything to it. If your motherboard
> could handle a higher FSB, it would be fun on there too.
>
> Wish I had your 3.06 P4-M
> Paul