Yonah has been pinmodded! - Page 4

hmm then we have a problem, is there a way to raise the standart resistance without isolating the pin? that could be it. At what values considers the chipset a pin as high or low resistence?

IS speed steep techology enabled in the bios?

Did you try with it disabled ?

I am not sure what you were using to see the CPU speed in linux but off my sabayon live cd it lists the processors by doing a cat /proc/cpuinfo and it will show the current speed, example:

processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 14
model name : Genuine Intel(R) CPU T2500 @ 2.00GHz
stepping : 8
cpu MHz : 2000.000
cache size : 2048 KB
physical id : 0
siblings : 2
core id : 0
cpu cores : 2
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 10
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc pni monitor vmx est tm2 xtpr
bogomips : 3994.27

processor : 1
vendor_id : GenuineIntel
cpu family : 6
model : 14
model name : Genuine Intel(R) CPU T2500 @ 2.00GHz
stepping : 8
cpu MHz : 1333.000
cache size : 2048 KB
physical id : 0
siblings : 2
core id : 1
cpu cores : 2
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 10
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc pni monitor vmx est tm2 xtpr
bogomips : 3989.42


Notice how the second core is at 1333 MHz....

news?

Eh, if you guys can wait I'm definitely going to be trying Merom 667FSB to 800FSB mods. As of now, I cant help much here as I dont have a Yonah CPU . How did the rest of you guys make out?

On that note Mr. K6..

Yonah doesn't play at my house, but merom (core 2 duo) does & I think that it should have good O/C'ing potential. Merom is higher binned than the desktop conroe chips so there should be lots of room to overclock the front side bus despite the multiplier lock. Likely the big impediment will be adequate cooling. For anyone who is thinking of trying this - the first thing you should do is ensure that you have decent thermal grease under the heatsink. It would also bear investigating to see if there are any ICT or manufacturing jumpers that might reveal some overclocking friendly options in the i945 based Dell motherboards.

The first thing I am going to try when I can find the time is to see if my merom is stable with a 800Mhz FSB. The Merom datasheet only show Bclks of 133Mhz (533 FSB) & 166Mhz (667 FSB). But there are some unexplored reserved settings.. Hmm, I wonder what those are? When we look at the Pentium D datasheet, it looks like the (Bsel) LHL probably selects a 200 Mhz clock.

Merom
Bsel2 Bsel1 Bsel0 Bclk
L L L reserved
L L H 133
L H L reserved
L H H 166

Pentium D
Bsel2 Bsel1 Bsel0 Bclk
L L L 266
L L H reserved
L H L 200
L H H reserved

So if we take a T7200 (Merom: 2.00Ghz Core 2 duo) for example, we can probably bump the FSB from 667 to 800 by pulling Bsel0 low. That would clock the chip at 2.4Ghz (200Mhz x 12). One way to do that would be by shorting Bsel0 to the adjacent Vss pin (short pins B22 & C22). If necessary, two Vid pins could also be shorted to bump the processor voltage up to 1.400v (Vid pins 6:0 0001000).

Caveats
Keep in mind that this is theoretical. I don't know anyone who has actually tried this yet. While I think that it will work, be warned that it could also DAMAGE YOUR PROCESSOR AND/OR MOTHERBOARD. Probably not, but I won't be buying you a new notebook if you short the wrong pins or if something goes wrong. So proceed with caution.

Second, if you find it necessary to bump the processor voltage up for stable operation at a 800 FSB, it might interfere (override) the power saving, low-voltage EIST modes. I doubt that this is a problem, but I don't know enough about EIST (sidestep) logic to be certain. In any event, the T7200 at least should be within fairly easy reach of 20% overclocks. Won't that be nice?

E.g.
T7200 FSB 667 -> 800
Core 2.00 -> 2.4 should be easy

T7400 FSB 667 -> 800
Core 2.16 -> 2.6 probably

T7600 FSB 667 -> 800
Core 2.33 -> 2.8 ?

I wonder if you could do that to a Core Duo 2.0Ghz Yonah?

Floating pins, that is unconnected pins, are not considered reliable. For this reason, they are sometimes tied high or low internally which will default their state unless you force the logic state with an external voltage. You should check the yonah datasheet to see if that is the case here. Logic high & low varies for various processor pins. I glanced at the yonah datasheet and the open drain specs are; 1 (high) = 1.00 to 1.10v 0 (low) = 0.00 to 0.20v However, you should still check the datasheet. You didn't specify what pins you were talking about & your pin(s) may be open drain or may fall under a different set of logic state voltages. HTH

There are no notebook motherboards that support an 800Mhz FSB though. This might be possible in a desktop, but not with any current notebooks

No official or explicit support. But if the electrical design has sufficient headroom, you should still be able to run an 800 FSB. The only way to know for sure, is to try it. Pin modding e.g. forcing control signals at the processor is one possible way of doing an end run around what the M/B wants to allow. Merom is almost certaintly capable. Whether the M/B can cope, thermal issues & unexpected twists yet to be determined.

Have a look at this screenshot to see what kind of O/C potential yonah has. On a desktop board at least it appears that the yonah T2600 core can get into the 3Ghz range (231 x 13). That's an amazing number for an air-cooled processor - a 40% O/C. Considering thermal & memory issues, a more reasonable target for a notebook would probably be 2.6Ghz at no more than 1.4v (T2600: 200 x 13), with PC5400 ram. Preferably at stock voltage, if that is stable. The merom/core 2 parts are higher binned than yonah & probably will make more reliable overclockers. But the yonah results are nothing to sneeze at. I don't have any yonah parts (just merom), the proof of concept & screenshot come from the good folks at xtreview.

So the core chips appear willing & the thermal load is probably managable with a 200Mhz clock. The two possible obstacles that I can envision are Calistoga locking up/locking out the 200Mhz clock (won't be an issue for Crestline/965 - which will explicitly support a 800 FSB). And /or any booby traps that Dell might have laid in the BIOS.

How are you going to get the FSB to 800mhz when the i945 chipset only operates at 400/533/667mhz??? If you can figure a way to get the FSB to budge higher than 667mhz I'm all ears and I will be a test dummy.

I'm not Mr. K6 or the OP. But why assume that the chipset is incapable of a 200Mhz clock (800 FSB)? That may or may not be true. Just because it is undocumented or reserved doesn't mean much. It wouldn't be the first time that a "future" feature was hidden & undocumented in an intel part. Intel originally intimated that the "refreshes" i.e. the forthcoming 800 FSB santa rosa merom would be backward compatible with socket M. They are now telling us otherwise - socket P to go with the 965PM crestline chipset. But chip design has lead time & if the original intent was to support it.. who knows what support is really in the 945PM chipset? Even if there is no strap recognition for 200Mhz - it might still be possible to force by resetting the PLL in software or by a bios runaround. We really don't know anything until we dig around & see what's really there. BTW, does anyone happen to know what PLL clock gen chip Dell is using?

me too.unless theres some bios update that allows hidden boosts.

man i want this to work tho. CSS sucks.

Well considering the i945 chipset was being developed long before the roadmaps for Merom, I don't see how they would reserve 2 generations in a chipset that was designed for 100/133 FSB. The fact that they reserved 166 was futureproofing enough for a single chipset family.

There will be i965m chipsets coming with true 667/800 support, and maybe they'll have 1066 reserved for futureproofing.

But if you can show me a chipset in Intel's history that had 2 forward reserved FSB speeds and not 1 forward and 1 backward, then I would give this idea some hope.

Remember... these are laptops, there is only so much we can do within reason to them before we fry them.

When we got the i915 chipsets we were allowed 100/133 operation, then Intel made the decision to recycle low quality samples of 133 cores as lower 100fsb products. It wasn't hard to reverse a downgrade. And eventually we may figure out a safe way to do the same with the T2x50 chips which are downgrades of low quality T2500/T2600 chips.

We'd have better luck trying to replace parts on the motherboard to unlock the multiplier or CPU frequency before we would ever be able to push an i945 chipset to 200.

Keep in mind that Intel originally intimated that the "napa refresh" i.e. the forthcoming 800 FSB santa rosa merom would be backward pin compatible with the 1st merom iteration. Now they are singing a different tune - a new, incompatible socket P & crestline 965 chipset. But if there is any truth to the original position & notwithstanding the design lead times for hardware design.. there could well be a 200Mhz strap support buried in the 945PM chipset. Intel is famous for hiding features that they don't want us to have now, but they might want available in the future.

Well, there are several things that we can try. They may or may not work, but they are plausible & worth trying. Of course none of these steps would be necessary if Dell simply provided a few basic O/C BIOS features. Anyway -

1. Download a windows based PLL reset program like Clock Gen. And see if the FSB can be bumped to 800 temporarily. Best attempted with PC5400 (667Mhz) ram. And more likely to work if you know what PLL chip Dell uses on their motherboards. You might find this somewhere on the net, but it's generally necessary to look at the motherboard and see what is stamped on the chip. It will probably be one of these brands: Cypress, ICS, IDT, Pericom, RealTek or WinBond.

2. Open up your case and short the clock select pin that selects for a 200Mhz clock (800 FSB). For merom, this would be shorting (placing a thin wire jumper) between pins B22 (Bsel0) & C22 (Vss). You can find a map of which pins are which here (merom). Or the corresponding datasheet for yonah here. Keep in mind that whether this works or not, you will probably have to do a CMOS/BIOS reset afterwards.

3. If #2 is not successful, the processor may need a bit more voltage to achieve a 200Mhz clock. So leave the B22-C22 jumper in place & add some voltage jumpers to force the higher voltage. Off the top of my head, the merom nominal voltage with sidestep disabled is 1.3v. So to bump this to say 1.4v, we need to change logic 0010000 to 0001000 (Vid pins 6:0). So short pins AE3 (Vid4) to AF3 (Vss) and AF4 (Vid3) to say AF9 (Vcc).

4. Let us know what did or didn't work for you & we'll go from there.

They always showed on the roadmap that Yonah would be a small step to Merom, then Merom2 would introduce the new chipset and socket type. That roadmap never changed. From they beginning they had an entirely new platform in the works. They just gave us Merom1 on the existing architecture to give us a half-assed taste of things to come.


Intel is known for it, but like I said... they already expanded i945 to 667, that was the "future-proofing" of the system. i915 chipset expanded the i8x5 chipsets to 533. They were the same chipset but rebranded. And since laptop motherboards do not have the capacity for modification, most brands are direct ports of reference designs ratified by Intel.


This is all fine and well, but no one has ever gotten any of the Dell Inspiron motherboards to change the FSB thru a PLL modification. We can flip the CPU to another FSB, but the motherboard is a locked component to us. Always has, and likely always will unless you want to find some i965 chips and try soldering them to your i945 board. Then there is the total lack of printed circuits to handle the additional speed and bandwidth.

You really think the board is over-built? Yea... and I have a bridge I'd like to sell you.

This isn't a Gigabyte or ASUS motherboard in our Dell's.