1810t TME MOD

The M11x and Asus UL30VT/80VT implements the FSB mod to overclock their SU4100/SU7300 CPU. Earlier on the M11x where only getting 1.6Ghz, but after some bios fixes they got their 1.73Ghz. So it would likely be something the bios is setting up.

Could be an idea to get a MSR dump from a M11x owner to compare CPU states. You could confirm if Linux's reported 1.73Ghz is true by comparing Bogomips info if you can force 1.6Ghz by locking EIST.

Unfortunately the SU4100/SU7300 doesn't have IDA so can't do dual-IDA overclocking.

 

Interesting to see that the M11x had the same issues. So apparently it's a BIOS issue. Do you know if this nando fellow might be inclined to try a BIOS mod?

I'm pretty sure the frequency is 1.73 in Linux because you can see it switching between 1.6 and 1.73 in PowerTOP, depending on the load. In any case, 1.6 isn't too shabby either. The one thing I'm going to do now is to try and wire the FSB pin to the Bluetooth switch.

 

Alright, I've now connected pin 64 of the PLL to the pin of the Bluetooth switch that switches to ground. Since the PLL only samples the FS pins during powerup, I can enter overclocked mode by holding the Bluetooth switch for 1 sec while powering on the notebook. Very elegant solution, if I say so myself. The TME mod is still there, in case something comes up in the future. All that's needed now is a BIOS fix to get full speed in Windows.

 

Is there any dos tool that will write to the MSR?

Maybe if you try changing it before you get into windows?

 

Here's code that checked for the locked EIST bit and tried to unlock it that could customize. Use DJGPGG to compile. I had the binary for this so can't help with the compilation. Could then use the DIY ViDock Setup 1.x to run the DOS binary and then chainload your OS.

Code:

#include <stdio.h>
typedef unsigned long long uint64_t;
typedef unsigned uint32_t;
typedef unsigned uint;


uint64_t ReadMSR(uint reg_index)
{
  uint32_t lo, hi;

  asm volatile("rdmsr" 
    : "=a"(lo), "=d"(hi)
    : "c"(reg_index));
  return (uint64_t)lo | ((uint64_t)hi << 32);
}


void WriteMSR(uint reg_index, uint64_t v)
{
  uint32_t lo = v & 0xffffffff,
           hi = (v >> 32) & 0xffffffff;

  asm volatile("wrmsr" 
   :
   : "c"(reg_index), "a"(lo), "d"(hi));
}


int main()
{
  printf("reading MSR 0x1a0\n");
  uint64_t v = ReadMSR(0x1a0);
  printf("value = %llx,  bit16=%lld bit20=%lld", v, (v >> 16) & 0x1, (v >> 20) & 0x1);
  v &= ~(1ull << 20);
  v &= ~(1ull << 16);
  printf("writing MSR\n");
  WriteMSR(0x1a0, v);
 
  v = ReadMSR(0x1a0);
  printf("readback = %llx,  bit16=%lld bit20=%lld", v, (v >> 16) & 0x1, (v >> 20) & 0x1);
 
  return 0;
}

DSDT tables also entries that change CPU MSR bits. Could make the mod there and do a DSDT override OR swap out the existing DSDt module in the bios. PhoenixTool can do that for you.

 

It seems very unlikely Linux is running at 1.733 even though it might be asking for it. Try looking at MSR 198. If it reads ..xx06?? then it's at 1.6GHz, only if it reads ..xx46?? is it at 1.733GHz. Since EIST is disabled and locked there is no way to select a different multi from the BIOS handover. You would need a BIOS change.

 

Hey does anyone know if I can use conductive silver paint fot this mod, rather than soldering? It was suggested by a friend who has access to conductive paint, not micro soldering tools...

 

Hi, when you say "Remove this", you mean just remove the entire resistor and leave it unconnected?

 

Same Question here. But im quiet sure that silver paint works. But whats with a normal pencil like in the Athlon XP Times.

What happens if i remove the resistor and do NOT connect the 2 points? The Laptop still works? If yes, i would remove the resistor and try out pencil and tell you.

 

I have the SLG PLL and u4100 @1.3...I already flash the memory to 800MHz, I just need to do this mod. but, I am not sure how it is still done, if someone can clear things up a little bit, he will be more than welcomed.
Is connecting pin64 to gnd enough to get overclocking?
Do we have to connect it to the bluetooth switch etc.?
And what is the overall performance gain?

 

So has anyone confirmed that this mod works with the SLG PLL? It seems as though it should work if the pins are the same.

And also, has anyone noticed decreased battery life when overclocking if they aren't touching the voltage? I am thinking about doing that FSB mod to auto overclock to 1.6Ghz but am concerned of loss of battery life. I absolutely love the battery life of this laptop. (~10 hours typing notes)

 

I have the wire in place, but I apparently wasn't able to remove the leg of the transister or resister or whatever that is I was trying to remove. Anyone have any advice for making sure the leg of that piece is completely separated from the pad?

 

Hey,

sorry if I'm asking something thats already been told, but I'm from Germany and I do not know if I have understood everything that was posted here.

I have the 1810TZ with SU4100. My question: does the tme mod from the post #1 work here? Or do I have to do the mod like Gaylord Fister in post #53? In this case, I can't see which PIN is the correct one on the BT module.

Some help/detailed information would be appreciated! Thanks.

 

I just did this last night on my 1810tz with a SU4100. If you're decent at soldering, and have a soldering iron, it's not the hardest thing to do. The resistor is about as big as a grain of rice, and undsoldering was a little tricky. I just used some wire wrap (30 AWG wire) and soldered it to the left pad of "PAD 2". You can ground it anywhere, that was the easiest place for me. We'll see how the benchmarks and stress testing go!

 

Can everybody explain a way to increase the FSB in Linux for an modded 1810tz, please!?
I have modded my 18010tz and with SetFSB in Windows it's no problem to set a new frequency. Then I have tried lfsb to set the frequency in Linux, but the ICS9LPRS365BGLF is not supportet
But here I see that someone have done it, e.g. the post from Gaylord Fister with the /proc/cpuinfo.