M11x clock generator ICS9LPRS387BKLF

@Over - here is the shot with cputruespeed. It shows 1733. If you feel that your R1 is not achieving what alienware advertised as the actual overclock speed I would reach out to Brian AKA ERAWNEILA who is on the alienware M11x quality team. Again I am just confirming what alienware advertised. Best wishes, StevenX



Uploaded with ImageShack.us

 

CPUTrueSpeed just takes the bus frequency and multiplies it by the max multiplier, I tested this out on several computers. It doesn't read the current multiplier.
Since you get stuck on the minimum multiplier with a BSEL mod, it's not the real speed at all.
CpuZ reads the multiplier and clock speed correctly when not overclocked, I trust it's reading thing correctly when overclocked.

 

@over - Sorry I may not be understanding you correctly. I am OC in bios, it's either enabled or disabled for R1. So therefore are you saying you believe the reading since I am OC and you trust the reading when it is OC?

Either way, it is running at 1733. Like I said, feel free to reach out to Brian if you feel your R1 is not running at advertised speeds. Best wishes, StevenX

 

i saw in news that there are new version of setfsb, can anyone explain me in simple words why can't we over clock with setfsb?

 

@vorob - checked website and it doesn't list the pll as one of the supported models. It might be worth reaching out to Nando to see if this release changes anything for us. Good Catch, StevenX

 

The PLL(clock generator) in the m11x is set to not receive programming via software.
You need to physically modify the board in order to change this clock (overclock) using any software.
The overclock done through the bios is a pin-mod on the cpu/northbridge, so it bypasses this restriction - although it limits you to the lowest multiplier on the CPU.

 

so its just impossible to use setfsb? any version? and there is nothing we can do except this hardware?

 

@vorb - this seems to be the case, re-solder your PLL and you can use SetFSB.
FSB doesn't look good. Other option is multiplier. I suggested disabling speedstep and trying CrystalCPUid to change multiplier from 6 to 6.5. Looks like no one has done it yet and since this is my only computer at the moment i won't chance it.
cpuz, crystalcpuid and hwinfo32 seems to be consistent if it comes to reading fsb, multiplier and voltage inboth oc and stock mode.

Most users are of the opinion that we are getting 1.7GHz. You choose....

 

Ok folks.. the pin 14 pll mod has already been discussed in this thread so no reason to go over it again. (Anotherwords you have to short 14 to be able to softclock the cpu).

The P-states. The su4100 and su7300 have two P-states. 6 and 6.5 These are chosen based upon demand, voltage, power-plan within windows 7 and if you have intel speedstep enabled. If you dont want your M11x to switch to the lower P-state, just disable Intel Speedstep and Make sure you are on max performance for power state. You can even go under advanced settings to make sure the cpu operates at 100 percent whether on battery or AC.

I would really like Brian (aka erawneila) to chime in on the p-state speed and trigger so we can put this to bed.

Best wishes, StevenX

 

The P-states. The su4100 and su7300 have two P-states. 6 and 6.5 These are chosen based upon demand, voltage, power-plan within windows 7 and if you have intel speedstep enabled. If you dont want your M11x to switch to the lower P-state, just disable Intel Speedstep and Make sure you are on max performance for power state. You can even go under advanced settings to make sure the cpu operates at 100 percent whether on battery or AC.

Anyway, speedstep is not working on OCed cpu...

 

If your BIOS is doing the BSEL mod to overclock, EIST will be disabled and so you will be stuck at the lowest multi, at least that's what happened with my BSEL mod.

 

Last 2 posts are exactly what I'm talking about.
We can't physically measure the clock speed inside a CPU, best we can do is use what it's reporting, or just use benchmarks to Guesstimate.
For example, here are my results from IntelBurnTest and calculations:
@0.8Ghz
5.104 GFlops
=6.38 GFlops / Ghz
@1.2Ghz
7.72 Gflops
=6.48 Gflops / Ghz
@1.3Ghz
8.636 GFlops
=6.64 GFlops / Ghz
With OC in bios
10.62 GFlops
if we presume 1.6ghz, 6.64 Gflops/Ghz
if we presume 1.73ghz, 6.13 Gflops/Ghz (lower perf/clock than any other setting)

It is very unlikely to be running at 1.73.

 

If you run setfsb it puts us at 1.73. By all means.. keep up with the testing. It doesnt change any of my benchmarks at all.

Powered by R1.

StevenX

 

Even if SetFSB could be used after a pinmod, you have no way to change the VCore. Might not even reach 1.73 GHz this way. Some Asus laptops with the SU7300 CPU have reached 1.9 GHz overclocked, though. But Asus provides more voltage to those CPUs.

While the "science" of changing the FSB and clock speed of a CPU seem to be well known, I have never seen anything mentioned about how to increase the VCore for a CPU.

 

You know there is another SU cpu that may fit ...

SU9600

 

If you can find the schematics for your notebook, you will see the chip which controls the voltage for your CPU. Then a pencil mod would help you increase the CPU voltage.

 

Yep. I think we've got a little wiggle room on voltage with these chips, they run pretty darned cool.
I think a pencil mod would be much easier than desoldering/resoldering the cpu

 

You can borrow my easy bake oven.

 

I didn't think anyone else still ran a Prescott PC

 

I have an afternoon to myself. All revved up and ready to mod.
Then I run into a snag... I have a different PLL
SLG8SP556V

 

Yikes! I have no idea on that one. Wonder if Nand0 can chime in?

Congrats on the courage to do that. Tough to be the first one over the bridge.

Best Wishes, StevenX

 

@OverZealot - did you ever get the PLL, TME clock disabled? Best Wishes, StevenX

 

Bump to keep this alive. I installed the newest version of setFSB and Abo(from setFSB) asked that I run a diagnosis and some info to him. I sent all the stuff and he had me run some settings back and forth. As stated before the TME is locked and we can't update... I asked if there were any other methods (out of the box thinking) type of stuff and he said they can't support it. So... If we could get the R1 up to 2.0 ghz I think we would be in a good place.

Anyone have any other updates? Best Wishes, StevenX

 

Been busy studying and playing games that don't need the extra cpu speed
Attempts to change clock speeds using similar PLL presets in setFSB fail, so I presume that it's locked.
I found anecdotal evidence both for and against the ability to change frequency via software... mostly against though.
The datasheet doesn't describe a TME pin, nor a method for soft-clocking.
Maybe I'll do an FSLx mod in the future, or try the PCI2 pin. Not today though.

 

Fresh info from SetFSB site:

It can also work in M11x - but you must unlock TME (I'm not 100% sure - you might try without TME mod).

 

And how can we unlock this tme?

 

I'm wonderin too. This might help me in optimizing my R1 XD

 

You would have to do a pin mod to unlock the TME. Then you can OC, but that requires cutting and soldering a resister in.

 

Why it is so hard? Why there are so many laptops which can be overclocked without such pain in аss? Maybe there is some software way?)

 

TME is locked on the M11x PLL. Only way to then get software PLL programming is with a pinmod (see sig). Before doing that, consider Asus UL30VT/80VT owners without such restrictions were able to get their SU4100/SU7300 up to 1.9Ghz. So is the extra 0.17 worth voiding your warranty for? I don't think so.

 

Hey guys, I know it's been a long time but I have a question to ask: Is TME unlocked on this M11x?

EDIT: Looks like it is. I followed Dufus's guide a few pages back and I can't set byte 0E.

 

If it's using an ICS PLL, then the TME_READBACK register 9 bit 6=0 which indicates that it's TME-unlocked.
Can see a TME-locked PLL setfsb diagnosis screenshot here.

Have you tried to move the slider to see if can set a different FSB?

 

I tried moving the slider and it hard locks the laptop.

 

Try manually adjusting the PLL config registers as shown: http://forum.notebookreview.com/ali...generator-ics9lprs387bklf-12.html#post6188065 .

 

I did follow that guide- I get stuck at changing byte 0E. It won't change value.
EDIT: I fail. I was so excited that I forgot to click Update then Apply. Now testing.

 

This is the best I can do ATM:

At least I got SpeedStep and the 6.5 multiplier back .
Problems with overclocking:
1. I think RAM is holding me back- I tried higher but it crashes. I need to edit some SPDs.
2. I don't understand how Register 14 increases clockspeed- I just added 10 to 68 for 78 to get a 30mhz increase. Adding 10 again got me to 260mhz. But putting in 69 or 98 gives me a blackscreen.

 

If you contact abo (creator of setfsb) he can probably fix whatever setfsb is doing wrong that causes the lockup.

Looks like everyone with the m11x should check the TME bit.

 

@DavyGT - congrats on getting setfsb to work on the M11x. The Asus UL80VT folk got their SU7300 systems cranked up to a stable 1.9Ghz from a base of 1.73Ghz (266Mhz FSB BSEL turbo33 O/C). Something to aim for.

 

He added support for the 387 already but it's only in the shareware version, which means I can't automate the overclock until I get it.

Rather, I have to edit RAM SPD. Interestingly, it supports running at 1218mhz, which means I just have to edit those timings into the slowest speed.
More important, I have to work out how does register 14 increase speed- I'm just guessing ATM.

 

Does your m11x have ram settings in the bios?

If not, thaiphoon burner will edit and flash the SPD on most ram.

An old trial version that lets you flash the SPD is attached to this post:
http://forum.notebookreview.com/har...pll-pinmod-overclocking-methods-examples.html

 

The M11x doesn't have RAM settings in BIOS. Wish I could have the 2:3 ratio back . I'll have to use Thaiphoon Burner to get the desired timings. Thanks.

 

Wow, nice work Davy. Did you pinmod your M11x? I don't understand how you unlocked the TME?

 

Thanks, I disabled TME through soldering a resistor to pin 14.

I managed to work out Register 14 partially- it starts at 68 and increases by 2mhz with every letter replacing 8. Eg: 6A will give 268 if 266 is stock. It all works until the letter G is reached. It will freeze at G- which means to keep increasing, the first digit needs to be increased by 1. Adding 1 will add 30mhz. That's all I know from observation, I'll figure out a mathematical formula and reasoning later.

I managed to break 1.9Ghz by negating the RAM limitation through using the provided "Overclock" mode for the 2:3 RAM ratio. At 324Mhz FSB, RAM is running at 974Mhz, well within the previous overclock. Too bad the RAM limitation is looking to stay with me- Thaiphoon Burner doesn't work (Gives an error message "Could not install system drivers" and SPDTool has trouble reading my RAM SPD.) Time for stability testing.

 

I see a pinmod in my future.


*Anxiously awaits stability tests* What method are you using? Prime95? And for how long?

 

@Davy - what is your 3dmark06 score now? Just to give a general "idea" of score. Also, is it stable when overclocking video simultaneously?

A huge congrats! I believe you are the first so you deserve some props for sure.

Good Luck and Best Wishes, StevenX

 

I do a quick test of 5 runs in IntelBurnTest to prove basic stability then I do a overnight run of Prime95 and finished off by 20 IntelBurnTest runs.

EDIT: I haven't done too many benches yet as wprime gives an error and I haven't bothered to do 3dmark so I dont know what to expect. I did do some quick runs of Cinebench r10 and it scores 4295 at 2.13ghz. So far, I managed to breakthrough 2Ghz but cannot pass 2.1ghz stable. This is either due to the RAM running at 1040mhz at 6-6-6-15 timings (too fast) or not enough vcore. Both fixable.
Sorry for the taped reply, mobile browsing is so... Slow and messy.
EDIT EDIT: I haven't tried gaming but my current GPU overclock (575/800) holds fine so far.
BTW, thanks for the rep guys. Be sure to rep nando4 and Dufus as well, as they figured out this mod in the first place.

 

awesome sir!

+1

i am sending back my R1 and waiting for a good deal on a R2 now. but this great news!

 

DAVYGT you the man.

 

can you post a photo or a paint diagram (lol) about how you soldered the resistor to pin 14? whats the value of the resistor you used, 1K, 2K or another?

 

Wow, congrats DavyGT, Nice work man. Son you got some big b@11$. +rep