How to Unlock the Core 2 Extreme Multiplier in Windows

The 3dmark06 for the Q9200 should be at least the same as for the P8800.
When you view the result online(I assume its the 3dmark06 test version) what is the CPU score?

 

CPU Score
3624

EDIT: I've noticed the guy in this test probably OC'd the cpu?

http://forum.notebookreview.com/sager-clevo/364578-m860tu-qx9200-es-test-coming-soon.html

My core clock (cpuz) is at 1596 Mhz
multiplies x6

The other guys setup:
core clock: 2394
9 multiplier

 

I used throttlestop to unlock the multplier but the results aren't great:

My specs:

3DMark Score
6417 3DMarks
SM 2.0 Score
2581
SM 3.0 Score
2313
CPU Score
3549

GPU: nvidia GTX 260M
CPU: Q9200
Multiplier x9
Core speed 2393

http://forum.notebookreview.com/sager-clevo/364578-m860tu-qx9200-es-test-coming-soon.html

GPU: nvidia 9800M GTX
CPU: Q9200
Multiplier x9
Core speed 2393

3DMark Score
10676 3DMarks
SM 2.0 Score
4600
SM 3.0 Score
4406

Same specs but he's got 4000+ points more

 

The other guy had a resolution of 1280x1024, wasn't yours 1680x1050?
That makes a difference.

 

Not this time, my latest 3dmark06 score is using everything on default.

 

As far as I can tell, the 9 multiplier is the default maximum multiplier for a QX9200. You haven't unlocked anything.

I just got a PM message from a happy QX9200 owner.

Now that's taking advantage of an unlocked multiplier.

Try running a GPU-Z log file while running 3DMark to make sure your GPU is running at the correct speed. It might be running in 2D mode and killing your scores. The QX9200 has a lot of potential once you crank up the voltage and multiplier.

 

Hi! thanks for helping me out!

I increased the multiplier to 12.5 or so and then voila BSOD.

Rebooted back to windows 7 then I got a massive bsod once more, tried my archlinux partition then it said something about cpu cache not being written and a complete froze.

5 minute later linux and windows 7 works again.

I will try the GPU-z log suggestion, thanks

BTW all of my cores according to CPUID hw monitor is running at 52C idle if I run the multiplier over 10 and vid 1.275v

I'll report back as soon as I find out more

Is it "safe" to try to find out the limit of my cpu? what exactly happens when 2 Operating systems fails to boot due to a CPU oc? Everytime I try to go past 13 my system bsods and linux/windows refuses to boot/hardware freeze.

Overheating?



Trying:

12 Multiplier
1.3 VID

Prime95 bench

If the temp peaks at 80 or max 90c is that okay?

What's a safe temp to stay at during full load?

I've started reading this excellent thread from page 1, trying to find if most of my questions have already been answered.

TZ0 in CPUID HW monitor goes to 90-100c everytime I run prime95, hmm

Using stock CPU settings:

CPU CORE: 2.6Ghz
TZ0 temp in CPUID: 93C and rising

isn't that insane?

 

There's no Intel warranty on a QX9200 ES CPU so you'll have to use your own judgement on how far you want to try and push it. My #1 suggestion is to do some thorough stability testing with any multiplier before trying to step up to a higher multiplier. If you randomly increase the multiplier and you don't have the voltage set high enough to support that multiplier then you can end up with a BSOD and that can nuke a Windows install if it is writing something to the hard drive when it crashes.

The guy that just wrote to me needed a VID setting of 1.40 to achieve enough stability to get through a Vantage run at 3.48 GHz. Your CPU will probably need similar. Ultimately you will be limited by heat. That might be overkill for your heatsink and fan to keep up with. I like starting with too much voltage and then I decrease it after stability is proven. That has always been safer for me when exploring the outer limits of what a CPU can do. Run some stability tests like wPrime on all 4 cores when testing to give you an idea if your CPU is stable or not.

Once you come up with a safe maximum overclock then you can decide if you want to come up with a not so aggressive profile that you can switch to for daily use and keep the high OC profile for gaming.

 

A big thing with 3dmark06 scores is the operating system you're running.
I know for a fact XP vs Vista can be almost a 400 point difference.

When you check kaltmonds OS, you'll see he is running server 2008 the best OS made by windows. It is vista without the huge performance drop.
I think w7 has a similar if not bigger performance decrease.

90c is unsafe. You want to baby that chip!
Its not overheating, as unclwebb said, its instability. But yes, a 1.275 with a multi higher than 11 you will certainly overheat under stress of a few minutes.

Unfortunately the BSOD issue has to do with throttlestop but as unclewebb and I discussed, the program is not exactly responsible for this. When your computer crashes so quickly, it doesn't always have time to reset the multi of your CPU, so it restarts at the same unstable overclock. Often unplugging all power sources and holding the power button for a while can fix problems like this.

There are some great ways to avoid getting your multi above 11. I don't think I will ever need to do that, because I finally discovered how to stabilize FSB overclocking with the free version of setFSB.

If you click the link in my signature where my q9200 is at 3.65ghz, you'll see its only at a 11.5 multiplier. This is because I have my FSB overclocked from 266mhz effective 1066mhz bus speed all the way to 333mhz effective 1333, so its 11.5x333.

This doesn't mean that you dont have to increase the voltage for stability, on the contrary many more things are unstable such as RAM which must be downclocked to compensate, but its my opinion that FSB overclocking is generally more efficient and will let you run at lower voltages with less heat. Its just quite a hassle to get it stable in the m860tu especially at 1333.

Fortunately for you there is a much more stable frequency that can get you some pretty high clocks with a lower multi and possibly less voltage. Its 1149mhz. which overclocks a stock q9200 handsomely at 2.58ghz. There is an easier and very similar 1120mhz but if you can get another 30mhz, then why not?

Heres how its done, I made a little guide
lol I guess I don't want to be hypocritical with the spoiler tag thing

Spoiler

I'm not sure how many people with an m860tu know about this... but I don't really feel the need to post a thread on it, its from an older thread.

 

Thanks for the cpu crash explanation. Yeah of course every tampering on my own hardware is my own responsebility. I was just curious if the cpu could break everytime we peak a number or BSOD.

That's insane, i'm not sure if he's on stock cooling or better then my own. But i'm getting over 84C just by going over 2.5Ghz but it seems to be stable and now it went down to 82C. Still 2.5 to 3.6Ghz is one gigantic leap

Thanks for the application

And thanks for all your help!

 

Oh I didn't know that, thanks for letting me know sir!

Yeah like I was saying before everyting we do to our own hardware = our own responsebility.

Whoa thanks for the guide!!! I just followed your guide and there was a smell increase when my FSB increased. But how did you reach a stable 1333Mhz, did you also undervolt anything, flash your ram?

Maybe I have bad luck or something. But it seems like every setting I try, the temperature goes insane just after a minute or so using the torture test on prime.

Right now i'm trying

Multiplier: 11
VID: 1.2500

And after 30 seconds i'm already hitting 80C
now i'm over 90C.

Maybe I should try increasing more to my fsb? Do you know how to at least juice out 300 more on my bus speed? Right now I have 287 after following your guide.

Now the temperature have gone down 2C, seems to be stable around 88C using multiplier 11, VID 1.2500, wrong now it's back to 90C.

But none of my cores went past 60C? It seems like TZ0 is always peaking at 90C no matter what.

Do you know of any cooling mods I can look into it that doesn't require rambo soldering skills?

I'm gonna try a multiplier of 11 and 1.2VID

Thanks again

Multiplier: 11
1.2875 90c+
1.2750 90c+
1.2625 90c+
1.2500 90c+
1.2375 90c+
1.2250 fatal error in prime95
1.2125 fatal error in prime95
1.2000 BSOD

*sigh*

I guess there's no point in increasing the multiplier and vid since if it overheats at multiplier 11 and any vid under 1.3?

Increading voltage = more heat right?

Again none of my cores goes past 80C, but TZ0 goes to 9-92C

Prime95 torture test after 5 minutes
Core Speed 2585.5Mhz
Multiplier 9
Bus speed 287
Rated FSB 1149
89C

Really? That's it? My TZ0 stays at rougly 5-60C idle, all cores around 50C, are those okay numbers? This can't be right, stock is like 2.4Ghz right, a small oc (setfsb) to 2.5Ghz and it's almost 89C?

Can it be my cooling, I mean too much/little artic, still wouldn't it overheat really fast just by stock idling if that was really the case?

I'm the only one who can't seem to OC this CPU using only multiplier without overheating?

 

Hmm. No I don't think you are the only one. the m860tu is a very small notebook for this kind of stuff.

I do think there is something wrong though. Is your fan spinning up all the way? Maybe you are stuck in silent mode.
Are 2 cores running a lot hotter than the others?
Mine idles below 50c always. I would maybe recommend a different cooling paste, you do mean arctic silver 5 right? I have mx-3.
Did you check the box in throttlestop options that says core 2 mobile cpu? What are you using to monitor temps?

Occasionally you will find chips that just run hotter than most others. They all do have heat and stability differences.

I'm really not sure if the 1333mhz is cooler or more stable. I will test it in a minute. Yes I did have to flash my RAM to lower timings, as well as downclock my graphics card, I believe that was one thing that fixed it, but i'm really not sure what did it.
I am using the 6th position on the slider, I'm not sure if before I was just using a position that was less stable and changing more of my clocks like SATA and stuff.

cooling mods dont generally require any soldering, the ramsinks i suggested you stick on just have thermal transfer tape...

 

running orthos right now with 9.5x333mhz=3.15ghz 1.265v

I only stress 2 cores

both of them hit 80c after 2 minutes 45 seconds. And I'm in silent mode, my fan is locked at an extra low speed.

FSB overclocking does seem to be efficient, as normally to run at 3ghz I have to do 1.275v. Though I only got through 2 tests. I dont do as much stress testing as I should.
How long does it take you to get a fatal error at the lower voltages?

 

When it goes past 80C the fan really starts to kick in, so i'm sure the fan is doing it's job, although it's one poor fan and shroud if you ask me.

All 4 cores are showing approx the same temperature, i'm using CPUID hardware monitor and realtemp to monitor the temp. Throttlestop says "intel core 2 extreme"

Yeah I know about the ramsinks but I doubt they will decrease the temp from 90C down to 7-80c

I just did a test again after re-applying more paste on the CPU/GPU and cleaing out leftovers, at stock 2.4Ghz running prime95 torture test after 10 minutes the temp is now 80C (TZ0) all 4 cores around 70C

Are you telling me your fan is spinning even slower and it's only maintaining a steady 80C? Wow if that's the case then i'm impressed especially if you're past 3Ghz.

The errors at lower voltages takes rougly less then a second to show in prime95 and after a few minutes prime95 crashes "fatal error".

I'm using artic cooling 5, not the MX3, i've checked several tests and at max the only difference is roughly 1C compared to artic cooling 5, and in some tests AC5 prevailed over the MX3 by 0.5C.

I'd be happy if I could just reach 3Ghz at a stable rate but now it seems pretty far away especially if prime95 at stock settings (2.4Ghz) yields 80C stressed

Thanks again for helping me out bro!

EDIT: I have no idea what just happened but after pulling every component out from my laptop and cleaning my CMOS, this is my new 3dmark score:

3dmark06 score 10945
SM2.0 Score 4582
HDR/SM3.0 Score 4539

 

just received a qx9200 from laptopmonkey. here are the results

 

i have actually hit up to 3.55 stable. i believe at 3.55 and 1.4V i received a 13,850 CPU score

 

Good work greenfish, whatever you did.

Now if you can just find a way to keep your CPU cool, maybe you can catch up to what wB` is doing with his QX9200. Amazing bang for the buck.

laptopmonkey still has lots of QX9200 CPUs available at a great price.

INTEL CORE QUAD Q9200 2.40GHZ- QAVR - better than Q9100 - eBay (item 140476668630 end time Dec-10-10 06:38:01 PST)

 

i would purchase some mx-4 off of ebay. i have had some great luck with it. also make sure that there is contact between the cpu and the cpu heatsink....

 

Hi! Thank you

Right now i'm holding steady at 3Ghz using a 11.0 multiplier and 1.3250 VID temps never go above 70-72C if i'm using ORTHOS (2 cores) prime95 is a whole different matter, it takes less than 30 seconds for me to reach 90c regardless of what VID/multiplier i'm using.

Even at stock 2.4Ghz I find this to be one fast CPU, of course I notice a difference running it at 3Ghz, especially when i'm loading data in games.

Thanks for the benchmark, very nice results

Have you also tried prime95? I can get 3Ghz from this baby if I run just ORTHOS (2 cores) my temp stays steadily around 70-72C, but 10 seconds on a prime95 torture test voila 80C in less than seconds..20 seconds more 90C.

Regarding the thermal paste, artic cooling 5 is good enough for me. But I'm gonna buy a couple of batches of MX-4 to use in the future, thanks for the suggestions

I've also installed heatsinks on my CUP heatsink like niffcreature recomended, cheers.

 

right now im at 3507mhz my max temps in CPUID Hardware Monitor have been
78C-Core 0
75C Core 1
85C Core 2
84C Core 3

the temps in Throttlestop seem a tad bit higher but the computer keeps going full speed!

 

What is your laptop? and graphics card(s)?
It would be more interesting if you posted more information.

Also, it looks like your heatsink might not be quite level. A 10c difference in the 2 dual cores is pretty big!
If it is level, then you have 1 dualcore that runs hotter than the other. Perhaps it would be a good idea to give the hotter side of the CPU better contact in an attempt to level things out at around 80c.

 

niffcreature,

i have an m17x with dual 4870s (non overclocked)

i have honestly taken this thing apart 4 or 5 times and re-applied mx-4 compound to try to even things out but the temps have literally stayed the same.

i also have a an area-51 alx with a 980x hex core and have the same issue where the cores temperatures fluctuate quite drastically from one another.

 

The sensors that Intel uses are not 100% accurate temperature monitoring devices. It is also very likely that Intel deliberately sets TJMax differently by 5C or 10C between the two different dual cores within a Core 2 Quad. This is a an easy way to prevent all 4 cores reaching the thermal throttling point at the exact same time which would be more noticeable to a user. It's a lot easier to just throttle one or two cores to control heat. This is usually enough to keep the other two cores from reaching the thermal throttling point.

Your temperature spread is perfectly normal in a Core 2 Quad and it usually has nothing to do with how you applied your paste or how square your heatsink is. These sensors were only designed by Intel to control thermal throttling and thermal shutdown which is why Intel does not recommend you use these sensors for accurate temperature reporting. They have never fully documented their calibration procedure so accurate temperatures from these sensors are impossible.

Can you check the ThrottleStop.ini configuration file? Is it using TJMax=100 or 105? For a QX9200 it should be using 100. If it is not then you can edit that line. I wouldn't be surprised if it is using the wrong TJMax for some of these ES processors.

A 10C spread or more is also very typical of the Intel 32nm 6 core CPUs. Perfectly calibrated sensors would cost Intel too much money so they don't bother. Sometimes you get lucky and get a CPU with well balanced sensors but most of the time you don't.

 

Ah, so that explains the wacky temperature differentials from all four cores. Thanks for that knowledge! I have a Q9200, it's an ES I know, but those sensors are either really wacky, or one core runs about 10-12C cooler than the others at load. Probably the former.

 

I got 6000 in score using 3dmarkadvantage default test. I noticed your gpu score is doubled my own.

My fps was like 15 and yours is over 30 in the tests.

What gpu are you running?

 

tenchar

 

oh my

BTW thanks for helping me out sir
ram heatsinks installed

 

im using 4870s in crossfireX on my m17x r1
ya its set at 100C
i have been running prime all day stable. at 3.55 it wont run prime stable for more than a couple of hours before having an error but i have played games flawlessly for longer than that. i like keep it around 3.1-3.2 all the time

 

I have 8-9C difference between my hottest and coolest cores. So I think around 10C difference is probably normal. Ran some tests already which suggest I can do 3.5ghz stable with 1.225V and 70C load temps however my power supply is lacking so 3.4ghz is probably as high as I can go with my QX until I upgrade said power supply..

A great program Unclewebb I look forward to pushing this chip to the limit soon.

 

does anyone notice a high pitched whine coming from the qx9200 cpu? its hard to hear it unless its very quiet or if you put your ear up to the bottom of the laptop....

 

The whine won’t be coming from the CPU, more likely some other components (capacitors maybe, could be chokes, not sure). But not the CPU.

 

My CPU whines all the time, It's too hot in here, you are working me too hard, there is never enough break time and the list goes on..............

 

3Ghz, gaming black ops for 2 hours 93C, funny enough stalk clear sky using dx10 and with ultra high res textures (cs complete) i barely go over 80c and stalker is known for it's poor coding (mostly cpu).

Would buying one of those ugly zalman 2000 laptop cooler work for me?

 

A laptop cooler primarilly works for idle cooling when the fans are not churning away. They will help a little under load just making sure there is fresh air under the laptop for the fans to pull...........

 

Enabling the C States option in ThrottleStop seems to have fixed the whining noise that wB` was hearing. I've heard that this helped another user or two a long time ago. Not sure why but it's worth a try if you have a similar noise issue.

 

anyone have a recommendation on a good starting point with a Q9100??
stock temps seem to be 60*ish
"cpu 2" alwase seems 4-5* cooler than the rest

 

Temperature sensors on the 45nm Core 2 CPUs are not 100% accurate. Reported differences of 5C or 10C from core to core is not unusual at all. That doesn't necessarily represent an actual difference in temperature so the reported difference is usually meaningless. Intel only designed these sensors for thermal throttling and thermal shutdown control so they are not very accurate for anything else.

The Q9100 is not an Extreme CPU as far as I know so you won't be able to use ThrottleStop to unlock the multiplier. Post a CPU-Z screen shot. If it doesn't say Extreme in the Specifications box then your CPU is locked.

 

Its more like a specification line, the box with the picture only says extreme with qx9300.

q9100 has to be an ES cpu for unlocked multipliers.

 

Sorry about the lack of info, I had been up for a loooong time....

the q9100 I have is an ES, I can use TS to unlock the multiplier/voltage and cpuz recognizes it as an extreme series cpu

I will have a pic of cpuz and TS up in an hour or so.

 

Unclewebb, I cannot get throttlestop to reset FID/VID on exit.

Its kind of worrying to think you are running your cpu at stock frequencies and then notice its at 3ghz

 

I just got a x9100 QS from a friend and but I can't seem to get any new menus from right click, no option to unlock the multipliers? Also the CPU-Z didn't show the Core2extreme(black/white)logo, instead it's a core2duo logo? The revision bar is also grayed out, not sure if it's E0 or C0. Windows and hwmonitor detect it as a x9100 no problem as you can see in the screenshot.

Besides these confusion, the chip runs very well so far and I am able to undervolt it to 1.1125v from the stock 1.2625v, maxed at 60C after the undervolt.

 

Just checking..
make sure ur right-clicking these areas of ThrottleStop

 

bah, I guess my 1.41 cpu-z is way outdated Downloaded the latest 1.56 version, it correctly shows everything, its a C0 revision But still I got no option to unlock the extreme processor's features?

 

Thanks I thought I was supposed to click the TS icon in the taskbar. So I clicked "unlock max FID/VID", but it didn't show anything or notify me any changes? I right click again, the option "unlock max FID/VID" is still there?

 

Make sure EIST is ticked

 

Thanks for the help, but "EIST" is grayed out in the menu.

 

Ok might be best to ask the ThrottleStop questions on the ThrottleStop Guide thread, unclewebb is the creator...of ThrottleStop

 

Funny I was just talking about that.

You probably just need to add extreme=1 to the throttlestop.ini config file.

 

Edit: NVM, Problem fixed Trying to OC this x9100 now.

 

have you tried turning TS on and adjusting the multiplier???