Yippie.. First support for Extreme CPU's & P78xx

Hi,
I appologize ahead of time to the mods, I do not mean to double post but thought this deserves its own thread. Also if you could please sticky this too and if you want delete my other posts.....

It appears one of the Acer board posters, "Unclewebb", has created a program that lets P78xx users overclock their Extreme CPU's in windows. Not to steal his thunder or great work below is the threads link. You must have an Extreme CPU though to realize the overclock potential.

As a side note since the P79xx's are so similar to the P78xx's it may work for that series as well...............

http://forum.notebookreview.com/acer/489140-how-unlock-core-2-extreme-multiplier-windows.html

 

This is the best news ever. Just wondering how far is everyone overclocking. So far I'm at 3.45ghz stable.

 

My only question would be.

Is it actually Overclocking the CPU or just "saying" its overclocking the CPU?

Might want to run some actual benchmarks before we say its sucessful.

Not sure if it will work on the older models (68xx series) but i'll try it tonight if i can get it working and compare benchmark scores. It will have to beat my 11142 3dmark06 score on my 6860 to prove its actually doing something.
(3dmark is nice because its CPU bound)

But hey give some stock benchmark scores and some OC'ed ones (any benchmark will do really) to see what happens

 

Yeah if this works on the 68xx series that would awesome.

 

Thanks a lot for the info. Will wait and see now.

 

Yes it is overclocking the CPU. Last night I did some 3dmark06 runs to see and I gain 300+ points at 3.45ghz

 

Holy crap!!! this works! benchmarked using EVEREST

So far got to 3.6 Ghz with 13.5 multiplier

 

Omexis what was your voltage at 3.6?

 

Can someone email me the program? Work blocks the file download.

 

done..........

Editt; what are people getting with the OEM C0 compared to the E0 OEM or ES chips?

 

I'm running 3.45ghz with 1.250v on an C0 ES. Anything over this gives me a bsod.

 

Lucky I've got a good sense of humor or I might be offended by a statement like that.

I wouldn't put my name on it if it wasn't the real deal. There's a register hiding in these CPUs that Intel has been quiet about for a long, long time. I think it's called the steroid register.

I was wondering how much core voltage you guys are using? A tester on the MSI forum proved that VID settings all the way up to the maximum 1.500 setting was adding voltage to his CPU. For the 45nm CPUs, I like to stay under 1.40 volts actual but 1.50 volts is OK for the 65nm CPUs if you can find a way to keep them cool.

This new trick should work for any Core 2 based Extreme CPU including desktop Extreme CPUs and the older 65nm Extreme CPUs.

 

I was running at 1.25V, i haven't check for stability yet so i have to do some testing tonight and see what this chip can do!

 

Thank you sir

And thank you as well

No offence ment, its just with as locked down as these systems are i have to play devils advocate and tend to go with pessimism (we have had many fail attempts to date)

I will confirm that this does indeed work. I'm going to have to play with it a bit because i didnt have much time last night due to work and drills and everything else going on right now. But initial testing looks very promising.

I havent been able to run any benchmarks yet, but i was able to go straight to 17 multiplier and hit 3.4gHz on stock voltage before i Blue Screened. I'm going to have to play with the settings a bit and see what i can come up with as my theoritical maximum.

The problem is (what most will run into) is multipliers cant take you very high with out the blue screens. we really need to be able to adjust the FSB to reach the crazy numbers. BUt even able to run at 3.4-3.6 is pretty amazing

So all in all i say good job unclewebb and we all appreciate the work put into the tool.

Thanks

 

Thanks a lot You're a hero unclewebb

 

Wow, this is great news. I'm glad all those that got the X9100 can now release...the full power of this battle station...er, um... laptop.

 

Most laptops are pretty conservative with the default voltage so if you want to run 3.4GHz and up reliably then it's likely going to take some extra voltage.

Be careful with just cranking the multiplier up until you hit a BSOD. When you are overclocking in Windows like this, that can be dangerous. It's easy for a BSOD to corrupt some important file. Go up in small steps and do some Prime95 Small FFTs testing at each step. Even 5 or 10 minutes of Small FFTs will give you some idea how stable your new setting is.

There have already been reports of users screwing up their computers because they tried to go too high too fast. I know when the multiplier starts going up and up it's hard to hold back but try to use some restraint so you don't accidentally nuke Windows.

 

unclewebb, does your software allow other software like RMClock to take advantage of the new available multipliers?

 

Yeah i suppose i should have mentioned that as a warning as well. I just did it to see where my BSOD level was. I have no intention of using it that way for anything but it was neat to see it happen

In all things overclocking babysteps are always the best way to go about it.

Luckily i get a few seconds before i BSOD so i was able to readjust everything back to normal so i could actually relaunch the program

Would you happen to know what files are getting corrupted? As its strange that it would corrupt them.

Might be a good idea to create a restore point before diving too deep into the OC of the computer so you have something quick and easy to go back to should the worst happen

 

The guy that reported problems wrote back today and said everything is fine. Hopefully he will tell me what happened. ThrottleStop might have got the blame for something else that happened. Of course he had system restore turned off so his SSD could be a little faster.

The way ThrottleStop is designed is that when you are playing around, your new settings won't be saved until you click on the Save button. This way if you have a BSOD, you should be able to reboot without any problems and restart ThrottleStop without any problems as it will use your previously saved values and not your BSOD values. No matter how many safety features I put in, there is always some user that finds a way to screw things up.

Most Windows system files are well protected these days with a backup copy in case anything gets corrupted. I've heard about issues like this but have no direct experience with these problems. I've done some crazy overclocking and have never had any issues like this but other users on XtremeSystems where I used to hang out say they have.

Once you have used ThrottleStop to unlock your CPU to use higher multipliers or higher VID, you don't even need to leave ThrottleStop running anymore. It writes information directly to your CPU and will remain unlocked until the next time you reboot. I found that when using standby or hibernate mode that the FID/VID settings would be reset to default. I built an option into ThrottleStop that tries to detect that and if you leave ThrottleStop running in the system tray, it should be able to automatically restore your unlocked CPU settings after your resume from standby or hibernate.

Once unlocked, any software you run will be able to use the higher settings. After so many years of being tied to a pole, this is almost like a gift from heaven for enthusiasts.

Edit: I don't like picking on other programs unless there is a problem and when testing my QX9650 with RMClock in charge of FID/VID, I found a major problem.



At full load, two cores were able to use the maximum multiplier and two cores were stuck at the minimum multiplier and would not go any higher. ThrottleStop is in monitoring mode only in the screen shot. Even without ThrottleStop being used, you can confirm this problem by right clicking on CPU-Z and looking at each individual core. Two cores can get stuck with a 6.0 multiplier even when fully loaded.

The default multiplier for this CPU is 9.0 so RMC does recognize the higher multipliers, 10.0 and 11.0, but be careful when using it because on this Quad Core CPU, it has a few issues. It also locks up when I try to exit it. RMC used to work fine with other CPUs on this same board but it doesn't like this QX CPU, even when not using the FID/VID unlocking trick.

 

Running stable at 3.4 GHz and 1.225V. Thanks uncelwebb.

 

Yep its working like a charm.

I was able to get 3.6gHz stable last night by uping the voltage just a bit.

Once i get home and put this thing back on my cooler and can start playing with the GPU clocks i'm shooting for a 6860 that hits 12K in 3dmark06

Not bad for 3-1/2 generation old tech

Boy i wish i had this a year or two ago lol

 

And at the clocks you would have been running you would now be staring at a pile of ash, not a laptop.........

 

What are the max "safe" CPU temp for a X9100 ?
80° aiming seems OK ?

Another question:
If i do not change the voltage (and if it's stable), does running @ 3.06 ghz or 3.2 ghz will do the same temperatures ?
That's means changing voltage changes temp ? not the frequency ?

 

Increasing frequency but keeping the voltage the same will increase temperature as the CPU switches faster and draws more current hence the temperature will increase

 

What 4gHz on an x9000 with an 8800gts clocked at 700/900/1800 would have been glorious for the 1 run of 3dmark i could have gotten out of it before it burst into flames

Sacrifices have to be made for the overclocking gods to be happy

But no its just as much effort and wasted time we put into BIOS editing and dealing with Gateway (and lets not forget the attempts to flash the 8800GTS to an 8800GTX) its just nice to have the capability with out anything but stability checks.

I mean i'm not actually going to run the system at anything higher than 3.0gHz, but its nice to have the ability

 

k4b4l74: As long as an Intel CPU is not thermal throttling then you are running it within spec. Intel's CPUs do a great job of looking after themselves. I wish all hardware was as well made.

What I found when overclocking is that you will lose stability before your CPU will get to the thermal throttling point which is about 100C. The more you overclock, the cooler you will have to run your CPU so it remains stable.

Knowing this, all you really have to concentrate on is stability. As long as your CPU can pass some stability tests without rebooting or giving you errors and is not thermal throttling then you don't need to worry about the core temperature.

Here's how a CPU looks when you push it a little too far. I turned off the CPU fan on my desktop E8400 to see what would happen while running Prime95 Small FFTs. I overclocked it 20% and gave it lots of core voltage just to make sure it would heat up.

After 3 hours what happened? Absolutely nothing. I got bored so I stopped the test but Prime95 was still running just fine without any errors. The CPU bounced off the thermal throttle the whole time but it didn't do a nuclear melt down or anything like that. After that test I came to the conclusion that people worry too much about the core temperature of their CPU. The thermal shutdown temperature of an Intel CPU isn't until 125C to 130C so obviously Intel believes in their products ability to operate reliably even at very high temperatures.



If the RealTemp Thermal Status area says OK then that means you have not had one thermal throttling episode for even a millisecond since you last booted up. This information is stored in the CPU so you don't even have to have RealTemp running. It will check your CPU as soon as you start RealTemp and tell you if there have been any problems since you powered up. When OK turns to LOG that means at least one incident was logged. When LOG turns to HOT that means thermal throttling is in progress. Keep that on OK and you'll be OK.

http://www.fileden.com/files/2008/3/3/1794507/RealTempBeta.zip

 

Im really tempted to get one of these x9100 chips, can you guys link me a requatable seller of the chip, Thanks

 

Ok thanks for your answer unclewebb.

I've just pushed it to x12.5 (aka 3.325 Ghz). It's a C0 version.
Stable (i just test 30min with prima) with just 1.20 V

Temp CPU = 78/79°C

 

3.325 GHz is a great result with that VID voltage for a C0 chip. If you run your laptop in a meat locker you will have no problem adding some more voltage and increasing the MHz. The only problem will be keeping it cool.

As my picture shows, a 45nm Core 2 can run reliably up to about 3.6 GHz at pretty much any temperature but going higher than 3.325 GHz is going to start taking more voltage which creates more heat. For general use or gaming, most users with good chips like yours should be able to run reliably at about 3500 MHz. I just wouldn't try to run Prime 24/7 at those settings. With the right voltage I'm sure it can handle it but long tests like that create a lot of heat and a lot of stress.

 

appartenly stable @ 3.45GHz on 1.225v for my E0 X9100 (ES). Stress testing with orthos and/or OCCT causes BSODs though.

having said that, i'm able to do CPU intensive taskes like encoding/gaming for hours on end at 3.45GHz. Upping the voltage let me get up to 3.72GHz (14x multiplier) @ 1.3v before BSODing (no stress testing).

Aything higher than 1.225v under orthos/occt made my X9100 exceed its TJmax (running on a modded cryo LX (3x sythe s-flexs 1600rpm on 12v)) thereby initiating auto-shutdown. No BSOD. Extreme mega cooling could potentially let my X9100 run at a 700MHz OC.

 

TJMax for an X9100 is 105C and then the Intel specified thermal shut down temperature isn't until 125C to 130C. I doubt you are getting it that hot.

http://ark.intel.com/Product.aspx?id=35431&code=x9100

What happens when you start overclocking a Core 2 CPU, the maximum temperature it is able to reach will be reduced. It usually results in a random reboot. If you are not overclocking then these CPUs will run all day at 105C but as the clocks go up, you've got to keep them cool to prevent a reboot.

Keep in mind that VID is a requested voltage. There is vdroop built into the Intel design so typically a request of 1.30 volts only results in your CPU getting 1.25 volts. Every motherboard is a little different. A request of only 1.30V for 3.7GHz usually isn't enough.

Some manufacturer is going to have to come up with a mobile, liquid cooling rig to help you guys out with this new trick.

 

It's probably something to do with my EO X9100 being ES but i am stable at 3.6Ghz (juiced with 1.325v; idle temps approx. 68C - 70C). Intense gaming/SETI screensaver/general multitasking all do not induce a BSOD.

however, trying to stress test with orthos or OCCT is near impossible as my rig only lasts about 25secs before initiating auto-shutdown due to temps exceeding 100C. Individual cores hovers around 98C with overall temp being around 101C which is still below x9100 TJMax of 105C.

what's up with that?!

 

Is the auto shutdown orderly where programs are shutdown by Windows or does your laptop just instantly stop and you have to push the power on button to get it going again? It's possible that the bios has a built in feature to shut things down based on a temperature that is lower than the Intel specified temperature for thermal shutdown.

Your idle temperatures seem high. Is the heatsink making good contact with the CPU? If you can find a way to make it tighter, you might be able to improve those temperatures. Some heatsinks are poorly designed in laptops and don't make firm contact so heat transfer is poor.

 

Throttling 50% is controlled by operating system. Temperature point is 95°C,
OS shut down at 100°C; H/W shut down at 105°C

 

hmmm dont know if the heatsink to CPU contact is good enough; the last time i took it off to apply ICD7, i screwed it back on relatively tightly. To do it any tighter and i would run the risk of stripping the screw heads.

i'll have a look again and see what i can do to rectify the situation^^

 

Some times when you remove a thermal pad and replace it with ICD7, you can decrease the amount of contact pressure between the heatsink and CPU. Definitely worth taking apart and having a good look. Sometimes you have to do some re-engineering.

Is this an Acer design spec? The Intel spec listed in their datasheet isn't until 125C to 130C for thermal shutdown.

 

Yes it is.

 

3.7GHz @ 1.35v stable.

gonna try for 3.8GHz or 3.9GHz!

 

So just to verify, you have a program that will overclock the X series cpu's in the gateway fx. Has this been tested on the 68XX boards also? What chips do you guys recommend then for the 68xx boards.

 

Yes it works in the 68xx series. If you are going to OC past the 3.0GHz limit I would recomend the X9000 over the x7800/x7900.

 

Where can I find the x9000? I have seen one on ebay for around 550 if I recall and thats the only one I have seen. I wouldn't mind pushing the envelopes of this 6831fx but at the same time I don't want to drop 550 on just a processor when I could get an i7 gaming rig for not much more lol.

 

This is the link I was speaking of on ebay just as a heads up.

Intel® Core?2 Extreme Mobile Processor X9000 SLAQJ OEM - eBay (item 170505767380 end time Jul-26-10 22:25:37 PDT)

Where did you get your x9000 and what did you pay bro? Hows the Lappy feel with a processor that quick?

 

I got my X9000 from eBay for a little over $300. It is an "ES" so it was cheaper than the OEM in that link. For normal usage I cant really tell a difference between the stock CPU and my X9000 but in-game performance has seen a really nice boost.

 

I was thinking about grabbing a x7800, do you think the heat difference will be that drastic between the x7800 say at 3ghz vs the x9000 at say 3.4ghz?

 

I mean hell for less then 170 bucks shipped to my door this is starting to look pretty good =D

SLA6Z Intel® Core?2 Extreme Mobile Processor X7800 - eBay (item 320513970529 end time Jul-06-10 15:58:08 PDT)

 

The x7800 will get hot at 3.0GHz. The x7800/900 are the reason gateway had a 3.0GHz limit I think. I would recomend a T9300/T9500 over the X7800.

 

The only processors we can overclock though is the Extreme chips correct? I enjoy overclocking but again like I said, half a grand for the x9000 is just a bit more than I want to spend.

I hate this stock processor though, hell even wow can lag at times. I have it undervolted and it runs fairly cool, but still, I want something faster. I game a lot and do some video decoding/encoding at times and could use something quicker.

What are your recommendations? Oh and I do enjoy overclocking =D.

 

Yes the Extreme CPUs are the only thing we can overclock more than 200Mhz. If you want to overclock I would wait for a good deal on a X9000. I wouldnt spend $500+ for a X9000.

 

Yep it works beautifully on the 68xx. With my cooler i hit 3.8gHz effortlessly (but there were a few instability issues but not enough to majorly hurt the system). If i can hook up my ac to it i bet i could get to 4GHz with out too many issues (though it would likely become unstable)