The Definitive Guide to Acer Aspire Gemstone Blue Throttling Issues

Joegonz - undervolting is what we recommend in order to alleviate the problem. We just have suggested a different software to do it (RMClock). No hardware damages have ever been recorded through undervolting - the worst thing that happens is an unstable system. We've been hoping for a BIOS update too, but Acer's stubborn and insists that it's not a systematic issue, just an individual one.

Nickc24 - For Pete's sake, read the first post in this thread! It's called the Definitive Guide for a reason!

Chrism - Possible. I've seen that the 6930G has less instances of throttling than the 6935G. It is a nasty little bug, but the system runs fine when the CPU isn't stressed to the limit. I'd recommend the 5930G or 5935G instead - neither has even been reported to throttle.

 

Another issue of the acer 6935G and other similar models is the disabling of the intel VT(for use in virtualization).
There was a guy from marcansoft.com that was able to fix this problem enabling intel VT by patching the InsydeH2O BIOS
of acer. I wonder if the same can be done for the Throttling Issue. Someone can study this and find the address offset(in the bios flash file) for the temperature limit of the gemstone notebooks.
I know this is a longshot but since ACER refuses to fix the intel VT and does NOT acknowledge the Throttling Issues,
It may be the only solution w/o resorting to undervolting. A Bios patch by the concerned users. It would be the cleaner approach for me (but more dangerous also).
So for those with the stomach and know-how for Bios patching see link below.

http://marcansoft.com/blog/2009/06/enabling-intel-vt-on-the-aspire-8930g/

 

Hi,
I read the post but it says Throttling Issues, what does throttling actually mean? Ive actually read 12 pages of this thread. I have dowloaded that rightmark, tried to set my cpu as mobile, it closes but doesnt restart, when I start it again it is back to desktop. Also which colour is throttle? The one on the left is 1995 and 1995 on right, although it goes down to 1994 on the right a few times. Cheers

 

To really fix the throttle issue, you got to deal with the dust that is trap inside the heat sink's fin, by cleaning the Notebook fan, it won't help to fix the throttle issue. But you have to remove the fan, and remove the dust. My throttle issue gone even without the CPU undervolt tweak.

 

Joe - possible, but the 6920G uses a Phoenix BIOS.

Nick - Throttling means that the CPU is running at a given multiplier, but using on-die clock modulation (ODCM) to force it to run at a slower speed. It's less useful than Intel SpeedStep, which changes the multiplier and CPU voltage as more or less processing power is needed. The throttle can also be engaged if the CPU is getting too hot and needs emergency cooling - for the 6920G, that happens altogether too early. And the throttle is the pink line (clock is red).

Law - Again, different things work for different people. There is no one-size-fits-all solution until Acer gets off its behind and releases a BIOS update that fixes the problem. And that won't happen unless everyone who gets the problem calls Acer and tells them about it!
As far as I know, I'm the only one on NBR who's told Acer this happens!

 

Will this mean we need more NBR members to tell about it?
If you all want to take some risks, you all can try modded BIOS which adjusted to stop the buggy throttle problem.
weinter(an user here) know how to make it but I forgot 1 important link(modded BIOS) he gave me last time.

 

YES PLEASE, DarkSilver.
The bigger a stink we make about it, the sooner Acer will release the fix we need.

 

Problem is Acer isn't going to fix a old model they are no longer selling that is business minded behaviour.
Anyway here is the modified DSDT try overloading using Microsoft ASL Compiler
No one has told me if this works or not.

 

RMClock in the CPU Info page reports my CPU's minimal and maximal Req.Vcore (VID) as 1.0500V to 1.2500V.
The CPU is a T5800 (according to Intel, VID = 1.0750V to 1.1750V)
Does anyone know how to get the BIOS to recognize the proper VID?

 

Yup - go to the Advanced Settings page, and look on the bottom for "CPU Type Selection." Change that from "Desktop" to "Mobile," apply, and restart RMClock.

 

ok have the acer 6920 g t8300 bios v1.14..
i did run rightmark and orthos
run orthos for about 20 minutes..
and did not find any change in the core clock and throttle..
did stay the same about 2395mhz..
here is a recording
http://www.youtube.com/watch?v=X02v7RnRyQ8

 

You should play some games like Left 4 Dead or GRID instead of stressing it about 20minutes only. Play these games(both CPU and GPU intensive) is better than stressing test.
If you really want to do a stress test, use the small FTT(if not spell wrongly) and do it for a few hours(at least 1 hour).
Then, check the clock and throttle again.

 

I forgot to mention its already on the "Mobile" setting!

 

Cyberpunk, it's not really a BIOS issue so much as RMClock misreading it.
Feel free to go in and modify it yourself.

 

elevating your laptop helps a lot to prevent throttling. use a laptop cooler and allow some space below your laptop for the air to circulate.

sometimes my 6935g throttles even when just surfing the web when using my laptop on bed. i believe its because the fabric is blocking the air circulation below my laptop.

just wanna share.

 

Running Windows 7 -> 0.925v using the guide, no throttling happening.

Played left 4 dead for 50 mins with zero issues

 

WoW! You can achieve 0.925V, it is impressive. >.<
With these temperature, I am sure you won't throttle on Windows 7.
Make sure you do clean up the dust regularly. Dust can be 1 of the culprit too.

 

I was going to give him a quick once with my can of air.

This laptop has some amazing powers.

 

This method will work.

I have helped some people on this thread to stop throttling:
http://forum.notebookreview.com/showthread.php?t=298342&page=22&highlight=tzs0

Just read that thread ^^^ backwards to find how it's done.

It's very simple to do it, if anyone want help feel free to ask.
I will provide the program to extract the dsdt table. Also you may get lucky and be able to change fan speed values.

 

I'm definitely giving this a shot. Will try this myself but I never tried changing a dsdt table before. Heck I didn't even know there was such a thing .

But if this works you will be my lifesaver.

 

I was just looking at what another user posted on page 4 of this thread.

It seems you dsdt table does have fan code.
Lets focus on the TZ scope.

Here is what I'm just thinging so far:

Scope(_TZ_)
{
PowerResource(FN00, 0x0, 0x0) //this is fan code
{
Method(_STA, 0x0, Serialized)
{
If(LEqual(FANS, Zero))
{
Return(Zero)
}
Else
{
Return(One)
}
}
Method(_ON_, 0x0, Serialized)
{
}
Method(_OFF, 0x0, Serialized)
{
}
}
Device(FAN0)
{
Name(_HID, 0xb0cd041)
Name(_UID, Zero)
Name(_PR0, Package(0x1)
{
FN00
})
} ThermalZone(TZ00)
{
Method(_AC0, 0x0, Serialized)
{
Return(Add(0xa48, Multiply(ACTT, 0xa, ), ))
}
Name(_AL0, Package(0x1)
{
FAN0 //more fan code
})
Method(_CRT, 0x0, Serialized) //Critical temp
{
If(DTSE)
{
Return(Add(0xa48, Multiply(CRTT, 0xa, ), ))
}
Return(0xed0)
} Method(_SCP, 0x1, Serialized)
{
Store(Arg0, CTYP)
}
Method(_TMP, 0x0, Serialized)
{
Store(CRTP, Local0)
If(And(Local0, 0x80, ))
{
Subtract(Local0, 0x100, Local0)
}
Return(Add(0xa48, Multiply(Local0, 0xa, ), ))
}
Method(_PSL, 0x0, Serialized) // Passive cooling list (throttling devices are listed)
{
If(MPEN)
{
Return(Package(0x2)
{
\_PR_.CPU0,
\_PR_.CPU1
})
}
Return(Package(0x1)
{
\_PR_.CPU0
})
} Method(_PSV, 0x0, Serialized)
{
Return(Add(0xa48, Multiply(PSVT, 0xa, ), )) //This value is the temp at which throttling is triggered. I suggest changing it to 100C (E93).
} Method(_TSP, 0x0, Serialized)
{
Return(0x96)
}
}
ThermalZone(TZVR)
{
Method(_TMP, 0x0, Serialized)
{
Store(VRTP, Local0)
If(And(Local0, 0x80, ))
{
Subtract(Local0, 0x100, Local0)
}
Return(Add(0xa48, Multiply(Local0, 0xa, ), ))
}
Method(_CRT, 0x0, NotSerialized)
{
Return(0xed0)
}
}
ThermalZone(TZVL)
{
Method(_TMP, 0x0, Serialized)
{
Store(VLTP, Local0)
If(And(Local0, 0x80, ))
{
Subtract(Local0, 0x100, Local0)
}
Return(Add(0xa48, Multiply(Local0, 0xa, ), ))
}
Method(_CRT, 0x0, NotSerialized)
{
Return(0xed0)
}
}
ThermalZone(TZ01)
{
Method(_CRT, 0x0, Serialized)
{
Return(0xed0)
}
Method(_TMP, 0x0, Serialized)
{
Store(CLTP, Local0)
If(And(Local0, 0x80, ))
{
Subtract(Local0, 0x100, Local0)
}
Return(Add(0xa48, Multiply(Local0, 0xa, ), ))
}
}
}

 

Basically I would change this:
Return(Add(0xa48, Multiply(PSVT, 0xa, ), ))

to this:
Return(0xe93)


Now it should set the point at which throttling starts to 100C.
This has been tested in the thread I gave a link to and it worked.

If you want, just upload your *.asl file and I will mod it for you.
In order to upload the file, you will either have to zip it or just rename it to *.txt.

 

I think this will be the BEST solution if it works for everyone.
I remembered Tehsuigi did test this solution before but he failed.
I am not sure if that test was operated correctly.
It is too advance for me. When I am free, I will test out this solution.
Now, I am using my own solution and it works for me luckily.

 

I think Tehsuigi was trying to lower the temp reading.
While my advice is to raise the point at which throttling occurs.

My approach has worked for 3 forum users on the thread I posted a link to.

I can give simple and clear instructions if anyone needs them.
Also I can modify the file for anyone who is too lazy or doesn't want to waste time.

 

I'm @ work right now but tomorrow I'm giving this a shot

Will keep you posted!

And again, if it works, Thank you!!!

 

Moral Hazard, this could be the one-size-fits-all patch that could solve the problem!
Please post two things in this topic:
1. An easy-to-follow guide on how to patch the DSDT.
2. Screenshots showing your system's CPU temperature as high as it can go under a stress test - something in the 60s or 70s would be perfect. I just need proof that this does clear up the throttling issue (proof that other users found it useful would be nice, too).

 

Agreed. I smell it does work.
After all, we don't need to worry the throttle problem again.
So, I don't have to worry about the environment changes too.
I don't have any throttle with my tweak and undervolt as well as dust free in normal or cold area. However(didn't try it before), I don't know if the place is very hot. LOL.
Hopefully, this method will work.

 

Ok so here is the way to do it.

follow this guide untill step 7:
http://notebookequus.blogspot.com/

Now when you get to step 7, Please upload your *.asl file here.
Then download notepad ++.
Open the asl file with notepad ++.

At this point I have to stop this guide because I don't have your notebook so I need to see someones *.asl file.
I will continue this guide when someone uploads the file.

But basically you would change one line in the code to set the throttling trip point to 100C (or whatever you want).

once you've changed the code. continue following the guide ^^ from step 13 to step 15.

Please someone upload your asl file so I can make a better guide

EDIT: BTW I have attached a PDF + program that makes this easy.

 

I'm sorry I do not have your notebook. Also my notebook doesn't have throttling code in the dsdt. So i cannot give first hand proof.

I can give you other users results:
http://forum.notebookreview.com/showpost.php?p=5311525&postcount=143
http://forum.notebookreview.com/showpost.php?p=5325313&postcount=183
http://forum.notebookreview.com/showpost.php?p=5324388&postcount=182
http://forum.notebookreview.com/showpost.php?p=5326200&postcount=185
http://forum.notebookreview.com/showpost.php?p=5355653&postcount=203
http://forum.notebookreview.com/showpost.php?p=5355655&postcount=204
http://forum.notebookreview.com/showpost.php?p=5355669&postcount=205


It's all in this thread:
http://forum.notebookreview.com/showthread.php?t=298342&page=22

 

Ok, I've done all the steps and have an dsdt.asl

View attachment dsdt.txt

I had to rename it or it wouldn't upload.

Btw. this is from a aspire 8930g with latest bios 1.13 and win vista 32.

BTW I tried looking for your suggestion but I couldn't found it.

I think it had to be in here somewhere.

 

Hi,

I would like to give it a try too.

8920g

Bios: v1.14
Win Vista 32 Bit

I attached the file, renamed to a *.txt too.

 

I think, if i understood u correctly, this line has to be edited:

Method(_PSV, 0x0, Serialized)
{
Return(Add(0xaac, Multiply(PSVT, 0xa, ), ))
}
Method(_TSP, 0x0, Serialized)
{
Return(0x96)

to this:



Method(_PSV, 0x0, Serialized)
{
Return(0xe93), ))
}
Method(_TSP, 0x0, Serialized)
{
Return(0x96)


After editing it, i should rename it to *.asl and do steps 13 to 15?

If this is correct, i could do it by myself.

 

Posting my DSDT, 6920G BIOS 1.14.
It certainly does have the thermal trip information in there, but with a different hex code

Might give it a test run tonight after my video encode finishes.
Got a backup prepped and everything.

 

@spaanplaat, your code is slightly different. I think you can definatly still stop throtting but I will have to look at the ACPI datasheet again when I get home from uni.

@Abraxis, Yes you're doing it right.

@TehSuigi, I think you already know what you're doing

@All, before trying any of this, make a backup of your registry and anything else important to you.

If you get a BSOD when you restart, holt F8 and boot "last known good config".

 

Moral Hazard, if this thing works, I could not possibly give you enough rep.

 

Wow!! Moral_Hazard, Don't forget about me. I will apply you method sooner or later too(but need your help because I am noob on this). I also don't have enough REP for you.

 

As someone who dabbled in DSDT I like to point out the AAC = 2732
Does the number sound familar?
Celsius [°C] = [K] − 273.15 [K] = [°C] + 273.15
I just pointed this out for you guys to keep that in mind.
AAC is not the value you should be editing as it is a constant.
Return(Add(0xaac, Multiply(PSVT, 0xa, ), ))
By rights you should be editing PSVT because 0xa is a multiplier.
But because PSVT seems to a variable I am not sure it should be edited.
Before you do anything please download ACPI specification documentation.
I find HP implementation of ACPI is vastly different from Acer in terms of the structure.

 

Well, the idea is to remove the entire addition line and simply tell it to use E93, which is hex for 100 degrees Celsius in tenths of a Kelvin.
The code as is says to ad 2732 (tenths of a Kelvin) to the temperature of PSVT, whatever that is.

OK, I'm running the updated DSDT code, and the first thing I've noticed is that the fan actually responds to CPU load now!
Oh, and yes:

It's working so far.

EDIT: Goddammit, I spoke too soon! It STILL throttled after 7 minutes of ORTHOS on stock voltages.
I'm pulling a System Restore back to my original DSDT.

 

I think I customed a DSDT weeks ago no one tested it.
I found it here
My approach was different I suspected the throttling was due to _CRTT
In ACPI spec _CRTT meant CPU reaching critical temperature hence throttling starts.
Hence I modified it from 71 deg to 99 deg.
Maybe you would want to test it by overloading the tables while you are at it now.

 

@ Moral Hazard & Weinter

i've the AAC as well as Suigi. So editing it to e93 could be a solution or not?

And where can i change the CRTT? Because i'd like to increase it to 65° Celius, feels like it is lower.

Mine is an 8920G not 6920G, so could i use ur file Weinter?

 

Erm no you shouldn't DSDT for the 2 laptops are different.
I will try to get the AML out for you

 

Please ignore my previous post CRT refers to critical shutdown temperature which shouldn't be an issue.

Never mind forget what I said I still need more time to digest ACPI, anyway any anyone obtain the temperature at the point throttling starts?
Use OCCT if you have to it will be easier than understanding ACPI specification.

 

Well then, isn't that my luck .

As I have no idea what the values mean I guess I will be patient and let the real pro's handle this. But as I do not have this value but I have the throttle how is it triggerd on my lappie? Perhaps this is different then from your guide?

Perhaps if you can mod my dsdt to stop throttling you can do the same for TehSuigi with this approach. Maybe acer's dsdt is slightly different from Gateway's and eMachines.

Yes, I didn't forget about your dsdt mod, but knowing I have a different dsdt from the others I will wait and see what moral hazard can find to alter in mine.

 

can someone download and run OCCT.
Then post the graph to see which temperature throttling starts.
Rather than guessing the DSDT this will be more practical.

 

For anyone who needs it, here is the the ACPI spec datasheet:
http://www.acpi.info/DOWNLOADS/ACPIspec40.pdf

@spaanplaat, there are many ways to stop throtting.

One thing you might wan't to do is remove the throttling device list code:
Method(_PSL, 0x0, Serialized)
{
If(CMPE)
{
Return(Package(0x2)
{
\_PR_.CPU0,
\_PR_.CPU1
})
}
Return(Package(0x1)
{
\_PR_.CPU0
})

Not sure if it will work though. It's worth a try, but if you get a blue screen on startup remember to hold F8 and boot last known good config to fix it.

Anyway I attached the modified file with the passive cooling list deleted.
Probably will blue screen, I will read your code again hope to find a better way.

I'm guessing that If I cap your TZ temp to 80C it should stop throttling. Give me a min and I will upload another modified file for you, hopefully it should work.

 

ok, to cap the TZ temp I need to know what you would like the max temp to be.

It should be a value just under the temperature at which throttling starts.

I will add this code:
If(LNot(LLess(Local0, 0x?)))
{
Store(0x?, Local0)
}

Where "?" depends on what you want the temp to be capped at.

Let me know what you would like the max temp to be, and I will try to cap it there


This method should work for everyone, but it's better to use my first method where you just raise the throttling trip point.

Since spaanplaat doesn't have the PSV (throttling trip point) code, I gave him this second option of capping the TZ temp.

 

I extracted and decomplied the AML from aspire 8920G surprise!
It is the same as 6920G.

 

@weinter, do you have throttling?

If so, what temp does it throttle at?

 

I think my 8920G starts throttling when the CPU Temp is between 55° and 60° Celsius what is definitve too low because while gaming 55° is reached very fast.

So if i edit the aac -> e93 the Temp when throttling starts is at 100° C or what is the right way for me. Or is there another place where i should edit?

My file is attached few posts above.

I am a little bit confused now.

 

@Abraxis,

you should change this line:
Return(Add(0xaac, Multiply(PSVT, 0xa, ), ))

To this:
Return(0xe93)


That will change the trottling trip point to 100° C.

I modified your text file and attached it to this post, test it.

If you get the BSOD just boot last known good config.