Asus G73 GSOD - Perfect Immunity Configuration

The Theory Behind this Thread

- The Asus G73 GSOD problem is a puzzle to us. Lots of people are messing around trying to figure out the puzzle. Moving pieces around, making changes, experimenting, testing.

But think about doing a real puzzle the kind with a whole lot of pieces and the picture that you are trying to put together.

Would it not be many times easier for you to figure out the unsolved puzzle if you had a completed version right next to it already finished for comparison?

That is exactly what we are doing here, my G73 is the completed puzzle you can work backwards from the finished product many times easier than you can try to build from all the pieces with no point of reference.

Hope that analogy made sense

____________________________________________________

System Details

-Asus G73-A2

-Purchased @ Release from GentechPC.com

-Bios came with first release but I updated to 209 recently

-Drivers have been updated two times to some leaked driver releases. I am not using stock drivers.

- No GSOD in any situation ever

- My Review of the G73 - HERE

_______________________________________________


I have seen a lot of threads about a G73 GSOD problem, and they are causing a lot of negative impressions on what could be one of the best if not the best laptops out there.

I peek into these threads here and there, as well as the "how to fix it" threads trying to figure out what is the real problem.

Is it hardware? Is it software? Drivers? Temps?

I do not think anybody knows for sure, but I have had to post more than once in those threads that not every machine has a problem, there are some without any problems at all including mine.

Its not just people that do not play games, do not change drivers, or do not know about computers that are not suffering the problems. So lets figure out why it works for some, and not for others.

Firstly I noticed one very important trend that many people say they only get GSOD on new drivers and not the stock drivers. This to me is significant proof that its most probably a software/driver issue and not one of hardware. If it is a hardware issue there is probably nothing you can do other than RMA but if its drivers/software you can fix it yourself.

The problem with the stock drivers is they do not include some fixes for newer games like BFBC2 load times or the ability to run the FFXIV benchmark this is why many of you computer savvy guys need the ability to use newer drivers and have tried to update.

So here is what this thread is all about. Its my way of showing you a perfect system and explaining everything I can think of about my system so you can try to emulate it and hopefully get yours in the same state. If we can figure out why mine works and others do not we may just find the cure for the GSOD problem.

Ok so first off is Furmark this is a GPU killing test, this is the defacto way to cause your system to GSOD if you have a problem. Even a problem free system could have error from this test (it has fried gpus before and became known as a program to break gpus) by no means is any game or program ever going to put a load like this on your card, this is as bad as it gets.

I ram Furmark 1.8.2 the newest version with the default settings for 1 Hour temperatures got kind of scary but I let it run anyways and had no problems. Again this is no stock video card drivers anymore look at the screen shots for information & proof.





So 1 hour no issues. You can see the driver version on the GPU-Z and also the CCC its 8.74.

These drivers are what we called the "10.6 Alpha drivers" it was a leaked driver that contains all the VID&PID info for a direct install onto the system with no modification. So no mobility modder and it is not targeted to be a mobility driver.

Download Furmark 1.8.2 - HERE
Download the "10.6 Alpha Drivers" - HERE & HERE

These drivers do work for FFXIV (coming up next) and also have the BFBC2 loading time fix.

Since I know these drivers work I have shared them and a few people still had issues. So I think its HOW you install them or what you have done to your system that maybe matters.

To my best recollection of memory I simply installed these via running the installer I did not use driver cleaner, driver sweeper, or anything else.

I also never installed any modified drivers or official mobility drivers. I have a feeling that maybe doing this has corrupted or changed something in the stock Asus setup/drivers that is causing the GSOD for people. I only use the driver sweeper/driver cleaner process after I find I have problems with a normal install.

So what I am suggesting is try these drivers and see if it fixes your GSOD problem, if it does not do a total system restore with the install disks so you can get back to the stock factory state and then install these drivers. If it didnt work before and it does work after we know the GSOD problem is something to do with all those nasty non compatible drivers floating about. Its very possible they changed something that you cant easily change back without a full restore.

Here are some shots of the CCC and my settings.

CCC Hardware Info



CCC Software Info



CCC PowerPlay Settings



PowerPlay is important that it be on high performance if you find your scores low in benchmarks or games not running at the right frame rate.

I never run the system with Twin Turbo on. In my benchmarks when I first got the system I found that scores were lower with it on.

I also seem to remember there being some kind of Asus VGA driver installed in the system in the programs listing. I checked and I do not have that, so if you do you may want to un-install that from the add/remove programs in the control panel.


Next up after testing Furmark I decided to test the FFXIV benchmark so many people are having problems with.

But if its worth doing its worth over doing so I overclocked the system for this test to make it even harder.

Cpu Overclock to 1.75guz and GPU is at 850/1100

Results:



Settings:



So I ran it on High and had tons of stuff running in the background as well and it completed without any errors or artifacts.


Thats all I have for you guys for now. If you can think of any other screen shots, settings, or information you need to compare to my system let me know so we can crack this case.

Read the review too if you would like to see some pretty pictures and lots of benchmarks so you know what the G73 is capable of doing.

It pulls my heart strings to see this wonderful piece of equipment not working right for so many of you.

 

Reserved All updates will be posted here as we make progress.

 

Hopefully HellCry won't come in here and try to burn the place down . I'd like to see this thread play out with real co-operation and have lots of tests. I'll be installing that driver you provided tomorrow and see what happens. I've changed the drivers so much that the experiment might be a bit far from scientific. If I have issues, I suppose I'll use my backup disks and go from there.

 

Gary from Asus already acknowledged that it's a timing issue with the video BIOS in the G73JH. This is why some people get the GSOD and other's don't and everything in between. Their busy with a fix now.

The new video card landscape for laptops is that both Ati and NVidia decided that they have to release frequent drivers also for laptops because there are so many games and so many complex issues they simply have too. Every month there are lots of new games identifying bugs and performance issues. So the driver needs fixes.

Ati probably already released a new video BIOS months ago because it identified issues with getting better performance. Asus simply wasn't reacting on this. And since it was the first with a 5870m it has a BIOS that doesn't work with the newer drivers.

 

Yes I hope people actively participate and appreciate the thread. I gave up lots of very valuable starcraft 2 time to run the tests and put the thread together

I wouldnt bet on anything, if it was a timing issue that should mean everybody would have it and everybody would suffer the GSOD. The fact some do and some dont means its a variable that is not constant between each machine so it has to be either a user controlled variable or a random hardware issue.


I feel confident it could be a very simple but incredibly hard to find user variable because I have seen many very interesting "bugs" in video card drivers over the years that had the most obscure and easy fix.

Most of the time just discovered by dumb luck, it increases the chances of finding a fix if you are comparing a working unit to a non working one.

Some examples in just the last year.

Problem: ATI Eyefinity mouse cursor corruption, appears at random, only fix is to restart the machine. However some people never had it.

Solution: Mouse Pointer Trails enabled in the control panel. With this option on the cursor corruption never happens.


Problem: Video Card will not downclock to 2D clocks and stays in 3D clock speeds constantly.

Solution: Lock the ATI overdrive option in the CCC, there was a bug in a recent driver that with overdrive on it would keep the machine on 3d clocks full time even if you did not move the sliders.



Problem: Cant enable aspect scaling option in the CCC it is greyed out so my games when played at a lower resolution stretch to full screen.

Solution: The CCC wont let you change scaling options if your current desktop resolution is set to the native of your display. Change your desktop resolution to something lower than native for a moment and the aspect scaling option becomes available. Now when you go play a game with a lower resolution it will use the aspect scaling.


And here is a Asus G73 issue that was similar, the Creative Audio software crashing Steam and having issues with certian games. The fix for most was to uninstall the software but all you had to do was turn off bass boost or any of the other advanced creative options.

Its these strange minor settings and configuration issues that lurk on the computer that I think is causing the random GSOD issue with the G73. I think it has something to do with the drivers/software behind the scenes.

It could even be the Overdrive option for instance. If yours is unlocked then lock it and see if you get GSOD still. Try just the opposite if its locked unlock it.

If we explore things like this we may find a solution, an comparing to a machine that is working 100% is the best way to know what things to try.

 

I signed up after reading all of this.. you are the only person that I have seen during my 2-3 weeks of research to give hope in purchasing the G73! CHEERS BIG MAN! im going to try harder to beg my parents to buy me this now!

 

You seriously haven´t seen anyone else with 0 problems? I mean even I have "seen" atleast 4-5 + myself with 0 problems with our G73.

Am I really THAT lucky to have gotten a G73 with no problems, I think not. As so many times over stated this machine sells like butter ALL OVER THE WORLD and I can bet you my monthly salary that MOST people have no problems, atleast that they themselves didn´t do.

Sry for going OT. Nice thread Vicious, hope you guys get it solved!

 

Agreed, I haven't seen any of the video corruption issues reported on this forum for this particular model.

The only issue I have currently with this machine is a small bootup issue where it locks at the ASUS BIOS screen infrequently.

However, I wish all who DO INDEED have this issue, good luck!

Just to add to the thread and keep it OT:

G73JH-X2 BIOS 106
ATI 10.6 Official Drivers (upgraded from 10.1 which was upgraded from 9.6 original using the standard ATI driver uninstaller in Add/Remove Programs).

 

Agreed, I'm one of the lucky ones who've never GSOD'd. Good thread Vicious, I think that this will put to rest some of the bickering and insinuations that folks are either not knowledgeable or are outright lying about their laptops.

I'm running G73JH-X1 BIOS 209
ATI 10.7 Official Mobility Drivers
Ziddy's vBIOS

Works in: Crysis, Borderlands, Gothic 3, Dirt 2.

 

Hey Vicious, its nice that youve taken time to help out. Ive tried the drivers youve linked (again ) doing the following:
- Installing over stock 10.1 drivers: GSOD
- Uninstall the previous drivers via the control panel, reboot to safe mode, using regedit deleted all of the ATI hives in the registry (done it so many times ive memorised them lol), than installed the drivers: GSOD.

Ive already tried installing reference 10.5,10.6,10.7 and 10.7a drivers over the stock ones, on a fresh windows install, and after cleaning the registry and it made no difference for me. Only the 9.12 or 10.1 dont give me GSODs.
Did this on both GPUs ive owned so far (waiting for the 3rd).

Running bios 209 btw. Also tested 211 but didnt like it.

But Vicious i believe its very plausible it could be the vBios timings or the hardware itself - since i (and many others who have problems and have done this) get no GSODs at all with the MSI or Clevo vBios (but you loose HDMI and VGA with those) and as some reported there are supposed to be two different vRAMs used for our GPUs. All GPUs ive had so far had the Samsung vRAM and some people are reporting having Hynix vRAM, so it could be the Samsung one cant handle the timings set in Asus vBios on the newer drivers (which as ATI told me in an email response have some clock changes to adress 2D stability issues) and the Hynix ones could. Or perhaps it could be some manufacturing error. I must be real unlucky if its a hardware issue, since i had Ken test their laptops on stock in their warehouse and the service department here did the same favour for me and im waiting for the 3rd replacement GPU to arrive and i still havent found one that would work... Would make most sense if it would be the vBios AND the different vRAMs, since that would explain why the MSI/Clevo vBIOS work and why some people have no problems becouse there are different vRAMs from which some can handle the changes and some cant.

Have any other ideas? Id be happy to test them.

 

Are you sure the Hynix is the vRAM and not just the system memory? I think some folks had that twisted.

 

I dont know, i just heard reports from members and some techs too (though not the most reliable ones ive met), but its very possible someone got it mixed since many of us (including myslef) got the Hynix sodimms. Im going to keep ordering replacements until i get a working one though. Be it Samsung or Hynix vRAM. Ill let you know and post photos if i get a Hynix one.

 

"I wouldnt bet on anything, if it was a timing issue that should mean everybody would have it and everybody would suffer the GSOD. The fact some do and some dont means its a variable that is not constant between each machine so it has to be either a user controlled variable or a random hardware issue."


EXACTLY!!!! If it was timing, then everyone SHOULD have the same timing and EVERYONE should have the GSOD issues. Still putting money on it being hardware. Fixable in BIOS updates maybe, but still hardware.

 

I thought I saw Hynix as a writing on the video RAM. It is possible because if you look at this Clevo thread:

http://forum.notebookreview.com/sag.../465201-clevo-w870cu-review-ati-5870m-37.html

And then the picture mentioned there:

http://img.ruten.com.tw/s2/1/b2/80/21001147253376_26.jpg

It has Hynix video RAM.

 

Did you even look at what i posted? It could be explained by usage of different vRAM on the GPUs. Perhaps there are really some Hynix ones that work fine and just the Samsung ones are affected/uncompatible with those timings giving us GSODs as a result. Im not saying its so, but it is a possibility that has to be considered until we find out for sure if its true there are Hynix equiped GPUs floating around or not. Gary_Key could clear it up for us. Ill pm him and send an email to Ken and ATI to see if i can get an answer.

And a BIOS has nothing to do with it. The vBIOS is what is responsible for the GPU running.

 

The Clevo cards use some sort of Samsung vRAM. I had it on mine and theres a photo with it visible on the first link you posted.

 

The timing incompatibility issue could also be marginal, which is why some don't have an issue, and some do. Would also explain why overclocking the vram helps, since it can now keep up with the new timings. If there was a new vbios cooked up to match the ATI Catalyst drivers (which seems likely seeing the new MSI and Clevo vBIOS pop up around 10.3 release) then the new vBIOS Asus is requesting should fix things up.

 

Try my settings, I also had many GSOD issues and my g73 is running great nowadays

BIOS: 209
vBIOS: NO CHANGE
CCC: 10.7a (installed over 10.1 with P4GH uninstalled)
GPU Overclock: 705/1100 (seems like its a must)
P4GH: 1.1.38

(also using lastest intel turbo boost, but this doen`t seem to be part of the solution)

I hope it works and let you enjoy your g73 as much as I`m enjoyning mine : )

 

not everyone wants to overclock their video cards, and I have problem with overclock mine even just couple tens increase on memory will give me artifact. The idle temp is around 59~60 degree celcius so I believe temp is pretty good for ASUS TIM.

But MSI vbios really does the trick for me, no more GSOD crap anymore.

 

Overclocking only helps to make my GSODs less frequent, but doesnt solve the problem.

 

Sorry, HellCry, but I have Samsung video memory and my g73 does not GSOD. I too thought that maybe it was the video memory, but I guess that's not the case. I'm running the same vBIOS version as Vicious, what version are you running?

 

So its not different vRAMs after all but could just be that they arent made entirely equal by Samsung as Chastity suggested...
Ive tried the stock and every modded Asus vBios i could think of and none worked to get rid of the GSOD problems, some just helped to reduce the frequency at which it occured. The Clevo/MSI ones completely eliminate the problem, but they also leave me without HDMI and VGA which i need so im back to running a self modified version of a Asus vBios and 10.1 drivers for now.

I really wonder if theyll actually release a fully working solution, but im not holding my breath. I mean the keyboard missing keystrokes hasnt been solved for over 4 years and has plauged many of their models and the vBios update for the W90 model didnt completely solve their problems either... We can only hope it will be different this time round.

If Asus wont fix this i think ill be able to get a full refund or an exchange for the new G73JW here, since we have a law if the product still isnt fixed after three Service attempts/exchanges, you have the right to demand a full refund or an exchange for an equally valuable product.

 

Samsung VRAM suppose to be the best, I remember on the desktop Video Cards of ATI 9700, the first AGP 8X.

People try to modify their card from 9700 (or 9800 I can't remember) GS to 9700 Pro and found out only the card that has Samsung VRAM (the best) can be modify from 9700 GS to 9700 Pro via vBIOS update.

 

It was just a theory. Anyway there seam to be some good cards out there as people are reporting... I just cant find one myself... Be it extreme bad luck or the quantity of bad ones i dont know, but i just cant get one.

 

Judging from what you've said so far hellcry, it's obvious that it is a software/vbios problem, especially if you can use clevo/msi vbios with stability. I wonder if they'll just end up building on a clevo vbios and add in hdmi/vga

 

Actually, it's not obvious. Assuming we're all running the same vBIOS, it doesn't make sense why only some G73s can run reference drivers unless there's a hardware difference. But that's a big assumption. What version of the vBIOS is everyone running? If there is a difference, I'll see if I can copy my rom and post it here.

 

Forgot to post my version, I'm running 012.17.000.004.

 

We've investigated that option already here, and we all have the same vBIOS when shipped. The MSI/Clevo vBIOS is probably built on a reference vBIOS from ATI, and then modded for their hardware options. I expect the same for our Asus models. ATI will build one based on their reference one with out hardware config, and then all will be well.

 

It makes about as much sense why some people can use some pretty hefty overclocks vs people that either can't overclock at all/very little. Timings/whatever differ from chip to chip, this is not new and happens to pretty much every silicon production chip made in the past 10 years - what usually controls these timings and their thresholds are the bios/vbios, sometimes drivers are given access too.

If a clevo vbios is working when the asus one isn't, chances are the clevo one has a greater timing/threshold range and at the core is just the ATI chipset instructions, nothing else? It's just a thought but maybe clevo and msi use different motherboard controllers for hdmi and vga than asus and alienware do?

 

I realize that silicon dies differ from chip to chip, but there must be a relatively high degree of variance (some maybe even out of spec) for different G73s to have such a range of stabillity.

 

Is this actually true? Is Asus really working with ATi for a proper vBios? Or is this just wishful thinking?

Bleh, unfortunate, I still don't understand why Asus half-assed the vBios. I can't imagine it takes that long to do a proper one, could probably just assign one engineer and one day...

The Asus here I am using has perfectly working vBios with all the various automatic downclocking etc. Wonder how this got passed them and they don't seem to care.

 

I've been thinking there must be some sort of a trigger for the GSOD issue as well.

I've tried everything to see if my laptop will eventually succumb, but luckily for me, it remains rock solid.

I didn't take anywhere near as much time and effort as you have compiling all of this great information, but if you can come up with anything you'd like someone else with no GSODs (ever) to test with you, let me know.

I'm down for RMA'ing (if my card ever does crap out) if it might help solve the issue for those affected.

 

Every company uses their own controllers, thats why after flashing to the MSI/Clevo vBios our HDMI and VGA dont work. If we could just change that part of the vBios wed have a fully working laptop, but i have heard of no such tool released to the public to do so.

 

The new vBIOS and drivers were being tested at QuakeCon. Beta testers should have it very soon, if not this week. As reported in another post, Gary says they haven't crapped out yet.

 

Hi everyone,
I basically signed up here just because of this thread, as I got a G73 ( the one from best buy ) and evil seems to start awakening in it since 2 days.

Never installed drivers or Bios and everything was fine for a month, however the corrupted mouse glitch appeared ( thanks Vicious for the solution you gave, its very helpful ). At the same time I got another issue: screen getting completely messed up when going fullscreen in 2D games on my secondary display only ( couldnt find anyone getting the same bug though ).

After that I installed the drivers you give on the first post ( doing it with the express install ) and no better results.
The day after everything went back to normal and I thought the drivers did it.
Today the mouse is freaking out again. Its just so random and im not installing anything weird on my computer.

I just have two big question: is the mouse corruption usually coming with all the other trouble like computer crashing in games ? Has anyone been able to cure permanently a G73 diseased with GSOD ?
If this thing is the beginning of a whole serie I d rather RMA...


If there s anything I can do to help you out with this "puzzle" with my little bits of informations, I d be glad to help...

Thanks.

 

See the Sticky thread about Fixing your GSOD Blues