Updates from Asus

Man I want to play FFXIV beta like on the 31st. Needs to be fixed by then!

 

I think its pretty safe to say it wont be fixed by the 31st. I dont think they work during weekends and i dont believe monday will be enough for them to put together a working vBios if it already took them this long to slap together a non-working one.

 

Hi I'm Andrea from Italy i wanted to thank you Kaeland and Castity because i finally follow their istructions delete all the AMD ATI new drivers and reistalled the 10.1 asus catalyst driver with Bios 211 and i dont have any more GSOD or vertical lines...for my opinion is better dont install drivers from the ATI web site and wait for a new drivers officially from the Asus driver web site that can be tested and certicifated for the G73Jh-A1 !!! dont try now to install other drivers if not is a nightmare!!!! Anyway is really incredible how a super internation company like Asus take 4 or more months for send a new working driver expecially for their " best " laptop....this is sad yes..

Thanks again people

Andrea Italy

 

Why are you double posting everything in two threads?

 

Im sure she mean no harm, perhaps its beacuse she´s not up to date with the rules of the internets

 

I'd say she is a he. Andrea can be a male name in Italian.

 

To get back to the G73JH, just found something out.
After browsing Asus support pages i found the ATK drivers as separate installs and if im right theyre the same thing thats used in our ATK package and theyre old as hell, far predating our G73JH.
I downloaded and tested the following:
- ATK_Hotkey_WIN7_32_64_z100056; Digital Signature by Asus 31.7.2007
It gives the keyboard the same functionality as our ATK Package, but just the icons that are displayed are different and there is no icon when pressing the keyboard illumination hotkeys, but they do work (also it comes with the annoying missing keystrokes if touchpad is enabled in bios).
- ATKOSD2_WIN7_32_64_z700010; Digital Signature by Asus 31.7.2007
This changes the nice on screen display to what we have in our package, the keyboard now functions and the on screen display that are displayed are what we get with out ATK Package, the illumination hotkey also displays the same on screen icon (but of course i still get missing keystrokes with the touchpad enabled in bios).
Theres also ATKMedia_WIN7_32_64_z200007 also dated 31.7.2007 but i dont notice any changes after installing that. Everything was already working before and works the same after installing it and keystrokes are still missing.

If i uninstall all ATK software and drivers (which means i get no backlight and the hotkeys arent working) it seams i no longer get missing keystrokes with the touchpad enabled (unless i just had a lucky day when i was testing this?, so can someone else thats also having these issues test uninstalling all ATK software and drivers too and reporting back).

Now if im not wrong its very possible that the culprit for the missing keystrokes isnt Synaptics at all, but the ATK drivers that Asus as it seams very likely hasnt updated for over 3 years (i mean they even have those old drivers from 2007 on their site as for Windows 7 which is also evident from their name, but its clear they were made far before Windows 7), so we are very likely actually using 3 year old drivers that were made before the G73JH was even designed that were just packed into a single package and the package could be given a new digital signature with a newer date and this could be the root of all troubles since as old as it is it is very likely that it would be incompatible with the newer touchpads and synaptics drivers from what i see (it was already curious that as many users report just uninstalling Synaptics drivers isnt enough, but you have to disable the touchpad entirely which just didnt make sense if it were really synaptics).

Gary mind giving more info on this now?

 

Update -
1. I am back, sorry about any delays but took a few hours off yesterday to take care of some outside obligations.
2. I will be setting up additional systems to replicate certain user configurations here in regards to the GSOD problems and we are working with AMD on reported bugs with the 10.8 drivers.
3. Missing keystroke and heat reported problems are at the top of the list also. More on Monday as I am traveling to HQ this weekend but will keep tabs on the forum in case there is something I can solve without needing an EE degree.

 

Don't forget to try Furmark and Left 4 Dead 2 (Hard Rain map) this time, both gives xSOD with this new vBIOS in a few minutes
I uploaded the memory dump analyses here, hope they can be usefull in some way.

 

Hellcry, thanks a lot for this; I no longer have any missed keystrokes and I can use my touchpad completely. Unfortunately, I often plug in my laptop to projectors and such but now I can't do that without the ATK package installed.

 

I didnt really mean not using ATK was a solution, to me its even worse than not being able to use the touchpad (i also want the function hotkeys and i want the backlight to work). I only posted that as an argument of proof that such old ATK could be the culprit causing the problem, but if you prefer to use it this way i guess it is an improvement if you only need touchpad and keyboard to fully work and dont need the backlights and hotkeys as much. Btw you can switch to external display in CCC, actually everything you can do with keyboard function keys can be done without them (those keys just use the functions in windows and other software - so they tap into device manager to disable wireless and bluetooth for example and use windows sound to change the sound volume and if you uninstall some software like synaptics, the touchpad function keys wont work, if you uninstall Catalyst drivers, external display function keys wont work and so on...), it just takes longer and is a pain in the rear to have to do it without ATK.

In either case we need a real solution and that means a new ATK or Synaptics driver - i still cant be sure without knowing if im right about ATK or not. With our temporary "solutions" we always loose some functionality in order to avoid missing keystrokes and thats just not acceptible.

Gary can you confirm this? Is the ATK as old as i think it is and could it be the problem and not Synaptics or touchpad hardware or have you got any other info on this?

 

The ATK uninstallation is better for me since I find it very inconvenient to disable touchpad in BIOS since this laptop is used primarily for school work (particle physics etc) and not gaming.
Deleting Synaptic's drivers also helps but I have to disable the touchpad in bios so I'm not sure if there's something with ATK and the touchpad that's messing stuff up. I know the U series uses the Elan touchpads so those probably don't have this missed keystroke issues.
I wish I had bought a U series and a desktop since this is so painful

 

Also test YouTube, since I got a WSOD from watching a video but not one from playing Combat Arms for 30 minutes or running Unigine twice.

I have my laptop at 705/1100 to prevent the WSOD, but every time I watch a video, the card changes clocks to like 500/1000 or something, and it makes part of the screen flash, which is not only annoying, but embarrassing if you're trying to show off your beast. :3

 

The ATK driver is very modular. It has independent executables to do various things. I would imagine the basic ATK stateful key inspection is probably old, but the package is probably rebuilt for each new model of laptop to include OSD icons for the new stuff or to make sure it calls the right backlight enabler, etc. I know nobody asked me, but I am a software guy in a different industry and this is how we do this kind of thing. You modularize it so you only have to rewrite specific parts when you ship for a different configuration. When we do a new install, it mostly works. I just have to hand write code for specific hardware differences.

 

But the thing is i installed the seperate ATK drivers from 2007 and everything is EXACTLY the same - the OSD, processes... so i suspect they just packed those individual installs into one package without changing anything and properly testing it... I cant proove it though.

 

I am agreeing with you. They are probably using exactly the same source code for stateful key inspection since 2007. They just redo the libraries that fire when certain keys are pressed and the icons that are displayed. This would explain why people with other similar laptop series have had exactly the same problem we do. I think you are 100% on the root cause being the ATK package. When you rebuild a source code project and only modify some called libraries, the compiler will usually touch the base exe files too since the things it is accessing changed. It is source identical to previous builds, but gets a new date time stamp.

I don't envy the programmer who has to fix this. It will not be fun. Either there is an if/else branch that is accidentally throwing away a keystroke or it's a timing thing where it is taking too long to process and the windows API is throwing away the keypress because it is late. It could also be passing every keystroke every time, but one conditional branch is passing it to windows incorrectly. I hope they have a kick butt regression tester on staff.

The fact that it doesn't happen with the TP is disabled in BIOS is a clue, Gary. The keypress is likely being lost in the section that gets bypassed when ATK determines that the TP is not present and doesn't need to parse that section of code. That is the first place I would look. The TP enable/disable is a separate executable so it could be right there where it is determining whether to fork a thread to call that exe or not.

OK... I'll stop. I don't have the source and I'm sure the people that wrote it know a heck of a lot more about this than I do. Furthermore, they probably could care less what I think about it anyway.

 

Sadly it all comes down to how much they even care as you said.
We know it can be fixed since it isnt hardware, but if theyll appoint enough resources to actually do it is another matter...

Well see if theyll even properly patch up the vBios first...

We should really stop doing someone elses job. Time for me to get some shuteye again.

 

i agree.. this must be the issue.. it seems Asus needs to make a whole new ATK..

 

Agree. My screen just flickers when before the video starts to stream. I'm on 10.1 and stock vbios

 

do both of u all have X1? I nver heard of this issue other than from both of u.. seriously , this vBIOS is epically screwed.. need a new one ASAP.. anyways enough ranting on my part for a day.. good night.

 

I bet there's some power setting behind this, are you using p4g or windows power management?

 

That's your GPU downclocking from 700/1000 to 200/500 when you have hardware acceleration enabled for flash.

It is not a fault of the vBios.

 

Just disable hardware acceleration in flash. Right click go to properties or something and un-tick hardware acceleration. Then your videos will run at normal clocks. Reason it downclocks to 500/1000 is due to UDV

 

wow jeez you know a lot and other people here!

still waiting for the updates to come ^^

 

wow no problems? could you try running the FF14 benchmark to see if it GSODs on you still?

 

I had to solve this for mine when i did the firepro softmod. Because my youtube videos and all would just appear as a green screen.

 

Did, still flickers.

 

I hope they make a final vBios that changes the way the clocks work or something, the flickering is embarrassing and annoying, though it only seems to happen on the 705/1100 clocks.

If the dang computer would work right on 700/1000... *rants*

 

yea I decided to bite the bullet and reinstall the stock vbios along with the 10.1 drivers.

 

The flickering happens if you make vRAM change clocks, especially by so much as from 1000 to 500 and back. I alerted Chastity its not a wise idea to change memory clocks from what they were in the default vBios since than you get the annoying flicker youre seeing (in worst case you can crash since not all GPUs will tolerate changing memory clocks the same), but i guess she wanted to do it anyway and if you downloaded and flashed her moded beta vBios your card is constantly changing not only the core but also the vRAM clocks. I dont see a point in underclocking the vRAM anyway since the main impact on temperatures is the core. I tried changing the vRAM clock from 1000 to 200 and it did nothing to lower my temps. Lowering the core for a step of 200 results in around 1-2C lower temps on both the core and vRAM and these results are consistent and repeatable, so it looks like the vRAM is only heated up by the core which is normal since its on the same heatsink. Anyway im using a moded vBios on which vRAM clocks are a constant 1000 and i have no flickering problems whatsoever.
IF you also flicker when just changing the core clock than something is weird on your GPU becouse none of my three GPUs ever did that, but as i said only flickered when changing the memory clocks.

 

Ok, so I decided to give the beta vbios one more chance. This time having changing the core and memory clocks to 705/1100 in amd gpu tool. Ran mafia 2 for about 20 minutes with no crashes, and on average I would crash after 5 minutes usually with stock clocks. So far so good.

 

You're right- I changed the core clock and no flicker. Change the memory clock and insta-flicker.

It's not like a huge part of the screen is flickering. I have my taskbar on the top of the screen and the flicker encompasses pretty much just the taskbar.
Regardless, it's still annoying.

 

This just reads as that my suspicion about the Hynix vs Samsung memory is right. I have Hynix video RAM and my results:

- down clocking to 500 MHz has a significant impact on temperatures; about as much as the core going back to 300 MHz.
- I get no flickering what so ever when I go back up;

If I do go lower than 500 MHz and go back up I get a grey screen. This is a problem Ati has for years now. From the 4000 series to now they still can't seem to get it working properly.

EDIT: some extra data

Started the laptop and waited 30 minutes. Room temp about 20C and TSS1 54C

I then downclocked the memory to 500 MHz with the AMD GPU Clock tool and left the core at 700 MHz. Within 5 minutes TSS1 dropped 4C and is now 50C.

So for me, it does make a lot of sense that the memory is down clocked.

 

Yep, thats the flicker i was talking about. Really annoyed me.

But how much is the difference between the TSS0 and TSS1? Becouse in my expirience the memory temperature cannot fall below the core whatever you do. If your TSS1 was decently higher than TSS0 than it of corse has room to get cooler otherwise i dont see logic in how it could fall below what the heatsink is heated up to. I have my memory just around 2C hotter than the core since the pads are properly installed and have sufficient contact, so the heat from the core heats up the heatsink and prevents the memory to drop lower than the core.
Show before and after screenshots of GPU-z and hwinfo32?

 

I have a similar behaviour.

If I downclock just the core 700 -> 150 I get a temp drop from 66 to 59.
If I downclock core to 150 and memory to 300 it drops to 52
If I leave core at 700 and downclock memory to 300 it drops to instanteneously from 66 to 61 and then slowly further to 58

I don't know which vRAM I have. How do you find out?

Also, I started playing with powersettings and powerplay. Although it reportedly doesn't work on G73, it seems to make a difference for me, but I need more data to figure it out. Either that or it is the pci-e link...

BTW: I get flickering no matter what I change, core or mem.

 

Post some before and after screenshots so i can see what youre getting on all sensors.

Ive had Samsung vRAM on all the GPUs ive had. You can find out by taking the laptop apart and looking at the chips.
The powerplay works on the beta vBios, but not on the older stock one where only UVD worked.

 

Well, things are certainly looking up compared to when I first bought this laptop 3 months ago in terms of the GSOD issues.

I'm lookin forward to seeing some more updates from the ASUS rep! I wanna get some ATI drivers up and running on this thing already so I can get some enhanced SC2 performance!

 

I meant the other temperatures also drop. The reason I use TSS1 because it's always warmer and the closest to the memory chips.

My room temperature is somewhat warmer now so the temps are now:

Core 700 Mem 1000
TSS0 55C
TSS1 58C
TSS2 55C


Core 700 Mem 500
TSS0 51.5C
TSS1 53.5C
TSS2 51C

 

Heres what i get. I used AMD GPU Clock Tool to set the clocks, let it idle for a few minutes and opened up GPU-z and took screenshots.


300/1000


300/500

As you can see i get practically no improvement from downclocking vRAM from 1000 to 500, but i guess you could get different results...
Even if it did lower temperatures for me as it does for you i dont think id do it since i really hate to see that flicker that it causes.

 

It might be that Hynix video memory runs somewhat hotter than Samsung memory and adds more heat to the heatsink. So when I drop the frequency it adds less heat to the heatsink.

 

I second PulsatingQuasar.

Downclocking vRAM leads to lower temps on my TSS 1. By as much as 8-10C.

Weird thing is I flicker when changing both core and memory clocks, and I've only had one GSOD during my entire ownership of the G73. And that's by overclocking a bit too far to play BFBC2.

 

i am on stock vbios and ccc 10.8
the screen doesnt flicker when i downclock gpu core and memory 300/300. it used to when i was still on catalyst 10.1
not sure what this means tho

 

Hmm weird... Cant say whats different.

 

how did you install the 10.8? when was your G73 manufactured?

 

been playing another hour with no gsod on the beta vbios with 705/1100 clocks. Can't believe I didn't try this before.

 

i didnt have to do anything fancy to get it working, just installed it to try it out and i was afraid it was going to GSOD. after 2 days of gaming and testing its gone without issues.
my g73 is manufactured in june btw

 

you just installed the 10.8 over the 10.1 or did you uninstall the 10.1 first?

 

yep 10.8 over 10.1

 

SO previous posts said the new VBios was a slight improvment, if any at all.....these newer posts seem to say that the new bios is doing some good, if you change clock speeds?
Me personaly:
Stock Drivers (10.1)
211 bios
stock vbios

I dont want to jump on the vbios especialy of there is a real fix comming from asus/ati soon..

with this config i am stable but i can not do the following:
cant run furmark (it jsut wont launch)
cant run a few games (prey, and a few others)

my 3dmark is comming back at about 12700ish which is about the same as the 10.8 drivers would come back with (if i didnt crash mid test)

Best buy told me to go screw myself, laptop return policy is 14 days. And there is no way i want to RMA back to Asus. lossing my g73 for 4 weeks jsut to get it back in non-working order, is not my idea of fun.

I am watching this and a few other forums, keeping my fingers crossed that a fix is shortly incomming. I have my botable flash driver ready to go with both the MSI and beta vbios, but i jsut dont want to pull that trigger yet.

 

​

Hi

I am in the exact same boat as you...

Stock Drivers (10.1)
211 bios
stock vbios

When I saw the first post about a potential fix it mentioned that it would not work with bios 209, so I upgraded to 211 in anticipation.

A few days later there was a beta vbios with incompatibility with 211

Not really interested in beta testing so just checking in every day for an official fix.