ASUS G73JH Problem - Very slow after using battery

Thanks. I know you're pretty active on the forums and appreciate the work you're doing. I have not used Throttlestop before (never had a need to) but looks like I'll be a follower soon enough!

 

It's been established that BIOS 205 and 206 make no difference towards the issue.

Basically what you received is book one, chapter one, page one, first verse of the script. Chances are good they didn't read and if they did, understand, your message.

The part I find very questionable is the recommended method of updating the BIOS, and then having the user go into BIOS afterwards. Sheesh. If you are going to update the BIOS, and you have to go into the BIOS anyway, why not do it from there, it would be safer.

Isn't CSR roulette fun.

 

That's what I needed to know. It makes it easier for me to ask that question than review 20 pages of this forum.
More time to work on ThrottleStop.

 

ThrottleStop 2.00 Beta 11
http://www.sendspace.com/file/wmsn1s

I still need to do some more testing and reboot 101 times to see that everything is OK but I think I've made some more progress.

I created a separate default profile for AC and for Battery power. When ThrottleStop starts up, if you are using this option, it should go to the profile that you have selected. If you are not using this option, it should start with the previous profile that you were using.

It will also use these options and go to these profiles if they are set when you switch between AC and Battery or between Battery and AC. The profile won't change when switching power sources if these options are not being used.

I adjusted how the C1E State works so it should lock it to whatever you prefer, on or off, and it should use your preferred choice when you boot up and have ThrottleStop starting up in the Task Scheduler.

If you notice anything that doesn't seem right then let me know so I can fix it up.

 

Confirmed corpfan1, when on Battery power the cpu clock speed will remain constant, although less than its maximum potential, when using ThrottleStop with a "Set Multiplier" of 12. And can be manually clocked back up by using another profile setting that has a "Set Multiplier" of 13. Although i only tried this action well after the 30percent battery level had been reached...

Unclewebb, instead of this manual input to change the set multiplier, maybe a feature that could be added to ThrottleStop? Something like... when on battery power - level reaches ("variable" - in my case something lower than 30%) automatically reset "Set Multiplier" to 13.

 

Unclewebb,

Beta 11 works great.

Only minor glitch, but not a big deal is...

The system tray icon won't change numbers if you switch between AC and battery while the machine is running unless you manually switch the profiles. In other words, if it is showing 4 and I plug in the AC it remains 4 (but the profile does indeed switch to 1). And, vice versa. However, if you reboot, and make the change while the comp is off, it will recognize the change for the icon.

Like I said, small glitch, but doesn't affect performance at all.

Thanks!


Now, if only you could write a program that would enable the keyboard backlighting prior to the login screen, so I could type in my login password while being able to see what I am typing! LOL

 

If anyone talks to ASUS about this...

Please pass on, that under EVENT VIEWER > CUSTOM VIEWS > ADMINISTRATIVE EVENTS

...that every time the comp wants to or does throttle, this is the message that gets recorded:

The speed of processor X in group 0 is being limited by system firmware. The processor has been in this reduced performance state for 71 seconds since the last report.

It still shows up, despite ThrottleStop overriding the throttling effect. So, it is still "trying" to throttle.

 

I also use bios 206 and tried all asus suggestions last week. No change...

Asus doesnt really hurry to solve this issue. I guess i will keep working on ac for now. Got ThrottleStop 2.00 Beta 11 on lap if needed and keep waiting for permanent solution.

 

Could some one post screen from HWinfo showing how all 4 cores works?
On mine G73 all 4 cores are running to 2,8Ghz, is it normal? Shouldnt only one go to 2,8?

 

Yes you are right..and thats not right...believe if all four are maxed, they only produce 2.3~2.4...someone can chime in with the exact specs for the cores

 

This is how normally works mine i7.

 

Here shows all 4 cores at 2,8Ghz

 

This is how mine works too...I think it is normal.

 

http://www.intel.com/technology/turboboost/
check out: View the Intel Turbo Boost Technology demo

according to this demo it should give 4x 1,6Ghz Or 1x 2,8Ghz

 

HWinfo32 may not be a perfect representation of what is going on...

 

so what tool do you use to monitor cores?

 

I use the same one...but I am sure there are probably better.

Post this question in the owner's lounge or the general G73 lounge...I would like to hear more comments.

But, just to assure you, mine does the same according to HW32 if on 13x+ multiplier.

What we may be seeing is a delayed response of the cores...in other words, 1 core at a time may be at 2.8GHz...but it may be alternating between cores making it appear as ALL 4 are at that level.

 

The other thing is that I have read that if there are no thermal problems it can go higher with multiple cores. Not sure how high or at what temps but that is what I assumed was happening. If it's running nice and cool maybe it would run a bit faster.

 

Quite possible it is delyed. Funny thing happens when you hit Prt Scr button. All cores are freezing and like its showing current real core use. I have found Multi Meter 4 cores which i think shows better current use os cores, but only in %.

 

ThrottleStop 2.00 Beta 13
http://www.fileden.com/files/2008/3/3/1794507/ThrottleStop.zip



I told you guys a couple of pages ago that HWiNFO32 does not report these new CPUs 100% accurately. The Intel documentation shows the correct method to report these new CPUs in their November 2008 Turbo White Paper but hardly any software, including CPU-Z, uses that method because it is a little complex. ThrottleStop uses the correct method so you can trust what it is saying and RealTemp also uses the correct Intel recommended method.

The latest update includes a fix so hopefully the system tray icon will stay updated when switching between AC and Battery power.

I added daniel042's great suggestion of 2 separate profiles for battery use. One for when you switch from AC to battery and now you can also set up a second low battery profile so when your battery gets down to 30% or whatever level you choose it can switch to a different profile.

The Alarm Profiles are the highest priority and will take precedence over all other profile selections.

I also added battery level monitoring to the main screen.

I don't have a laptop available to test these new features out so let me know if there are any problems. It seems to work OK.

Edit: The way a Core i7-720 mobile CPU works is that it can only use the 21X multiplier when a single core is in the active state. The other 3 cores need to be asleep in the C3/C6/C7 state. As soon as a second core is needed to process some background information then the maximum multiplier will drop down to 18X. When a third or fourth core enters the active state, the maximum multiplier drops down to 13X. These CPUs are switching the multiplier hundreds of times a second based on load and how many cores are in the active state at any instant in time. That's why old school software has trouble accurately reporting these new CPUs during some situations.

During any 1 second sampling interval, it will be pretty much impossible to see the full 21X multiplier. There will always be background processes needing to be serviced which will wake up an additional core or cores so the average multiplier will be less than the maximum of 21. The ThrottleStop log file will report the highest multiplier on the hardest working core which I find best represents how fast your CPU is really running at. Reporting a high multiplier from a core that has spent 99% of its time asleep would be misleading.

Edit: Update to Beta 13. Fixed up the profile switching code so it is more reliable.

 

Thanks for application and good explanation of how i7 works!

 

Correction: The maximum possible multiplier is 21, 18, 13, 13 for when 1, 2, 3 or 4 cores are in the active state on a Core i7-720.

For the Core i7-820, the maximum multiplier is 23, 21, 15, 15 for when 1, 2, 3 or 4 cores are in the active state.

 

What about the 920-xm?

 

The chart shows the i7-920XM has maximums of 3.20 GHz, 3.06 GHz, 2.26 GHz and 2.26 GHz. Divide those numbers by the BLCK which is 133 MHz and you get maximum multipliers of 24, 23, 17 and 17 for 1, 2, 3 or 4 cores in the active state.

The i7-920XM has a multiplier you can manually adjust in the bios so when you do that, those maximums won't apply.

 

unclewebb unfortunately now ThrottleStop is failing to switch to the designated Battery profile when i unplug my charger... probably just a minor glitch in beta 12

 

Thanks daniel042. I'll have a look at that problem. I don't have access to a laptop to do some testing until next week so I've been doing some programming in the dark lately. I'll get this fixed up as soon as I can properly test it.

 

Just unplug your desktop ;D

 

I am sticking with Beta 11 for now, as things are just fine.

I don't really care about the differences at this point. I think once uncle tests things and gets an official version complete, I will stick with 11 as things are flawless now (except for the little number 1/4 showing up in the systray correctly on boot after a AC/Battery switch).

 

Any word on what BIOS 206 changed?

 

Does anyone know? Doubtful.

 

What settings do you have on Throttlestop? I think I got hit with this yesterday when I was playing a game and the cable came unplugged. Plugged it in and it never stopped lagging up.

 

I set Chipset% to 100% on both battery and Performance.

I set Multiplier to 13x for AC and 12x for battery.

I have start minimized and minimize on close checked as well as log file, more data and BD Prochot.

I have AC profile as 1 and Battery profile as 4.

That's it.

 

This is from the big Asus G73 thread. Can someone try and see if this works?

 

hm i dont see nothing like that.

 

I found one saying this...

Processor power management
-Minimum processor state
-On battery: 5%
-Plugged in: 100%

 

I mentioned that earlier in this thread and I was told that changing it does not help.

I've been meaning to post all of my Power4Gear settings. Please click here for more information.

 

ThrottleStop 2.00 Beta 13
http://www.fileden.com/files/2008/3/3/1794507/ThrottleStop.zip

I think I got the bugs worked out of Beta 12 so the automatic profile switching should work better now. I don't have a laptop handy to do some testing so let me know if there are any issues.

 

Well the automatic profile switching does work... but it seems as though a physical reconnect of the AC power is needed to switch the Set Multiplier back up from 7 to 13. Oh well, at least we don't have to restart the laptop to make this happen... just gotta plug the charger back in.

 

Is anyone else having a problem with the cpu being slow after coming out of sleep mode?

 

Asus' last reply:

Haven't we found that this isn't a battery issue?

 

Hey unclewebb, thanks for the utility, works great on my Sony. Sometimes it doesn't restore the multiplier when AC is connected, but I just start it again. Any chance getting the voltage adjustment working?

 

It's not the battery. According to the test I posted the first stage of throttling occurred with 75 percent of battery charge remaining. Besides, it would be the first time in human history that, all by itself, software (ie: ThrottleStop) repaired a battery.

Although Lathan didn't say free battery replacement, he said repair (yeah right) or request, it's commendable that ASUS is willing to take that step, but it is really, really sad that they have absolutely no clue what's happening. Especially since you included a link to this thread, which if read, would have explained the issue quite clearly. Instead, what it unfortunately does show was he wasn't paying attention or he just doesn't care.

 

TechGadget09: Unfortunately voltage adjustment only works on the Core 2 based processors. Intel got rid of that feature in the new Core i series so it's not possible anymore.

It's not a battery issue. It's a poorly designed bios issue. There's no good reason to use chipet clock modulation, ever.

 

I have had my Asus G73 since it first started selling on Newegg and it has been doing this since I got it.
As long as I keep it fully charged and plugged in while playing games I don’t seem to have any issues.
I wish Asus would release a bios fix soon.

 

Even if ASUS did release a new BIOS, the description would probably say, "New BIOS". My experience, from several years ago, the descriptions were somewhat useful. This latest release might as well been described as "different".

I'm still hopeful that unclewebb will have an opportunity to talk to JJ about 206. Or even better, what the future holds as 206 is probably pointless.

 

Thanks Kalim for reminding me. I just sent an email to JJ at Asus and I explained the problem that G73 users are having and directed him to this thread for some background information. JJ has always been great at bringing problems to the attention of the bios guys and coming up with solutions.

It might take a while but I know they'll be looking into this issue. Severe chipset clock modulation shouldn't be used for this purpose so hopefully they can remove it as soon as possible from the bios. Locking the multiplier at 7 after the battery got down to say 20% is reasonable but using clock modulation is not.

 

Thanks for sending the e-mail unclewebb. There are several of us just dying to know what BIOS 206 addressed. If it's anything like BIOS 205 which added support for the Winbond flash chip, the change, relatively speaking, is probably negligible. Well, at least that is what ASUS claimed BIOS 205 addressed.

As to the future, lets all hope BIOS 207 fixes the clock modulation issue and says so in the description.

 

Man this really sux, I had the trigger pulled on this laptop but this might be a show stopper. Might want to pass that on to JJ...

 

I have no such issue with this BB model. Can't be all the systems or there would be a much larger outcry.

Cheers,

 

I got a reply back from Asus again. First quote is my reply.

My reply back to him.