ASUS G73JH Problem - Very slow after using battery

What are the dangers of doing the above, if any?

As I said before - this ONLY happens when on battery power and when the battery drops to 50% or less.

I would "assume" that this is the case with ALL G73s....despite some people saying they DO NOT HAVE THIS PROBLEM? Or is there a specific reason that this could be happening to MY LAPTOP?

Is there anything else I should do? Should I monitor anything else? What about if using the THROTTLESTOP as you have outlined above?

Thanks for your help - it is nice to be getting some answers.

 

This throttling scheme seems to be built into the bios and is based on and triggered by the remaining battery percent.

Now that you're starting to understand what's going on here and can accurately measure chipset clock modulation you can get other users to do some testing and see what happens to their G73 when on battery power.

The HP Envy is designed to lock the multiplier to 7.0 when you switch from AC to Battery power. A Core i7 Quad when combined with a separate GPU can draw a huge amount of energy if both the CPU and GPU are allowed to run at their full capabilities. This is more than most batteries can handle so that's why they have to use throttling.

Excessive chipset clock modulation, and 12.5% is definitely excessive, kills performance. I would set this to 100% and start with a multiplier of 7.0 and see what performance is like. This reduces power consumption when on battery power without totally killing performance and making your laptop unusable.

I wouldn't run any crazy stress tests like Furmark + Prime95 when on battery power but you might be able to play some games when set up like this. You might even be able to use your computer on battery power with a multiplier higher than 7.0. See what ThrottleStop reports when your battery is fully charged and when you first switch from AC to battery. This will be a baseline for how high the multiplier can be when on battery power without damaging the battery.

Asus might be doing this throttling so when your battery starts getting low, it will help increase your remaining battery time. They might be doing this thinking that they are doing you a favor.

I don't own a G73 so I can't tell you what your battery is capable of. If you try to draw too many watts from a laptop battery, it is similar to connecting the positive and negative together directly. You can damage your battery.

That's why ThrottleStop lets you set up a power profile that you can automatically switch to when you switch to battery power. This should be able to let you create your own throttling scheme that isn't nearly as severe as what Asus is doing to you. Hopefully it is fairly straight forward to set ThrottleStop up to do this but if you have any problems just ask.

 

Ok, I get a lot of what you have said, etc...

But, a lot of it is also over my head, but I am a pretty quick learner.

Maybe you could refer me to some step by step stuff I should/can do to help the situation...don't get me wrong, what you have posted seems pretty straight forward, etc... It is just I have never really done "testing" per se with programs to see how far I can push CPUs / GPUs, etc... Not really sure about these technical things and how they work, etc...


What I really don't understand is why more people with the G73 haven't complained about this - it is SOO NOTICEABLE! - maybe it is because most G73 users don't run on battery that often?

Anyways, thanks again for your help...it is a huge step in the right direction...just don't get how the makers think it is helping to throttle the CPU when the machine essentially becomes for the most part unusable.

 

And, just to follow up, the machine does NOT return to normal when plugging in the AC once this starts.

 

uncle,

Can you also forward this to the ASUS rep?


Should I be contacting ASUS and referring them to this thread?

Thanks!

 

So tell me people, since Asus doesn't like you upgrading your cpu and everything what happens if you buy it from best buy or buy it from somewhere else, and you get the CPU upgraded from best buy will asus whine and complain about these things? Why should asus whine and complain if you want to upgrade the cpu anyway, its an enthusiasts machine and what the heck is their problem? Also can you put in Intel 6300 and not void your warranty or is Asus also going to whine and complain about it!

 

Is there a cause to the THROTTLING, or is it automatic?

Is there anything I can do to minimize it's effects? (ie screen brightness, settings, etc...)

 

Happens to me as well, all I do is turn off the computer, plug in the power cord, then boot it back up.

 

That's the situation here. The only time I run on battery is when I'm moving the G73 around the house. My G73 is never on battery long enough it would seem to encounter the issue.

Well, for one ASUS can claim the G73 runs longer on battery then it would otherwise and perhaps the Li-ion battery life is extended while recharge time is reduced.

 

Well why don't you try the option he mentioned and see if it fixes it?

 

Here's where you can download Prime95 for testing purposes.
Go down the page to Step 3.

http://www.mersenne.org/freesoft/

And here's where you can download Furmark.

http://www.ozone3d.net/benchmarks/fur/

corpfan1: Your initial testing seems to show that this throttling is directly related to the battery level. As the battery level drops, the clock modulation percent drops lower and lower which is killing performance when on battery power.

You could fully charge your battery and then run something like Prime95 on battery power. For an i7-720QM you would need to run 8 threads and I prefer the Small FFTs test for its load consistency. If it asks you to join GIMPS just say No, just stress testing. As soon as you run Prime95 it should let you select the Small FFTs option. If not then go into the Options menu and select Torture Test... and it will bring up the option of what test you want to run and how many threads of it. Run a ThrottleStop log when testing and keep track of when you start the test and what your battery level is. Whenever chipset clock modulation changes, write down what the battery level was at that time. Open up the Windows clock to get the exact time so it is comparable to the log file.

If this throttling scheme is totally battery level dependent then there isn't a hell of a lot you can do other than keep your laptop plugged in I guess.

I'm interested to see the results from the Prime95 test with a fully charged battery. It will give us an idea of what your battery is capable of. If it starts to immediately throttle and slow down then that's a sign that Asus doesn't have any confidence in the battery they are using to fully power your laptop. I think it might run fine at first and not slow down until the battery starts to discharge.

Before contacting Asus for a solution we need to be able to clearly document this problem so any user can do the same sort of testing and get the same results. We have to prove this is a design defect and not just your laptop and the way you have it setup. With some more data then we can both start complaining to Asus to see if they can come up with a fix.

Any user that has this problem, before you reboot to restore your computer's performance, run ThrottleStop and check the Chipset Clock Mod function and set it to 100% and see if that fixes things. If your computer is idle when you do this then it's not going to hurt anything. Avoid doing this when stress testing on battery power until we find out more.

 

Ask in another thread. This thread is about solving the issue with our laptops slowing down on low battery. Why do people keep trying to hijack this serious thread?

 

Throttling as power diminishes is probably by design, as well as the throttling for thermal reasons.

What I would think is not by design, is the CPU state not returning to normal when you go back to AC power. This is something that could quite possibly be fixable through some combination of firmware and/or drivers but Asus needs to know about it.

If I ever get mine back I'll certainly test this

 

Honestly, I cannot see how they would do this by design. I mean if it takes say 10x as long to do something because it is so slow (25mhz bus clock/180mhz cpu in my case) and you get 1% more battery life, then what is the point? For anyone who has not seen this yet, it is so slow you can barely do anything, even moving windows around on the screen is a problem. If I had not read this thread I would have called in an RMA thinking it was completely broken.

Luckily I don't run on battery much so it won't bother me a whole lot, but rebooting to fix it is a pain.

 

I'm guessing since corpfan1 made a post on another thread referring to this thread right after the question freedom16 posted there, freedom16 assumed corpfan1 was referring him to this thread, since freedom16 asked what this had to do with his question there.

Freedom16: Corpfan1 didn't quote you in his post on the other thread and it's unlikely he was replying to you just because his post happened to come right after yours

Hope that all made sense LOL

 

I didn't say it was a good design

My point was that the power management software may be buggy - stuff like that happens on first generation machines.

I haven't seen enough evidence that there is or isn't a problem with your machine or if this is a problem on all G73s. We need every owner to test it...I will certainly test mine when I get it back.

 

Yes - would love to see more people testing it - that's for sure.

Not a huge deal for me anyways, as I use 95% AC power as well...

But, when this started, I was showing my laptop to a friend and didn't plug it in...I told him about the "awesome" specs of the laptop and its blazing speed, etc...

Then, booted it up and let him play around a bit, and he said - "This laptop sucks - you call this blazing?" Of course this was the first time I noticed the issue!

Anyways, I will test some stuff and see what happens as per Uncle's instructions. If there is anything else I should do/try, please post!

 

DrBoost,

Could you change the thread title to:

ASUS G73 Problem - Slow on battery - CPU Throttling?

Thanks!

 

This wouldn't be the first scheme dreamed up by engineers that might be good in theory on a piece of paper but really sucks when it kills performance and makes your laptop completely unusable.

I'm pretty sure that when Asus gets some complaints, they will adjust this. If not, you should be able to use ThrottleStop to regain some control of the performance of your laptop when on battery power. Hopefully we can get some test results from a few different users so we can see exactly what's going on.

ThrottleStop has a new feature that lets you automatically switch power profiles when changing between AC and battery power. This feature could be very handy to control this issue while waiting for Asus to come up with a fix.

 

Uncle,

Why is it that when Throttle Stop is monitoring only...

1) The CPU sticks at 12x and does no turbo boosting?

2) HWiNFO32 reports GPU temps, but shows no current temps when running by itself?


Thanks!

 

I didn't see while re-reading the thread - if you just leave it on a single power profile, with NO power4gear installed ... does it still throttle?

 

Doesn't seem that power profile has anything to do with it...just whether it is on battery or not.

 

Make sure you set the Multi to 13x, It will tell the CPU to run at any speed it can. But since this is only to fix problems when running on battery, and hopefully you're not pushing hard when running on batter, it may not matter.

 

So, does it switch power profiles when it goes to battery? I think windows does this automatically, and maybe it's not switching back correctly.

 

Have you tried uninstalling Power4Gear and then going into bios, setting it to defaults?

I only ask because my G73 does not do what yours is doing. only throttles at 5% as expected it would.

 

Yes, did uninstall power 4 gear, but did not touch bios. Running everything stock. Reinstalled power 4 gear when I realized it didn't make a difference.

 

PRIME 95 REPORT

1. Started Prime95 at 100% battery power at 2:53pm.

2. Temps for CPU spiked to 63 degrees from about 45degrees.

3. Battery draining quickly. All clocks still at 100% at 79% battery. CPU Temps at high 60s.

4. 3:08pm (15mins in...). Chip% dropped to 50% at 70% battery level. Multiplier to 7x. CPU Temps immediately dropped to mid 50s. CO% at 54% approx. Clocks on HWiNFO32 dropped to 498MHz. CPU Status to 70MHz x 7x. Print Screen attached.

5. 3:19pm - as expected at 50% battery power, everything slows to a crawl. Chip% drops to 12.5%. CO% now at 16-17%. HWiNFO32 shows 24.3MHz x 7 = 170MHz. Keyboard typing is lagging. Print Screen attached.

6. Things are just to slow now to continue. Stopped all processes and going to reboot. This should be enough data.

 

Thanks for finding that bug. I added a new feature recently to restore the multiplier to its maximum when the program exited. That was working correctly for Core 2 owners but the register used to control this is a little different on the newer Core i CPUs so this feature was broken. This new version should be OK now.

ThrottleStop 2.00 Beta 8
http://www.sendspace.com/file/4d4g5n

Question 2 is a bigger mystery. I'm not sure what HWiNFO32 is up to.

With the latest version of ThrottleStop, I've added GPU temperature monitoring but I'm not sure if the ATI mobile driver is supported. It works OK on my ATI 5770 desktop card.

You will need to go into the ThrottleStop.ini configuration file and edit it and set:

GPU=2

if you have an ATI GPU.

For Nvidia use GPU=1 and if this feature doesn't work or you don't want it then use GPU=0. Let me know if this works for you.

corpfan1: Looks like you have solved part of the mystery with why this is happening. It seems to be all battery level related.

The next test would be to run ThrottleStop and force chipset clock modulation to 100% to see if this will work to fix this problem. The other thing you could try is to disable the bidirectional PROCHOT bit. Unlock this in the Options window and then disable it in the main screen. This alone might prevent this chipset clock modulation but probably not.

 

Don't you mean...

Profile=2

?

 

No, it does RESTORE when the program is exited.

It just doesn't allow Turbo, while the program is monitoring. The multiplier stays at 12x while the program is running/monitoring.

 

No. If you have an ATI card, which you do use this:

GPU=2

Open up the configuration file I mentioned and this will make more sense.

 

Did you reboot and is that with the latest version?

Did you run a load to check the multiplier? I'll check back in a few hours to see what's up.

 

There is no "GPU=" in the .ini file. There is a Profile=0.

Would I add a GPU= line?

 

What I mean, is if I run HWiNFO32 after a fresh boot (on battery) the CPU multiplier starts out at 12-21x.

As soon as Throttle Stop is launched, the multiplier goes to 12x immediately and stays there. It no longer moves around between 12 and 21x.

The CPU total MHz goes to stable amounts at about 1600MHz...where as before ThrottleStop is launched it fluctuates between 1600MHz and 2800MHz constantly both under load and when moving the mouse during idling.

 

Did you read my PRIME REPORT?

http://forum.notebookreview.com/showpost.php?p=6105937&postcount=127

 

I wonder if you have a bad battery? Your battery times are horrible. 50% about 30 minutes later?

I can watch a movie for close to 2 hours on this machine.

 

Did you download the latest version of ThrottleStop that I uploaded and posted a link to on the previous page?

Here's how the INI file looks when you first download it.



Now can you see the GPU= line?

I just got some feedback and this new feature seems to work OK on ATI mobile GPUs.

Your Prime95 report proves what is going on. I will send that to JJ at Asus. I hope other users can do some similar testing. I'll be back later. Download Beta 8 and do some more testing.

 

I'm pretty sure he was running a CPU intensive program meant to drain the battery as fast as possible. Watching a movie is barely any work for the laptop, just the drive spinning uses much of any amount of power.

 

Interesting idea...

But, I think that is because Prime95 was running.

I get about 2 hours as well if just watching a movie.

 

Yep - all is good with Beta8.

Using it now. Both glitches fixed.

Thanks!

So, when you say it PROVES what is going on...is it anything abnormal - or is it simple ASUS safeguards?

Wonder if BIOS 206 addresses it?

Too bad ASUS won't list what BIOS 206 does.

 

I'm running a test as I type this, I'll report back with my findings.

If you have an extra moment and don't mind, please ask JJ at Asus what BIOS firmware 206 is suppose to address, the description is useless. Maybe JJ or someone at ASUS could add a meaningful description.

Thanks.

 

My opinion is that it is a very poorly thought out way to reduce power consumption.

Some Dell laptops are throttling down to about 10% of their rated performance level and I thought that was bad but this is even worse. With chipset clock modulation dropping down to 12.5% and the multi stuck at 7.0, you're lucky to be getting 5% of the rated performance. If another user or two confirms this mess then it is definitely time to start complaining. That's a safety feature that has gone horribly wrong.

It seems all of the bios guys in the world speak some sort of special language that is impossible to interpret. The Dell bios guys also have a difficult time releasing any meaningful information about what they've actually done when they release a new bios version. I'll ask JJ but I don't expect to ever get a meaningful response.

 

Noticed something interesting while playing around with Throttle Stop...

It does override the CPU and multiplier, until the battery % gets under 30%...at that point, even engaging throttlestop and setting multiplier at 13x, WILL NOT override the CPU - multiplier stays at 7x - as indicated on Throttle Stop, CPU-Z and HWiNFO32.

Just thought that was interesting.

 

Have you tried disabling the BD PROCHOT bit that I talked about before? That might help.

Can ThrottleStop override the chipset clock modulation when the battery gets below 30%? That's the biggest issue at the moment. That's really killing performance. A laptop is usable with the 7X multiplier as long as chipset clock modulation isn't being used.

 

For some reason, when ThrottleStop is set at 13x multiplier, the mult drops to 7x under 30%, but the chip% didn't drop under 100%...but this was one isolated test.

I will do more later.

I tried disabling PROCH, but the issue still proceeded as usual.

 

unclewebb,

Maybe you can post some instructions of things to try out so I can do so later when I have some time...

Like how to properly use/disable PROCH...

Maybe what to try in terms of multipliers/chip%/clock% etc...

...and when to use them (e.g. battery power at x%, with AC, plug in AC at x%, plug out at x%, reboot, at such and such a point, etc..)

Do you think the BIOS upgrade (206) might address this?

Is there anything else worth trying with power settings / profiles, etc...?


I am assuming for now, that this is an inherent G73 issue and that the best thing to do is live with it and use software (e.g. ThrottleStop) to prevent the worst situations?

Anything else anyone can think of here?

 

Here's the test I did, values are approximate. I used both FAH SMP and GPU clients for testing to create a heavier load on the battery. Results follow:

** Test Started **
12:50 pm - Unplugged AC power
12:50 pm - 1.7GHz (ratio 12x), battery 100% (47 minutes remaining) and CPU-Z reports 13x multiplier.
01:00 pm - 1.7GHz (ratio 12x), battery 78% (35 minutes remaining).
01:02 pm - 500MHz (ratio 7x), battery 75% (30 minutes remaining) and CPU-Z reports 4x multiplier.
01:05 pm - 500MHz (ratio 7x), battery 70% (47 minutes remaining).
01:10 pm - 500MHz (ratio 7x), battery 65% (44 minutes remaining).
01:15 pm - 500MHz (ratio 7x), battery 57% (33 minutes remaining).
01:20 pm - 500MHz (ratio 7x), battery 49% (32 minutes remaining).
01:25 pm - 500MHz (ratio 7x), battery 41% (27 minutes remaining).
01:26 pm - 170MHz (ratio 7x), battery 40% (29 minutes remaining) and CPU-Z reports 1x multiplier.
01:30 pm - 170MHz (ratio 7x), battery 34% (25 minutes remaining).
01:35 pm - 170MHz (ratio 7x), battery 28% (21 minutes remaining).
01:40 pm - 170MHz (ratio 7x), battery 21% (15 minutes remaining).
01:45 pm - 170MHz (ratio 7x), battery 14% (10 minutes remaining).
01:47 pm - Low battery warning at 11%
01:50 pm - 170MHz (ratio 7x), battery 8% (5 minutes remaining).
01:51 pm - Critical battery at 5% would send G73JH into Hibernation, instead load test ended.
01:51 pm - Plugged AC power
01:51 pm - CPU speed did not return to normal on AC. Recharging battery.
02:00 pm - Charge 7%
02:15 pm - Charge 13%
02:30 pm - Charge 27%
02:45 pm - Charge 39%
03:00 pm - Charge 51% (charge level now above 2nd ratio change - ratio remains 7x)
03:15 pm - Charge 64%
03:30 pm - Charge 77% (charge level now above 1st ratio change - ratio remains 7x)
03:45 pm - Charge 89%
04:00 pm - Charge 95%
04:15 pm - Charge 98%
04:30 pm - Charge 99%
04:45 pm - Charge 99%
04:47 pm - Charge 100% (ratio remains 7x)
** Test Ended **

I couldn't wait to reboot the laptop and return everything back to normal. Thankfully I don't have to do that again.

 

Wow Kalim thanks thats good stuff-that SUCKS! I use mine as a desktop replacement and havent even taken the battery out of the wrapper...Why would they do that?...its like they give the owner/user NO control over their own system....it gimps down no matter what your settings are and the battery dictates the performance you will get, not your power/system settings..and then the kicker is, then it DOESNT EVEN RETURN TO ITS NORMAL STATE AFTER RECHARGE!!!thats just dissappointing..I plan on keeping mine on AC for 95% of its life after seeing this ; )

btw-I have the orginal .204 bios...I know I dont need to update to 205 or 206, but Im curious to what 205 contained (theres no description on ASUS site of 205) and Kalim I hope you post up about 206 when you get the word. Thanks!

 

yep, agreed Kalim, thanks for enduring that test! Great info there, and honestly, I struggle with a laptop that can't be used on battery when needed... and this one may not work.

+1 Rep

 

ehh still no solution. and i nned to work on this machine...