[Fixed/Workaround] Asus G51J(x) CPU throttling investigation

Thanks for the thorough review Peter. I did a lot of testing of ThrottleStop 2.00 on a few different Core 2 CPUs, mobile and desktop, so it's good to hear that it is also working correctly on your Core i7.

It was a big overhaul adding profiles to ThrottleStop but now that is done, adding Nvidia GPU monitoring to TS will be relatively easy. Hopefully before the weekend.

 

That's great - good job man. The only things I would have suggested is having documentation - but then I looked and you do have it - maybe you could add a Help or 'See Documentation for more details' with a link to the html file in the directory in case people miss it. In my mind I thought the DTS Alarm worked differently and couldn't figure out why no matter how high a number I put in, it always defaulted to 80 .

Having that GPU and GPU+CPU Cores combined throttle stops as a secondary alarm will really put the icing on the cake .

Thanks a lot,
Peter

 

ThrottleStop 2.00 Beta 4
http://www.fileden.com/files/2008/3/3/1794507/ThrottleStop.zip



I added GPU temperature monitoring to ThrottleStop for users with Nvidia GPUs and the ability to switch power profiles automatically based on your GPU temperature. I think this feature will be useful for Asus owners that want some more control over the bidirectional PROCHOT bit based on GPU temperature. It should be easy to set up a second profile to switch to when you reach the maximum GPU temperature level that you feel comfortable with.

Peter: Beta 3 that I sent you last night had a bug or two when switching profiles and returning to the original profile. After installing my Nvidia card today, I think I was able to get everything working 100%. Let me know if you see any issues.

As I told Peter before, a GPU alarm will take precedence over a CPU alarm. Intel does a good job with CPU thermal management and the problem with the Asus G51J is now GPU temperature related. If a CPU alarm is triggered first, it will automatically switch to the GPU alarm profile if that alarm goes off after the CPU alarm.

If the GPU alarm goes off first then a CPU alarm will be ignored until after the GPU temperature has returned to normal. Once this happens if the CPU is still too hot, then the CPU alarm will be triggered.

You can manually switch profiles during an alarm but when the alarm is finished, it will probably switch back to the original profile. I'm still thinking about a few details like this and what should happen in this situation so give me some feedback with your suggestions.

If you unplug and plug in your laptop 101 times while an alarm is going then I have no idea what profile you'll end up at or which profile you will return to when everything is back to normal.

Happy testing.

 

Excuse my techno-foolishness.

From the wiki:

using these values we can make a guess that asus has set the max draw at 90-100W before the CPU is cut. so far we have determined the cut is not initiated by PowerManagement or ACPI; disabling them has no effect. because it can draw more than the trip point, it is not a hardware limit either. see the BIOS disassembly page for technical details.

Therefore, if I was to reduce power usage by the laptop in other ways, would I be able to prevent needless throttling WITHOUT the risk of the computer overheating?

My thought is to remove the second HDD, which uses about 10W while idle.

Obviously not ideal for everyone, but I am very worried about cooking this laptop because I had so many problems with my last one (Macbook Pro don't even get me started with Apple), so I'd like to leave the thermal throttling mechanism in place.

I've also disconnected the LED lights in my laptop's screen, but I would guess that's only saving me a watt.

The reason I think this would work is because lowering the screen brightness seems to do the trick as well?

 

There's always some risk - it's not the computer overheating and shutting down - it's the PSU. 10W idle? Wow that's a very power hungry HDD - how much does it take on full blown RW?

See my reply below for a plan on reducing the risk . Note though that the 90-100W limit was an estimate from many tests - I got around 100W with efficiency factored in. Just having the throttling in though by removing the HDD - I never tested that since I only have 1 HDD slot. We know burning a CD, having max brightness and GPU overclocks and having USB devices hooked up can make it trigger at lower overall power use than expected.

Thanks Uncle Webb for the info on that, I did some testing with that version from last night, but not too much.

The only thing about it is that if the GPU was being pushed to the limits while only 1 core of the CPU was being utilized (and possibly not greatly) then this system would throttle the CPU.

Meanwhile if the CPU is being utilized heavily while the GPU is idle, then this would also throttle the CPU.

In neither of those cases would we want this throttling, since it is only when both the CPU and GPU are being utilized heavily - and in the one case it could result in the CPU being throttled (possibly heavily) while the GPU was being pushed - which may not help lower the GPU temp at all and therefore it wouldn't end until the GPU was allowed to cool down.

It's true that this system as it is helps to reduce CPU temps - but it wouldn't help with overall system power usage.

For me, for example, I don't mind if my GPU is at 101*C (well don't mind in that I wouldn't want the CPU throttled because of it, I'd rather nVidia underclocked the GPU for me at that point) - as long as my CPU was not over 75*C on multiple cores (approximate values).

That's why an overall combo option taking the total heat output of all the CPU Cores and GPUs and adding them up would give a picture of those most important of power drains and give an approximate overall system power usage in one simple option.

If we know that we have more USB devices hooked up, or full screen brightness, wifi, bluetooth, etc... and we experience problems or suspect we will (very hot PSU/Power Meter reading above 145W consistently) then we can reduce that combo trigger number.

With that - we'd be all set imho

Thanks a lot man for all your help on this,
Peter

 

I hope you have some time this weekend to do some high temperature testing because I think once setup, this updated version of ThrottleStop is going to be able to handle a lot more situations than you give it credit for.

The ability to throttle the CPU or GPU are independent. In your example if you were pushing the GPU and it was hitting 100C and that was as far as you wanted it to go then you could get it to switch profiles so that the bidirectional PROCHOT bit was enabled which would throttle the GPU. If the CPU wasn't too hot, it could still run at full speed. If the bios was trying to throttle the CPU at the same time, ThrottleStop should be working against that to nullify CPU throttling.

If the CPU is working hard and getting hot and you don't like that then have it switch to a profile to slow it down a little to keep the CPU heat under control. If you want Intel to be in charge of CPU throttling then disable the CPU alarm and just use a GPU alarm.

With this program, you will still have the built in CPU and GPU thermal throttling thresholds once you go beyond your preset ThrottleStop alarm temperatures. This is an extra layer of protection and you can use the TS profiles to manage the CPU heat output and the GPU heat output by toggling the bidirectional PROCHOT bit.

Play around with it and show me some real world examples where this present version of ThrottleStop is really deficient and I'll happily look into adding more features and control to it.

The amount of throttling is all up to how you set up ThrottleStop. I think most users are going to find it gives them a lot of control over their Asus G51J's maximum CPU and GPU temperatures.

 

I'm not 100% on the 10w idle measurement, but I believe hard drive's actually often use up to 14w idle, and these are 7200 rpm drives.

My battery life has increased substantially (like 20 - 25%) since removing the second drive. I hope it will also sort of help with CPU temps (by like a degree or two, nothing major ).

However, if throttling is caused prematurely by excessive draw on the PSU, then would removing this drive have helped?

 

http://www.tomshardware.com/charts/2.5-hard-drive-charts/Idle-Power-Consumption,681.html

I'm not sure what model you have but 7200 rpm 2.5" hard drives typically consume about 1 or 2 watts at idle. The less power consumption the better but most users are not going to find a 25% increase in battery life by decreasing power consumption by 1 or 2 watts.

 

The PSU does not cause the throttling - it just overheats and shuts down if you draw too much. From my tests it seemed like Turbo Boost (before disabling the bit) would only reach its full potential if there was 20W power left in the limit before it was triggered - if I went just above that and not quite enough power was available, then Turbo Boost wouldn't activate at all. Those were tests before disabling the bit though - so this is different.

I too think the HDD draw is way out of proportion - 3.5" drives can be power hungry, and maybe you got your data from those - now you know . Every little bit helps, but it wasn't quite as much as you thought, still an improvement though, which is good - though having one less drive is a big trade-off (depending how you use the system I suppose).

Cool, thanks for the info Uncle Webb - that's the key point for me - my understanding stems from the idea that PROCHOT only affects the Processor (it standing for 'Processor Hot'). If in this case the GPU is also being throttled then I was wrong in my understanding and that would accomplish what I want. In the FurMark tests it seemed like the GPU temps would just keep climbing regardless of whether the CPU was throttled or not (though the slightly more slowly since there is not as much pressure on the heatsink).

If the PROCHOT does underclock/volt the GPU as you say, then this app would do what I want. .

Peter

P.S.

I understand and have great respect for that . I'm not saying to remove any control, but with my understanding one more box (without taking any other boxes away) would solve things on the PSU side (the case where if both the CPU and GPU together get too hot ( possibly throttling with the GPU and CPU each cooler than I want any throttling if they are full blast on their own) then I want throttling, otherwise keep going as normal).

 

Update: In all I've read the BD Prochot only affects the CPU. It throttles the CPU if the system power usage gets too high - which could happen as the result of other factors than the GPU (though the GPU is a very large power consumer it's true).

I see how through our different methods we can arrive at a similar (but not identical) result:
The case I was thinking of was if the GPU gets too hot, then I would want the CPU to be throttled and prevented from getting too hot and therefore taking too much power it's true.

However, if the CPU is still operating at a low enough temperature then I don't mind it running at full throttle until it reaches a high temperature because that full throttle takes it there. Until the CPU actually uses that power though I want it to be going full speed. So even if the GPU is hot, I would want the CPU to be running full speed until it reaches a temperature running at full speed that takes it past the safety that I have set for the combined temperatures.

In this system, any time I use this setup, when the GPU reaches a certain temperature, the CPU will be throttled in some way (while the GPU is still free to keep getting hotter). You don't have any mechanism to cool down the GPU it's true - so this is how things would operate anyway. It's just that in this system, I can do what I need to do, but in a more complicated fashion.

For example, GPU at 100*C and CPU still running cool - if the CPU speeds up in short enough spikes in small enough bursts (or is only using one core) it will not get hot enough that it triggers the alarm because from the wall it is not a drain. So it is still allowed to operate without any throttling at all.

On it's own I would not want any throttling of the CPU based on GPU temperatures since this app doesn't do anything to cool down the GPU (since the BD ProcHot itself has no effect on the GPU). So I only want any throttling at all if the GPU and CPU temperatures are both too high (since that means high overall system power usage). I would want the the throttling to only kick in if both were that high.

So if the GPU hits 100*C then I want no throttling on the CPU unless the combined cores reach 300 (75*4) - then I want the CPU to be throttled down to say 9x multi to be safe.

But if the CPU is all alone getting too hot, then I want it to throttle when it reaches 80*C for example, to prolong the life of the system. I suppose I could sacrifice this setting and instead just use this alone for the GPU throttling. But in that case I would only want this to come into account if the GPU is hot...

Does ThrottleStop take into account all the cores btw and average them for that DTS Alarm value? Or is it just the highest (or in this case lowest) value from all cores that triggers the alarm?

Bleh, the more I try to explain it, the more I'm probably just repeating myself. At least though this corrects the misconception of my last post, so I'll post it now.

Thanks again ,
Peter

 

I started to get tied up in logical loops trying to work out a way to accomplish everything in the system as is in my post above, it may or may not make sense. Here's how the system I was proposing would work though:

So if the GPU is running hot and the CPU is using all cores heavily (or one core heavily and all the others moderately, or one core very heavily and the other cores little) then I want it to throttle. That's why the combined GPU+All CPU Core alarm box would allow me to fine tune this throttling by only coming into effect.

It would also take effect if the GPU were moderately hot while the CPU was very hot.

It would give fine control that you can't get by looking at one or the other value alone. It allows us to take all of them into account in that large number, which we could use to affect the throttling if we want to, or just use the other options themselves. It gives us a lot of flexibility.

This alarm would combine the Absolute Temps of all the CPU cores with the GPU temp. So it would just be a calculated value of: TJMax-Core 1 DTS + TJMax-Core 2 DTS, ... + GPU Temp = Combined Limit. If that Combined limit is reached, trip to specified Power State. That way we can let the system use the CPU and GPU as much as we want with absolutely no throttling - unless both are at their peaks and therefore beyond the combined limit - and then we get the throttle we want to prevent the PSU from shutting down.

The Combined Limit could also be defined by averaging all the cores and thereby avoiding any issues with hyperthreaded cores (which have their own same temp display in Throttle Stop). That would involve more calculations on each cycle though (more than adding/subtracting and one comparison) but would also reduce the ability to finely nature of the limit. Alternately you could also only add in true cores rather than hyperthreaded ones - or you could just inform users that it would add all those CPU temp #s + the GPU temp together as displayed on the first page - which might make it easier for them to understand anyway. Of course the calculation could use mathematical simplification to take into account the # of cores and so on, but the result of the combined value would be the same.

Does that make sense? It might not take much more than copy and paste programming from the code you already have to get the GPU and CPU Temps/Alarms and change the power state (depending on how it's programmed of course) - but would add a valuable feature not covered by one or the other system above individually. What do you think?

Thanks a lot,
Peter

P.S. Let me say again, thank you so much for your assistance on all this and adding these features and everything else that you've done - it's greatly appreciated - Thank you

 

Thanks Peter for your detailed explanations. I don't own an Asus G51J so it looks like I've made an assumption or two that are not correct. I'll have to read your posts carefully and go back through this thread to look for test results. ThrottleStop is heading in the right direction to give you guys more control of your laptops and that's the important thing. I'll definitely look into improving things further as time permits.

One thing to keep in mind is that these 4 cores are very close together and that during normal operation there is no significant difference in core temperature between them. A lot of the difference frequently seen is sensor error. The load tends to get continuously balanced between the cores so any gradient between cores doesn't tend to last very long as the heat instantly transfers from core to core. A good test is to run two threads of Prime95 Small FFTs and lock that to one core. Run a RealTemp log file so you can see what the worst case difference is and then end that test and go do some gaming and run another log file. I find that the core temps tend to stay fairly close together but I haven't tested a Core i7 mobile CPU so I'd be interested in seeing a test like that.

The Intel turbo boost feature also helps to balance out power consumption whether 1 or 4 cores are active.

ThrottleStop does compare the DTS values of all 4 cores and triggers the alarm based on the lowest DTS (highest core temperature).

 

Ah! I'm glad I mentioned it! I think the battery life measurement may be due to the fact that I put my swap file on the 2nd drive, meaning it's not often idle.

I ended up putting the drive back in now that I know its not going to do much with the throttling. It's much easier to keep all my important stuff on this second drive and back it up directly, and I was cutting it close space wise.

Thanks for the help guys! I might look into this solution myself, but only after you pros give it a good work-through. I'm too scared to break something I can't afford to replace... and I usually just game at 1280*720 right now anyways.

My true worry about this throttling issue would be for music production... this laptops other purpose. I splurged for the i7 specifically for the low latencies, and I'm not too happy to hear that it might not really work as advertised.

Would the CPU throttle at 100% CPU load, but with a low GPU load (powermizer puts the clocks at 100, 200, 450 or something).

 

Shouldn't have to worry at all - it's only if the GPU is being fully utilized, with full screen brightness, and maybe some USB devices hooked up. With the G51J it's a bit more of an issue since that GPU takes more power and you have the second HDD - but as long as you do not utilize the GPU much you should be fine .

I thought it would be nice to have the two HDDs at first, but then when I found out it was not RAID capable, I realized I would just end up using the second HDD as my backup - cause I'm lazy - and that way if I ever lose my laptop, it gets stolen or damaged or something I could lose everything. So I'll stick with my 3.5" TB external HDD instead .

Hey Uncle Webb,

Thanks for that understanding . The application is very close - as is, it is very useful, but if you can add that feature I mentioned above it would make the jump to essential for me .

Thinking on it more and comparing the TDP on the GPU/CPUs, they are quite close (45W for the i7-720QM and 50/55W for the GTS360/GTX260M GPUs) so having all the different cores of the i7 listed would dillute that value too much. So definitely an average would be the most accurate for the overall CPU power use - but the #s are close enough (max difference I've seen while idling has been about 7*C between cores, or about 14% of total temp) on HWMonitor - but I can believe there may be some sensor issues with that app. I'm curious now myself and will probably do the prime95/RealTemp tests sometime soon.

By the way, do you have a donation link somewhere? I don't have that much to spare, but I can at least buy you a coffee or two for all the effort you've put into this - thank you .

Peter

 

By fix do you mean fixed by asus or the fix in the guide that we have to do?

 

All you have to do is open a program and uncheck a checkbox.

 

Ah and this will solve the throttling completely? What program is it? Or is it in the guide? thnx too

 

Ah nevermind i read it through.

 

Anyway, is the 120W power limit still present? There seems to be an absence of discussion on that issue.

 

Because we already answered it . With the bit fix, the laptop draws up to 164W from the wall in my tests (142W continuous). This is beyond the 120W even when you add in the efficiency of the adapter. thalanix was drawing so much his PSU would fail after 4 minutes or less of FurMark. So yes, the bit will remove the limit. So now we should ask ASUS for stronger adapters if we want to get the most out of laptops without killing the PSU....

I wonder - is it only the laptop itself that is covered by the 3 year warranty, or is it just 1 year on the accessories (such as PSU).

Peter

 

Peter, the Core i7 desktop sensors are excellent and for them I believe there is a deliberate 5C error between the first core and the last core. The data coming from the newer 32nm Core i chips have some significant error at idle. I haven't seen enough data from the Core i7 Quad mobile chips to conclude anything but I do know that the accuracy of the sensors that Intel uses on their CPUs is not typically great when the core temperature is a long ways away from the throttling point. There are also a lot of other unknowns about these sensors and Intel's calibration procedures so keep that in mind when testing.

I also don't have a lot of faith in the Intel TDP ratings. You can have two CPUs with vastly different MHz ratings yet the exact same TDP rating. At times it does not accurately reflect actual power consumption and a GPU manufacturers TDP rating isn't really comparable to Intel's ratings. Core voltage also varies from one chip to the next.

Time for a holiday for a couple of days. I'm looking forward to seeing some data and test results when I get back.

 

Accessories are covered by 1 year warranty, I read somewhere. Where can I get like 170W Adaptors?

 

I hear you on the TDPs - especially if you compare with models from different manufacturers. However, the TDP can be vastly different if a different within a model series depending on the nanometer manufacturing process being used, along with other factors. Mhz is not an accurate estimate of power usage...

However, I am simplifying things again - we can use TDP and also all the test data points collected for the CPU Throttling experiments. Thalanix posted for GPU and CPU power usage correlated with my test data on power meter levels. It doesn't mean the CPU and GPU use exactly their TDP - but it means they are on the same Macro Scale in comparison to the rest of the system. Plus I was only using it to refer to preferring using the average/max Core Temp in the measurements.

I hope that makes sense - have a good Easter weekend.

Thanks again,
Peter

P.S. I hope you saw where I agreed with you there is no point taking anything other than that average or maximum CPU temp reading anyway - so the differences between cores is moot...

 

I've been away all week since that last post, I was watching my inbox to see if there was any reply here since but looks like not.

Either way, thanks very much unclewebb for your help on all this - I hope what I was asking about makes sense, if not I'm sure it's because of my inadequacies in not explaining it clearly.

Thanks,
Peter

 

I've been helping out the G73 users this week that have severe throttling issues while on battery power. They used to have issues but ThrottleStop is helping them out too now.

I added a few more features to deal with their issues and generally polished up ThrottleStop a little.



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

Nothing too exciting. I understand the feature you would like to see added to ThrottleStop but it hasn't made it to the top of the priority list yet. I'm still thinking about it.

 

It looks good, thank you. Let me know if there's anything you'd like me to test out or confirm relating to that feature or any other.

Thanks,
Peter

 

really nice thanks and if it is possible try to add fan control into it
voltage id not working on my i3 330M :-/

 

If i do the Fix from Linux(Ubuntu 9.1), there will fix all the computer? it mean i can go to Windows 7x64, and had fixed too?

 

Voltage VID is no longer supported on Intel's newer Core i chips. You'll have to complain to Intel about that.

Artharas: You will have to use some other software to do this fix in Linux. ThrottleStop is for Windows only.

 

I am trying to read through the pages and it's painful. Okay, I have an Asus G51J-A1. So from what I collect:

1. This fix eats up the 120W(?) PSU.
2. The CPU/GPU might get too hot?
3. Any compatible 150W PSUs and where to buy them?

 

So i just got a new G51JX notebooke and the gaming performance is awfull so i ran some 3Dmark06 tests. The score is an below average of 8k marks. The GPU scores are as bad as 3000 where they should be 4000+

After alot of testing and reading i found out that when dimming the screen the GPU scores go up. The CPU score always stay the same. I have to totally dim my screen and turn of the keyboard lights to get a decent but still bit low 9.5K score and some decent FPS in games.

But we all know, u can't play on a totally dimmed screen.

After reading this post i understand that when u disable throttle the CPU always stays clocked high and the GPU will down clock?

Now when i got the system i knew it was bought by someone else who returned it, u could see that from how it was wrapped.

Now i wonder if the sympton i have is a sympton u get when u disable throttle and thus that previous owner maybe disabled it?

Before i am returning the notebook i wonder if it this throttle setting can be enabled again so my GPU stays normal and i can play a game decent?

 

this was not at all my experience. i had a steady 9.5k score. even made a youtube vidoe on it YouTube - ASUS G51JX-A1 GAMING NOTEBOOK NVIDIA GTS 360M VIDEO REVIEW & BENCHMARKS
notice 3dmark came in at 9.5k. make sure you have p4g set to full performance. On a stock setup without any tweaks you will get 9.5k in 3dmark06

 

Thanks for your reply.. I have tested it with P4G set to the max and i have tried it without P4G installed but it makes no different.

The CPU scores do not drop in 3Dmark only the SM scores. With normal brightness and and keyboard lights on the score is a pathetic 7.8-8.3k and this is after totally tweaked in nvidia control panel for performance

I have to turn of the keyboard lights and totally dim the screen to get an exact 9.5 score

This is out of the box and after a clean install

It is really starting to annoy as im truly no hardware noob and my options are running out i can think of to fix thus

PS and yeah i know that review, i watched a few times before buying the notebook lol. nice review tho. funny its from you

 

try removing the battery with max brightness
any change in score?

 

g73 150w ps is 100% compatible but the 120W has plenty of juice to power the g51. it is not underpowered

 

Testing it now.

Nice youtube vid btw, i watched it like 10 times before buying this notebook

 

Nope no change

8.1k score
3k CPU score and again low SM scores

screen brightness affects like 1.5k of the score and like 1k on theSM scores wich is like 25% on the GPU what is huge

Running on max performance battery now, unplugged the AC.. kinda makes me wonder.. ---> Nah 3k score lol. 3k on CPU and 1K on SM

 

"after totally tweaked in nvidia control panel for performance"
^- reset defaults, things like triple buffering LOD clamp tend to hinder more than help.

get http://www.fileden.com/files/2008/3/3/1794507/ThrottleStop.zip.
make sure BD PROCHOT is unchecked.
run 3dmark again.

if your low performance is because of the system throttling, your score will improve. otherwise, i would look at temperatures/drivers/powermizer/system/fragmentation or anything else that might cause a lower score.

you could try a different battery test; discharge it a bit (80-90%), and then rerun with brightness on full + battery charging. this should bring your score down even lower.

 

I ment its custum tweaked, i know what the settings do and cause ^^
Unfortunatly i have to return the notebook. Normally not the problem but im kinda low on free time and the retailer has to pick it up and bring a new one the next day so that's 2 days.

Thanks for the reply tho

 

BTW throttlestop is for CPU throttling if i am not mistaken right?
It is not the CPU, the CPU seems 2 always score the maximum.
Always 3100 points in 3Dmark
In this case it is the GPU that "throttles" when using an screen brightness above 0%
Whith complete dimmed display and keyboard lights off the SM scores are around 4100 the CPU around 3050
When turning on brightness the SM scores drop the CPU score stays the same. With max brightness both SM scores are dropped to 3000, CPU stays on 3100.

When enabling extreme turbo the CPU score ups to 3250 but the SM scores drop even more.

Every bit of power seems to be drained from the GPU.

Besides, out of the box 3dmark score should be like 9.5k right? without screwing around with throttle stuf. My out of the box score is 7.8k - 8k

I mean basicly on stock settings and out of the box i should be able to get 9.5k points despite the throttle issue right? So if i am getting 1500+ less there should something be something wrong more

 

Okay i tried throttlestop and now i get 9.5k exactly.
Still out of the box without throttle stop i am getting 8k

Would u recommend using throttle stop or swap the system?
I mean reading all these post normally u would get 9.5k without throttle stop so maybe swap is better?

Can throttlestop harm your system when intensively gaming??

 

Just to clarify for all, while the GPU scores look much lower and the CPU scores stay the same, it is CPU throttling *not GPU throttling* that is causing it.

CPU test: v.little GPU use, power draw within limits, no CPU throttling.
GPU tests: 100% GPU use, combined power draw with CPU is over limit, CPU throttles to 933MHz, ***GPU is bottlenecked and can't work to 100% utilisation***, GPU score hurt.

Additionally, numerous tests have shown it's the CPU clocks that drop, while GPU clocks remains unchanged.

You can simulate what a 933MHz quadcore i7 feels like by using power4gear's "Battery saving" mode. :vomit:

 

Hey guys... I think I found another game where throttling occurs intensively - Starcraft 2. When I adjust settings to ultra and set resolution to 1920x1080 and i started to log the multiplier while gaming I noticed that there are lots of slowdowns during the game. And when I checked the multiplier in the logs it showed me that there was a lot of throttling:/ The multiplier was 7 during slowdowns...After that I used the throttling fix, and there were no slowdowns, there was about 20-30 fps all the time with ultra settings (the clocks were 580/870/1450) on my gtx 260m. The temps went up to 89C during about 15 minutes of gameplay...And I have max 85C on all my other games and overclocked..Does anyone have the same problem? Maybe we could request to change the system on warranty because of that?

 

Hi, i have ASUS G51JX-SZ167V with Intel® Core™ i5 Processor i5-520M. Does 1FC toggle works with i5 or it is only for i7? Thanks for answer.

 

I have an Asus G51J-A1. I have recently (last month) been experiencing lag issues in game that seems related to these throttling issues the notebooks are so notorious for. It is only whenever I play newer games such as SC2 & World of Warcraft. I downloaded throttlestop and changed the settings according to the first post on this thread and to no avail, it is still lagging in the game for about 2-3 minutes at a time every 5-10 minutes. Could it be a heating issue? I notice that my CPU gets up to 90 degrees on a high load and my GPU reaches even close to 105 degrees Celsius when I check throttlestop while playing these games. I'd like to nail the problem down so I can once again game in peace lol. These temperatures seem outrages when I think about it and my house is freaking cold (ambient temp around 60-70 degree Fahrenheit). I am running out of ideas and am thinking about reformatting my laptop to see if that will fix anything.

 

Your temps are probably making the GPU clock down, this probably isn't the same as the throttling issue discussed here. You might want to clean out your fan area to make sure it's not clogged with dust, and think about redoing your thermal paste. 105 is about the upper limit you'd see on a G51J.

 

have people returned their notebooks to asus for the newer version or do the newer ones also suffer similar problems? I havn't been in touch with this thread since i got my laptop.. but it lags during windows live movie player. like every 10 seconds of editing or playing a video, it will lag for 3 seconds. pretty annoying

 

I'm not sure what you are having trouble with -- did you install the package successfully?

 

Hello,

Sorry for my writing but I am not english...

First, I want to say thanks for this fix. It seem working !

However, I worry about this technique. Is it dangerous ? How to control it to don't break the computer ?

Some people says that I must look the temperature of the components during my games but if I can automatise all, it will be better (Maybe a software exist to control it ?).

So, is this dangerous for the computer and if yes, how do you control it please ?

Thanks a lot and have a nice day !