PDA

View Full Version : Cpu or Fan problem ???



Wishmaker77
12-08-2014, 02:41 PM
I tested my new G751 JY in benchmark firestriker and rest of them , make stress test of Cpu , play alot games with all on ultra and see some problem .... my high temps on Cpu alot times goes on 95 in same time Gpu high temps never goes even on 70 .... alot times i have difrence in temps 30 degrees between Cpu and Gpu ... my fans all time are very silence , so i ask its a problem with Cpu paste or spome trouble with fans , i try check Asus Gpu tweak , but there i see fans on 5800 and when i put make fan on full speed , nothing change so Gpu tweak doesnt work :(

hmscott
12-08-2014, 03:17 PM
I tested my new G751 JY in benchmark firestriker and rest of them , make stress test of Cpu , play alot games with all on ultra and see some problem .... my high temps on Cpu alot times goes on 95 in same time Gpu high temps never goes even on 70 .... alot times i have difrence in temps 30 degrees between Cpu and Gpu ... my fans all time are very silence , so i ask its a problem with Cpu paste or spome trouble with fans , i try check Asus Gpu tweak , but there i see fans on 5800 and when i put make fan on full speed , nothing change so Gpu tweak doesnt work :(

Wishmaker77, that's because the version that has the Full Fan option is for Optimus laptops - your JY doesn't have Optimus.

If your JY doesn't come with Asus GPU Tweak for Graphics Cards, it may not work. But, you can try, here is the download area specific to the non-Optimus GPU's

Asus GPU Tweak for Graphics Cards, currently at release 2.6.5(.3) for the G750JW/JX/JH and G751JT/JY(??)
http://support.asus.com/download.aspx?SLanguage=en&p=9&m=ASUS%20GPU%20Tweak%20for%20Graphics%20cards&hashedid=n/a

There are newer versions, up to 2.7.1.8, but they cause problems on G750's, but they might work on the G751JT/JY, if you find the best running version for your JY, please post back here :)

The peak temp of 95c for CPU is high, but you really need to monitor your temps in between peaks too, and monitor Thermal Throttling as that is really what the cooling system is trying to avoid. When the CPU thinks it is too hot it Throttle itself and kills performance - you want to avoid that. If you aren't seeing Thermal Throttling, then you aren't running too hot for the CPU.

Asus makes the fan curves for quiet operation, but cool enough for long life and under the thermal throttling limit.

Use a tool like hwinfo64 to log temperature readings, and thermal throttling events, and leave it running in the background. Then you can look at the log file, it's a csv file, with Excel and see how often the peak temperature was reached - more often then not it doesn't happen very often and most of the time you are running well below that even under heavy gaming/application loads.

Before loading the .csv logfile into Excel I use Notepad to read it, and globally replace the semi-colon's with comma's, and save to a different name.

You said you are seeing problems when playing games, what kind of problems?

BTW, the Full Fan option on the Optimus laptops is not to be envied, it is lame to require manual intervention to keep things cool. The non-Optimus laptops keep cool automatically :)

Please let us know how the correct Asus GPU tweak works for you, and what the hwinfo logs show for temperature and thermal throttling. :)

43989

43990

43991

Wishmaker77
12-08-2014, 04:15 PM
Ok i install this new version 2.7.1.8 but i dont see any fan speed .... i need some program to checks CPU and GPU fan speed when i stress them :) The bad think i have throltling on cpu i using hwinfo64 .... but averange temps when i start hwinfo64 and put game all on ultra i play 2 hours then i stop and check temps , i have averange on CPU 65-75 , on GPU i have 50-60 , but problem is that highest temps on CPU are 95 and i have thrttling on cores on yes :( Gpu never have highest even 75 :(

hmscott
12-08-2014, 04:22 PM
Ok i install this new version 2.7.1.8 but i dont see any fan speed .... i need some program to checks CPU and GPU fan speed when i stress them :) The bad think i have throltling on cpu i using hwinfo64 .... but averange temps when i start hwinfo64 and put game all on ultra i play 2 hours then i stop and check temps , i have averange on CPU 65-75 , on GPU i have 50-60 , but problem is that highest temps on CPU are 95 and i have thrttling on cores on yes :( Gpu never have highest even 75 :(

Wishmaker77, what are you running when it hits 95c? The samples are every 2 minutes by default, how many times did you hit Thermal Throttling? On 1 or more cores?

We don't really have a lot of data yet, if you can post what causes the temps to hit 95c and thermal throttling others can try it - it might just be the way it is, or your cooling system might be just a tad off, we don't really know yet.

Don't worry about it yet.

None of the ROG laptops have any fan control, just the Full Fan option, which is *really* loud, you aren't missing anything :)

Artixz
12-08-2014, 07:03 PM
I have a one question about cpu fan. When the fan turns on it will never stop, is this normal? My temps on idle 37 degrees but cpu fan still working.

hmscott
12-08-2014, 07:12 PM
I have a one question about cpu fan. When the fan turns on it will never stop, is this normal? My temps on idle 37 degrees but cpu fan still working.

Artixz, yes the fan will always be cooling even at idle. I think the fan drops down to 2300 rpm at idle, at least on the G750JH.

The only program I have found that actually reads the CPU Fan speed is AIDA64. It shows it in the Sensors summary, and in the stress testing section there is a Fan speed graph.

Here is my idle fan speed of 2300

44000

44001

Artixz
12-08-2014, 07:23 PM
yes. Aida64 shows fan rpm. On idle have 2300, on load 2700 rpm. My rog on idle could be quieter

hmscott
12-08-2014, 08:02 PM
yes. Aida64 shows fan rpm. On idle have 2300, on load 2700 rpm. My rog on idle could be quieter

Artixz, which model Asus ROG laptop do you have?

Do you have the Power Plan set to Balanced? Let the CPU cool down in between loads, set the CPU Min/Max to 0%/100%. That might help keep the fan from spinning up.

And, try setting the CPU voltage offset to reduce power through the CPU, reducing temperature. I have mine set to -45mV for CPU and cache.

Mine idles at 2300rpm, and the ambient noise drowns out the fan, it is inaudible. At full load, the fan runs at 3900rpm - quite audible :)

If you have a HDD it might also be adding to the noise, the vibration from spinning conducts through the chassis and might be echoed up from the surface it is sitting on.

Wishmaker77
12-08-2014, 10:04 PM
ok my fans checking with Aida64 2200/2300 idle , stres 2600/2700 . I make real time test i use riva tuner with hwinfo64 and i think i find why i all time see on hwinfo high 95 on cpu .... testing now on Civilization : beyond Earth .... i start game temps on cpu 60-65 , fps 60+ cpu using 20% , when i choose all and put start , then on 1 s my cpu goes 100% and temps on all cores on 92+ .... game start all cores max 65 , fps 60+ all stable ... so normally i test only hwinfo64 and when i end playing i see this high temps 92+ ... but i think its just 1 sek ....

hmscott
12-08-2014, 10:17 PM
ok my fans checking with Aida64 2200/2300 idle , stres 2600/2700 . I make real time test i use riva tuner with hwinfo64 and i think i find why i all time see on hwinfo high 95 on cpu .... testing now on Civilization : beyond Earth .... i start game temps on cpu 60-65 , fps 60+ cpu using 20% , when i choose all and put start , then on 1 s my cpu goes 100% and temps on all cores on 92+ .... game start all cores max 65 , fps 60+ all stable ... so normally i test only hwinfo64 and when i end playing i see this high temps 92+ ... but i think its just 1 sek ....

Wishmaker77, cool, that was what I was hoping, a single spike and then the rest of the time everything is running coolly.

That is why it is good to look at a logfile of the entire run of the test, with a sampling rate often enough to see the dips and peaks in the load / temperature.

Your temps are aok :)

Wishmaker77
12-08-2014, 10:30 PM
I hope u have right ;) One q hmscott , if i wanna repast cpu i loose guarantee ???

hmscott
12-08-2014, 11:51 PM
I hope u have right ;) One q hmscott , if i wanna repast cpu i loose guarantee ???

Wishmaker77, yes, unfortunately the kind of paste that Asus uses is distinctive looking and probably includes a heat curing process that you would have trouble duplicating - perhaps not, but it does cure solid and bubbly around the squished out edges that show. They would know you opened it, let alone being obvious it was re-pasted.

Usually people wait till just before/after the expiration of their warranty to open it up and re-paste. Often they will try to get it re-pasted before the warranty expires, but Asus is pretty clear it actually needs to be thermally throttling during normal operation before they will re-paste.

It's really Asus's job to fix it during the warranty, that is why they get the big bucks ;)

Wishmaker77
12-11-2014, 01:35 PM
Hmm today i make test on Wasteland 2 gog version full patched , i make all on max in game , and starts used to test (hwinfo64 with riva tuner and aiwa to looking on fan speed ) 30 minutes of game i see some temps even 96 in cpu ... gpu never goes high then 72 , fan speed i see even 3100 ... but all time averenge on cpu are 69-75 .... The funny think is that when i have temps 90 in game i make fast jump on window screen and i see how temps goes to 70 in 2 seconds ... or if i ends game i dont need wait 5 minutes to temps on my cpu goes to idle 40-45 but its take maybe 10-15s .... its so weird .

Sagi
12-11-2014, 03:23 PM
You should return it for RMA.

The G751 is not ment to run @ 90c or higher :rolleyes: 70-85c is the comfor zone according to ASUS's support.

hmscott
12-11-2014, 04:18 PM
Hmm today i make test on Wasteland 2 gog version full patched , i make all on max in game , and starts used to test (hwinfo64 with riva tuner and aiwa to looking on fan speed ) 30 minutes of game i see some temps even 96 in cpu ... gpu never goes high then 72 , fan speed i see even 3100 ... but all time averenge on cpu are 69-75 .... The funny think is that when i have temps 90 in game i make fast jump on window screen and i see how temps goes to 70 in 2 seconds ... or if i ends game i dont need wait 5 minutes to temps on my cpu goes to idle 40-45 but its take maybe 10-15s .... its so weird .

Wishmaker77, are the jumps into the 90's single readings saved in the log, or more than 1 reading in a row? If it briefly goes high it isn't an issue, and you didn't mention if Thermal Throttling goes from No to Yes, does it do that?

If you can still take it back to the seller, swap it for another one, run the same tests, and see if you come up with lower temperatures.

Wishmaker77
12-11-2014, 05:28 PM
Yes i have throttling on all cores .... the worst is g751 have so many others issues that i scare to change him on other .... here i have only this problem maybe better send him on repast .... cause i f i change in shop on other i dont know what they give me back :(

hmscott
12-11-2014, 05:39 PM
Yes i have throttling on all cores .... the worst is g751 have so many others issues that i scare to change him on other .... here i have only this problem maybe better send him on repast .... cause i f i change in shop on other i dont know what they give me back :(

Wishmaker77, too bad, there haven't been many reports of Thermal Throttling, so you did get one with a cooling issue. Re-paste isn't always the only problem, sometimes it is a cooling component - heat pipe - fitting - etc also involved.

If you still are in the return period with the seller, return to them for a new one. Run the exact same series of tests and compare the results :)

f you can't return/swap you can RMA with Asus for a repair to the cooling. Asus will want to see something with the Thermal Throttling, so you can include the lines from the log with Thermal Throttling, either as text in the Technical Inquiry description or attached as an Image or .csv file attached to the Technical Inquiry.

https://vip.asus.com <== create a login, register your laptop, and file the Technical Inquiry with a clear problem description and request an RMA repair.

Please let us know how it works out :)

Wishmaker77
12-11-2014, 09:09 PM
I make now this and zero throttling .... maybe this game have some bugs .... 44144 and 44145

hmscott
12-11-2014, 09:28 PM
I make now this and zero throttling .... maybe this game have some bugs ....

Wishmaker77, usually heat related issues aren't limited to just 1 game/application, the same load from other applications generating the same load will have the same heating problems.

The AIDA64 stress test isn't very stressful. Note how the fan speed only got to 2900, when I run prime95 Small FFT test on 8 threads the fan speed goes to 3700

Try running prime95 and watch the fan speed.

You should also use hwinfo64 with logging on, so you can read the log for temperaturs between peaks and also see how often you are hitting Thermal Throttling.

hwinfo64 download
http://www.hwinfo.com/download.php

prime95 download
http://www.mersenne.org/download/

prime95 running the Small FFT test will be a good source for results to prove you are Thermally Throttling to Asus if you need to RMA. You can also include the results from gaming as well, but it is easier for Asus to reproduce the prime95 test runs than to use your game example - they might not have the game or the right location in the game to generate the same results.

Wishmaker77
12-11-2014, 09:30 PM
But look when i make Stress FPU ...... 44147 , 44148

hmscott
12-11-2014, 09:37 PM
But look when i make Stress FPU ......

Wishmaker77, what program are you using to see the Thermal Throttling?

Wishmaker77
12-11-2014, 09:39 PM
hwinfo 64 , aida64

hmscott
12-11-2014, 09:52 PM
hwinfo 64 , aida64

Wishmaker77, enabling more stress testing features increases the load, but you can leave the disk test disabled - that might slow things down.

If you like enable all the CPU stress test options, and the GPU stress test, and then watch temps with hwinfo64 and GPU-Z for GPU.

Then compare against the load generated by primed95 + furmark 15 minute burn-in test - I run those simultaneously for maximum load on CPU/GPU for stress testing.

Here are the results from running after 5 minutes on my G750JH for comparison:

44149

Wishmaker77
12-11-2014, 10:14 PM
Hate this stupid 8.1 windows .... i work with him from g751 ... dont know anything even how prt scr .... i make all u wanna gpu 76 max better then u but cpu 95 and thrrolling

hmscott
12-11-2014, 10:32 PM
Hate this stupid 8.1 windows .... i work with him from g751 ... dont know anything even how prt scr .... i make all u wanna gpu 76 max better then u but cpu 95 and thrrolling

Wishmaker77, I use the Windows Snipping Tool for screen/dialog grabs, it lets you select the area of the screen to grab and you don't need to go into mspaint to crop/save.

https://www.google.com/search?num=50&safe=off&q=how+to+use+windows+snipping+tool&spell=1&sa=X&ei=bxyKVKqyDZTToATWyIGAAQ&ved=0CBwQvwUoAA&biw=1226&bih=585

If your CPU is hanging around 95c during the prime95 test then it is running too hot...

Wishmaker77
12-11-2014, 10:51 PM
44164

Wishmaker77
12-11-2014, 11:00 PM
And here u have full .... 44166

hmscott
12-11-2014, 11:02 PM
...

Wishmaker77, that is kinda on the borderline... you did Thermal Throttle on Core 2, but your current temps, where it hangs out most of the time, is in the high 80's and likely not Thermal Throttling.

prime95 small FFT is pretty harsh on CPU's, and although it would be best if you weren't Thermal Throttling at all, it isn't too bad.

But, you are Thermal Throttling in games as well, so I would swap it for another one... it's a pain, but you might as well get a clean running one from the start - there haven't been any other G751's with Thermal Throttling reported either, so you should get a good one on the next go around.

Any other complaints with this unit?

Wishmaker77
12-11-2014, 11:09 PM
Only this .... i thinking what to do ... cause is so many isues with G751 , so i afraid change to new one and have bad ips display ... trouble with toupad , wi fi .... or others ... The best is that cooling in my laptop is amaziong cause from this temps which i screen u he backs to 45 in 3 s .....

hmscott
12-11-2014, 11:14 PM
Only this .... i thinking what to do ... cause is so many isues with G751 , so i afraid change to new one and have bad ips display ... trouble with toupad , wi fi .... or others ... The best is that cooling in my laptop is amaziong cause from this temps which i screen u he backs to 45 in 3 s .....

Wishmaker77, yes once the load is removed the cooling brings the temps way down quickly.

That furmark run confirms the Throttling, this time on more cores.

It is a tough decision, with the other faults that might come about. IDK what to tell you :)

You have 1 or 2 years of warranty, so you could wait and RMA it for just the cooling RMA - and over time it might get worse so you wouldn't have trouble getting it RMA'd for the problem, but it is quicker/easier to swap it for a new one now during the seller return period.

Good luck either way :)

Wishmaker77
12-11-2014, 11:20 PM
I wait morning Friday and call to my shop , i buy 7 days ago so i have 3 days more to give back or change to new :) Thx alot hmscott :)

cykelman
12-11-2014, 11:23 PM
Did the same test (5 min furmark and prime95 small fft) on my computer as well, again, what do you think? (I know i might be overthinking it, get really anxious about investments such as this though) http://speedy.sh/Ee9j4/prime95-furmark.CSV

hmscott
12-11-2014, 11:36 PM
I wait morning Friday and call to my shop , i buy 7 days ago so i have 3 days more to give back or change to new :) Thx alot hmscott :)


Did the same test (5 min furmark and prime95 small fft) on my computer as well, again, what do you think? (I know i might be overthinking it, get really anxious about investments such as this though) http://speedy.sh/Ee9j4/prime95-furmark.CSV

Wishmaker77 / cykelman, it may be that the cooling system in the new G751's will Thermal Throttle during such an intensive test that loads both the GPU and CPU 100%.

The new G751's cooling systems are merged, instead of being separated / independent as in the G750.

Here are photo's of both cooling systems

First the G750JX:

44168

The G750JH/JZ are the same layout, but with bigger heatsinks/fans:

44170

Here is the G751JY, note how the CPU/GPU cooling pipes are connected to the same plate, and the heat from both the GPU and CPU shares the same cooling parts:

44169

From previous testing cykelman's results are good from other benchmarks, but in this combined CPU/GPU test he is actually Thermal Throttling the same as you are Wishmaker77...

Maybe this is as good as it gets...

I wouldn't trade in your laptops yet, maybe what you are seeing is as good as the G751 combined CPU/GPU cooling gets with heavy load on CPU/GPU simultaneously.

Wishmaker77
12-12-2014, 02:41 PM
No others people making same tests and have mac cpu 82 ... so its bad pasting or some worst isues :( I will sand to my shop for new laptop ....

hmscott
12-12-2014, 03:16 PM
No others people making same tests and have mac cpu 82 ... so its bad pasting or some worst isues :( I will sand to my shop for new laptop ....


u have 2 times bad luck .... i have same terrible trottling temps and i will send laptop to shop ... alot people makes prime with furmark and have max 82 on Cpu so we have bad pasted or even worst isues

Wishmaker77, I have seen a JY post 82c for running prime95 blend test, running alone without furmark, but a JT/JY can't do the combined test and get a CPU temp of 82c. At least we haven't seen it so far.

Heating up the CPU and GPU causes the JT/JY to run out of of cooling headroom, and that is why it is throttling the CPU on yours and others tests.

Your results are normal... until we get someone with a test run of the same tests and duration as you have run that does better

There are several test options in prime95, if you ran Small FFT that is the hottest running test, if someone else ran another test, it will read lower in temperature.

44185

prime95 Small FFT + furmark 15 minute burn-in tests simultaneously.

My G750JH did CPU 85c, GPU 85c.

If you have a JT/JY please run the test and post your results, so Wishmaker77 and others can compare, and hopefully avoid return/RMA.

prime95 Small FFT + furmark 15 minute burn-in together, looks like this:

44184

If you run prime95 Small FFT alone, the CPU temp will be lower, like this on the G750JH:

44186

Without the GPU heat load of furmark also running at the same time the CPU will run cooler.

Wishmaker77, please give the single prime95 Blend and FFT tests alone a run and post the results.

rmuniak
12-12-2014, 04:01 PM
I ran the prime95+uingine tests with all cores 100% and GPU 100%, and CPU wasn't throttled. It was working with basic 2,4ghz speed at all cores, keeping the temperatures at about 90 degree celcius with 91 as maximum. GPU stayed very cool at 75 max. Probably most of us will get very similar results.

Turbo is working on max if the CPU is not heavily loaded. Under heavily load it minimizes from 3,4ghz to 2,6ghz and if the heat gets on it stops at initial clock- be aware that most games won't stretch this CPU that much - Crysis 3 worked with all 4 cores running at max Turbo. I haven't noticed any terhmal throttling yet and hope to never notice it :)

hmscott
12-12-2014, 04:10 PM
I ran the prime95+uingine tests with all cores 100% and GPU 100%, and CPU wasn't throttled. It was working with basic 2,4ghz speed at all cores, keeping the temperatures at about 90 degree celcius with 91 as maximum. GPU stayed very cool at 75 max. Probably most of us will get very similar results.

Turbo is working on max if the CPU is not heavily loaded. Under heavily load it minimizes from 3,4ghz to 2,6ghz and if the heat gets on it stops at initial clock- be aware that most games won't stretch this CPU that much - Crysis 3 worked with all 4 cores running at max Turbo. I haven't noticed any terhmal throttling yet and hope to never notice it :)

rmuniak, which prime95 test did you run?, the default one that comes up - Blend - or did you select the more demanding hotter running Small FFT? Small FFT is what Wishmaker77 and others ran, simultaneously with furmark 15 minute burn-in - which also is more demanding than Unigine Heaven.

What did you mean that "working with basic 2,4ghz speed at all cores"? With a Power Plan with CPU Max set to 100%, it shouldn't have dropped below 3.2ghz - unless the CPU was Thermal Throttling...

Would you please do a run as discussed above, prime95 Small FFT + furmark 15 minute burn-in tests run together at the same time, you only need run it for 5 minutes - and capture the hwinfo results as well as the tests results as above.

Thank you!

rmuniak
12-12-2014, 04:57 PM
hmscott running it right now (prime95 small fft + fumark (i disabled burn-in in option) - 70 on GPU and 82-89 on CPUs after more than 5 mins.

You need to check what Thermal Throthling is. If you're CPU is not working on Turbo, this doesn't mean if it's throtthled. There is not enough power to work on max Turbo on 100% CPU and GPU - TDP point is being reached. We would have thermal throthling if base clock of any core would be lower than x24.

Result test:
44187
I find prime95+uingine more stressfull - GPU got to 74 on that test and all cores were at x24. Still 91 was max temperature registered on both test.

hmscott
12-12-2014, 05:31 PM
hmscott running it right now (prime95 small fft + fumark (i disabled burn-in in option) - 70 on GPU and 82-89 on CPUs after more than 5 mins.

You need to check what Thermal Throthling is. If you're CPU is not working on Turbo, this doesn't mean if it's throtthled. There is not enough power to work on max Turbo on 100% CPU and GPU - TDP point is being reached. We would have thermal throthling if base clock of any core would be lower than x24.

Result test:
44187
I find prime95+uingine more stressfull - GPU got to 74 on that test and all cores were at x24. Still 91 was max temperature registered on both test.


rmuniak, thank you for running the test, but you have altered the test and the CPU speed configuration, and therefore your test isn't the same as Wishmaker77, or I ran.

If your CPU is running at 24x instead of 32x-35x like Wishmaker77 your temperature readings aren't comparable - his CPU is running a lot faster and therefore hotter.

Please run again with the Power Plan set to High Performance, CPU Min/Max 100%/100%, and don't detune the CPU with XTU, in fact it looks like Wishmaker77 might have the multipliers pushed up with XTU, as he is running the same Max multiplier I am.

Also, please don't alter the furmark test paramters, your performance viewable in the screen shot is way below Wishmaker77 and also lower than my results - both of your results should be higher fps than mine.

Your CPU performance was detuned, so your GPU performance was held back as well.

If you run the same test that Wishmaker77 ran, at the same full speed CPU, your CPU/GPU temps are going to match his, and you will Thermal Throttle. :)

Look at Wishmaker77's post and screen shot here:
http://rog.asus.com/forum/showthread.php?54521-Cpu-or-Fan-problem&p=456743&viewfull=1#post456743

Compare your CPU speed / multi range to his, and his FPS to yours, you are running a much detuned test.

rmuniak
12-12-2014, 06:14 PM
Hey mate,

I'm not altering anything. You need to learn more about current generation processors and how they're working. Here is the nice document describing everything:
http://www.intel.com/content/dam/www/public/us/en/documents/datasheets/4th-gen-core-family-mobile-m-h-processor-lines-vol-1-datasheet.pdf

It is processor which determines whether to put turbo or not and what should be the correct multipler - for mine 24x is base one and everything above is controled by CPU - everything above is Turbo and it depends on variety of factors - in this case it is not temperature which sets the multiplier lower since i get the same results only on prime95 running alone - it is TDP 47 W points being reached. I assume 4170HQ is a better choice for this laptop.

Wishmaker probably has different processor or different BIOS (mine is 4860HQ and Bios 205).
IHe was also running less stressfull test on Furmark (mine was max msaa on full hd).

BTW: we have no proof that numbers shown on HWInfo are correct... With different sensor tool we might get different numbers...

hmscott
12-12-2014, 06:46 PM
rmuniak, you altered the furmark test. That changes the test and makes it non-comparable. Why on earth would you do that? The whole point of running the same tests and same parameters is to make the test runs the same, so they can be compared. :)

Your CPU isn't acting correctly, as I would expect it to, it should have higher multipliers showing. Higher Max, Higher Min, and Higher Current - the same or 1x greater than the 4710HQ.

The 4860HQ came out a full Quarter ahead of the 4710HQ release, and hwinfo is reading the multipliers fine for the 4710HQ, so hwinfo is reading the 4860HQ correctly too. Download a couple of other tools and compare to verify to yourself, but you will find they are correct.

You need to double check your CPU Min/Max settings on your active power plan. And, if you have XTU loaded, set back to defaults.

If you have Asus Power4Gear installed, try uninstalling it and use a Windows Power Plan instead.

Your 4860HQ CPU should be performing slightly better than a 4710HQ, not worse.

Have you run other benchmarks and compared them against other JYs? Are the results similar?

Sorry to have dug this up, but you need to get a handle on it - something might be wrong with your laptop...

rmuniak
12-12-2014, 06:57 PM
Your CPU isn't acting correctly, as I would expect it to, it should have higher multipliers showing. Higher Max, Higher Min, and Higher Current - the same or 1x greater than the 4710HQ.

The 4860HQ came out a full Quarter ahead of the 4710HQ release, and hwinfo is reading the multipliers fine for the 4710HQ, so hwinfo is reading the 4860HQ correctly too. Download a couple of other tools and compare to verify to yourself, but you will find they are correct.

Why is so? Why would you expect that it will have 1x greater multiplier? Because it's newer? :D

It doesn't matter whether you have high performance set or not - i tried both and cpu was doing the same thing - under load on normal setting cpu will get as many power as needed. I suppose i don't have the same configuration as Wishmaker - these 2 processors have a bit different architecture, that's why might behave differently and give different results - but we're getting very similar temperatures!

44191
prime small fft + uingine almost 15 minutes (2 tools at the same time - same temperatures)- i don't have heart to push more :) You can see that voltage for processor is almost 47 - it can't pass that value that's why it is running 25-26x multiplier.

Wishmaker77
12-12-2014, 07:33 PM
Here hmscott i give u from some other g751 JY owners test 44197 12 minutes on f44197ull zero throttling ...

rmuniak
12-12-2014, 07:38 PM
Wishmaker which CPU you have? The CPU on this test you pasted behaves as mine in terms of mulipliers. But the GPU temperature it's really high there!

I will do same fumark+prime test as yours and this guy tomorrow.

hmscott
12-12-2014, 07:53 PM
Wishmaker which CPU you have? The CPU on this test you pasted behaves as mine in terms of mulipliers. But the GPU temperature it's really high there!

I will do same fumark+prime test as yours and this guy tomorrow.

rmuniak, thanks, and I will look at them, but for Wishmaker77, it might be too late, he needs to swap his unit for another one today / this morning as above...

The last 2 images of runs he posted were from other JY's, not his, he was showing how they didn't Thermal Throttle. Their runs as yours ride the edge of Thermal Throttling, hitting 92c but not Thermal Throttling, because that starts at/after 93c, pretty darned close.

There was another user that posted results like Wishmaker77, same temps / thermal throttling.

I wrote a response to your earlier post, but lost it when going back and forth between all the results, I might redo it later...

The 1x difference in the processors is documented here:

http://ark.intel.com/compare/76089,78930

The 4860HQ multiplier range is 24x - 36x, and the 4710HQ is 25x-35x, yours is 1x better.

Have you installed XTU yet? What are the max OC multipliers for the 4860HQ?

hmscott
12-12-2014, 08:07 PM
Here hmscott i give u from some other g751 JY owners test 12 minutes on full zero throttling ...

Wishmaker77, if you have a 4710HQ processor, you might try swapping for a model with a 4860HQ processor, or the other way around. Both your run and cykelman's run showed Thermal Throttling, and the same temps.

cykelman's post includes a .csv file from hwinfo monitoring during his run
http://rog.asus.com/forum/showthread.php?54521-Cpu-or-Fan-problem&p=456751&viewfull=1#post456751

To early to tell, but it might be worth switching to the other available processor as long as you are going in to swap laptops.

Good luck :)

Wishmaker77
12-12-2014, 08:13 PM
I have I7 4860 ....

hmscott
12-12-2014, 08:19 PM
I have I7 4860 ....

Wishmaker77, if the others aren't thermal throttling and you are, you probably want to swap for another, as we originally concluded. I was concerned that there might be a reason for the throttling based on the new cooler design when hitting it with full CPU and GPU loads, and then all JY's would have exhibited the Thermal Throttling with this load.

Glad to see that isn't the case and we have 3 samples of people without Thermal Throttling, and 4 samples with.

Can't wait to see your new laptop test results :)

Update: going back again and looking at the temps on the 3 runs that didn't Thermal Throttle, they all had max package temp of 92c, which is only 1c below Thermal Throttling - an extra warm day could put those systems into the same thermal range as yours, and Thermal Throttle.

Thermal Throttling starts at 93c and above.

cykelman
12-12-2014, 08:19 PM
And i have 4710HQ, A bit worried about these results, do i have grounds for a possible RMA in the future? (My return period is over)

hmscott
12-12-2014, 08:24 PM
And i have 4710HQ, A bit worried about these results, do i have grounds for a possible RMA in the future? (My return period is over)

cykelman, generally if you can show Thermal Throttling, yes that is a key to getting Asus to fix the cooling - re-paste - replace parts - etc.

You are hitting Thermal Throttling only when hitting both CPU/GPU hard with a stress test, or do you see Thermal Throttling in games / apps?

cykelman
12-12-2014, 08:32 PM
cykelman, generally if you can show Thermal Throttling, yes that is a key to getting Asus to fix the cooling - re-paste - replace parts - etc.

You are hitting Thermal Throttling only when hitting both CPU/GPU hard with a stress test, or do you see Thermal Throttling in games / apps?

Mostly in stresstest, though i throttled when playing my modded version of skyrim in high performance mode.

hmscott
12-12-2014, 08:36 PM
Mostly in stresstest, though i throttled when playing my modded version of skyrim in high performance mode.

cykelman, if you run skyrim modded in Balanced Mode, will it cool down enough in between to not Thermally Throttle?

cykelman
12-12-2014, 08:49 PM
cykelman, if you run skyrim modded in Balanced Mode, will it cool down enough in between to not Thermally Throttle?

Yeah, that works and as such i dont really have an issue with throttling in my daily activities. What i have been wondering is more along the lines of is if my results are unusual and if it would be worth it to rma in the future.

hmscott
12-12-2014, 09:00 PM
Why is so? Why would you expect that it will have 1x greater multiplier? Because it's newer? :D

It doesn't matter whether you have high performance set or not - i tried both and cpu was doing the same thing - under load on normal setting cpu will get as many power as needed. I suppose i don't have the same configuration as Wishmaker - these 2 processors have a bit different architecture, that's why might behave differently and give different results - but we're getting very similar temperatures!

prime small fft + uingine almost 15 minutes (2 tools at the same time - same temperatures)- i don't have heart to push more :) You can see that voltage for processor is almost 47 - it can't pass that value that's why it is running 25-26x multiplier.

rmuniak, 2 things threw me off in my comparisons between all the runs.

1) My runs had the hwinfo started before I started the prime95 and furmark, so I had normal multipliers showing for Max.

2) WishMaker77's image showed high Current multiplier , but that was because he had Thermally Throttled and some of the prime95 threads had aborted, so some cores were higher than normal multipliers - even though CPU utilization was high.

3) Your run must have already been running before you started hwinfo, so the multiplier readings were for the narrow band of values during the run, and didn't include the full normal multipliers.

I am used to having the tools in front of me to do testing, and trying to gather that same level of info for comparison is tough spread out.

Thanks for hanging in there and working with us on getting the runs done. I know it helped Wishmaker77 and cykelman right now to help them decide what to do with their laptops.

I was concerned that all the JY/JT's might be susceptible to Thermal Throttling under simultaneous CPU/GPU load, but I am happy to see at least 3 of the JY's aren't Thermal Throttling with this test. But, at least 4 are...

It would have been nice to isolate it to one or the other available CPU... but it looks like it doesn't matter if it is the 4710HQ or the 4860HQ, either one can Thermally Throttle, or now.

hmscott
12-12-2014, 09:05 PM
Yeah, that works and as such i dont really have an issue with throttling in my daily activities. What i have been wondering is more along the lines of is if my results are unusual and if it would be worth it to rma in the future.

cykelman, well it looks like you and Wishmaker, and CajunAzn on notebookreview.com had 1 that he knows Thermally Throttled, and his first one was also likely Thermally Throttling - temps over 93c, but he didn't know to look for Thermal Throttling when he had that one.

And there is rmuniak and the 2 examples of JY's that didn't Thermally Throttle.

So that is 4 that Thermally Throttle and 3 that don't... in our small sample pool :)

So, I would say your results aren't abnormal, but some don't exhibit it - at least not in the tests we have tried so far.

Update: Now that rmuniak has a test where he has ever so slightly Thermal throttled himself, I am thinking this is normal, and noone needs to swap laptops...they are all within +-2c in tests - probably due to ambient temperature or other local environmental effects.

cykelman
12-12-2014, 09:24 PM
cykelman, well it looks like you and Wishmaker, and CajunAzn on notebookreview.com had 1 that he knows Thermally Throttled, and his first one was also likely Thermally Throttling - temps over 93c, but he didn't know to look for Thermal Throttling when he had that one.

And there is rmuniak and the 2 examples of JY's that didn't Thermally Throttle.

So that is 4 that Thermally Throttle and 3 that don't... in our small sample pool :)

So, I would say your results aren't abnormal, but some don't exhibit it - at least in the tests we have tried so far.

Even if the cooling is slightly worse on mine so long as it doesn't throttle during my regular activities there is no rush to RMA it right? (Better to wait a few months, collect more data and such)

hmscott
12-12-2014, 09:46 PM
Even if the cooling is slightly worse on mine so long as it doesn't throttle during my regular activities there is no rush to RMA it right? (Better to wait a few months, collect more data and such)

cykelman, right, you don't have an immediately failing situation to fix, it is more of an edge case which is so brief that even if it hit in normal usage you wouldn't notice.

If you needed to do long batch jobs for rendering, I have some XTU tweaks that worked for G750JM/JS/JZ users to reduce the multiplier on core 3/4 but keep the core 1/2 multi's high, infact OC them - so single and dual threaded jobs can run at full speed.

Sometimes a heatsink in use will seat and cure over time and the temps will get better... I have had that happen a few times, as much as 5c - but that was different materials - I don't know if there is any such behavior possible with Asus thermal compound / pads. Unfortunately you can't get in there to turn the screws to continue seating the parts.

I wouldn't baby it, use it full bore as you would if it wasn't throttling, and enjoy it :)

hmscott
12-13-2014, 09:39 AM
Guys, going back and looking at the 3 runs that didn't Thermal Throttle, they came very close to Thermal Throttling, only 1c under the Thermal Throttling point. Thermal Throttling starts at 93c and above, and their CPU core and package temps hit 92c, just 1c below the Thermal Throttling point.

An extra warm day, bringing up the ambient temperature a few degrees, might be enough to cause those same runs to exhibit Thermal Throttling.

Something to test for next Summer, or in an extra warm Winter house :)

Update: Now that rmuniak has a test where he has ever so slightly Thermal throttled himself, I am thinking this is normal, and noone needs to swap laptops...they are all within +-2c in tests - probably due to ambient temperature or other local environmental effects.

Wishmaker77
12-13-2014, 09:56 AM
Another test from other user .... ,btw i have very warm winter house ;))))) 44218 i think he has the best and ideal resulkts for g751 with 4860 and 980 m :)

hmscott
12-13-2014, 01:32 PM
.... btw i have very warm winter house ;)))))

Wishmaker77, that's it then, you need to run your benchmarks outside in the snow; bundle up and get tough :)

Seriously though you need to test in a cooler ambient environment, like 21c...

Update: Now that rmuniak has a test where he has ever so slightly Thermal throttled himself, I am thinking this is normal, and noone needs to swap laptops...they are all within +-2c in tests - probably due to ambient temperature or other local environmental effects.

rmuniak
12-13-2014, 02:21 PM
You're right guys- we all might have the same cooling and pasting, we're just testing in different environments. Cooler ambient helps, even the humidity counts :)

hmscott i4860hq has x36 is maximum multiplier but only for one core, for 2 cores running paralllel is x35 and x34 for all 4 cores.

Mine hitted 94 Celcius maximum and hwinfo didn't indicated Thermal Throtthling. Let's be frank - none of the current games will probably be able to put the CPU into that condition which Prime95 does, so probably we don't have to worry about. For now of course... As time will pass, dust will come to heating pipes and paste will loose it capabilities, thus we might start to thermal throttling under normal gaming conditions... I'm going to buy some cooling pad definitely - 2-3 Celcius less will be vital.
I also recommend to put the "Balanced" power plan - it will downclock the CPU to 8x when not in use, lowering the overall temperature and increasing lifespan of the cpu and paste. It's very intelligent - if some process needs a core, then it can hit max turbo. Maximum performance keeps max turbo on all cores even if there are not in used - thus increasing the temperature of the CPU.

UPDATE:
44221
Standard fumark with burn in + prime small fft. Even though i got thermal throthling indicator, you can see that multiplier haven't been lowered below the base value. Base clocks also stayed the same. Actual thermal throttling threshold set by Asus might be higher than the one idicated by HWInfo - otherwise we would see effects on minimum clock value or bus clock values. Please correct me if i'm wrong. Cooling pad for the win :)

hmscott
12-13-2014, 02:39 PM
You're right guys- we all might have the same cooling and pasting, we're just testing in different environments. Cooler ambient helps, even the humidity counts :)
hmscott i4860hq has x36 is maximum multiplier but only for one core, for 2 cores running paralllel is x35 and x34 for all 4 cores.
Mine hitted 94 Celcius maximum and hwinfo didn't indicated Thermal Throtthling. Let's be frank - none of the current games will probably be able to put the CPU into that condition which Prime95 does, so probably we don't have to worry about. For now of course... As time will pass, dust will come to heating pipes and paste will loose it capabilities, thus we might start to thermal throttling under normal gaming conditions... I'm going to buy some cooling pad definitely - 2-3 Celcius less will be vital.
I also recommend to put the "Balanced" power plan - it will downclock the CPU to 8x when not in use, lowering the overall temperature and increasing lifespan of the cpu and paste. It's very intelligent - if some process needs a core, then it can hit max turbo. Maximum performance keeps max turbo on all cores even if there are not in used - thus increasing the temperature of the CPU.

rmuniak, yes, I use the Balanced Power Plan myself for everything except Benchmark/Stress testing - High Performance generates a lot of heat all the time, and doesn't give the CPU a chance to cool between loads.

The 4700HQ default multi's are 34x,33x,32x,32x cores and 34x cache. With XTU I can bump them up to 36x,35x,34x,34x cores and 36x cache.

Are the values you listed for the 4860HQ XTU max values, or default values?

Gaming sometimes kicks temperatures up over prime95 readings, the multiple threads don't all hit the CPU cores at 100% at the same time to get the temps maxxed out - but prime95 is a good test everyone can download and set up the same to get comparable readings.

Except for a heavy load test like this, I don't think both CPU/GPU will evet be hit this hard, unless your application is doing high CPU/GPU rendering, and only a few user loads I have seen hit harder - tuning down the multiplier on cores 3/4 to 29x-30x - adjusting for highest performance just under thermal throttling point, gets them through those loads as well.

When you find a cooling pad that helps reduce CPU/GPU temperature readings, please let us know :)

rmuniak
12-13-2014, 02:48 PM
hmscott i think they are default values. Since the temperatures are already high (check update of my post above for the test), i'm definitely not plan to modify them and do OC. Which software we can use for setting the maximum multipliers for each core to prevent thermal throthling?

hmscott
12-13-2014, 03:01 PM
...
UPDATE:
44221
Standard fumark with burn in + prime small fft. Even though i got thermal throthling indicator, you can see that multiplier haven't been lowered below the base value. Base clocks also stayed the same. Actual thermal throttling threshold set by Asus might be higher than the one indicated by HWInfo - otherwise we would see effects on minimum clock value or bus clock values. Please correct me if i'm wrong. Cooling pad for the win :)

rmuniak, it is weird, sometimes a particular model of ROG laptop will thermal throttle at lower than expected temperature, and sometimes higher... I wonder if hwinfo is including TDP Power Limit Throttling events in with Thermal Throttling events.

It helps to send the readings to a logfile - having the min, max, average values in a single display is handy, but it doesn't tell you what went on in between those readings, and how often the peaks were sustained - most of the time when you look at the detailed logs, the peaks are very rarely and briefly hit - with the rest of the readings much lower.

You can click the green + mark at the bottom left, 2nd over, icon to start default setting logging or go into settings to set the level / frequency of logging. A .csv file is created - but the default uses semi-colon's instead of comma's so I edit the file with notepad to do a global replace of ;'s to ,'s - makes the columns line up nicer in Excel.

Thanks for posting the new update with thermal throttling - what was your ambient temp? Mine here is 22.2c - is your run today at a higher/lower ambient than yesterdays tests?

hmscott
12-13-2014, 03:16 PM
hmscott i think they are default values. Since the temperatures are already high (check update of my post above for the test), i'm definitely not plan to modify them and do OC. Which software we can use for setting the maximum multipliers for each core to prevent thermal throttling?

rmuniak, actually for most times, setting the multipliers up to maximum values is ok, because we don't load the CPU on all cores at 100%.

As you add threads, which in Windows is most of the time, the cores all drop down in multipliers - it is rare to have a single thread at 36x for an extended period, some other background process kicks in and drops the over all multipliers. You are using Balanced mode, so the CPU is constantly down-clocking.

The program is from Intel, and all it does is unlock the intended maximum core multipliers - what it would have shipped with if everyone had as awesome of a cooling solution as Asus.

Intel Extreme Tuning Utility - XTU has been out for a while now, and is stable. You set the values, Apply them to the current running environment, then Save the new values to a Profile, which gets loaded and run again at the next boot.

If / when you crash Windows, at the next reboot XTU won't apply it's settings - because it doesn't know if the settings set by the active profile caused the crash or not - so after a Windows crash you know wasn't caused by XTU, start it, load the desired profile, and Apply it.

The other cooling thing you can do with XTU is to set a Negative Dynamic CPU Offset voltage and Cache Voltage offset - this reduces the power used and heat generated by the CPU, resulting in lower temps and less fan action at idle.

Every CPU is different so your setting may or may not match mine or others. Between my 3 G750's they all had different stable negative offset voltages. My JW was -125mV, JX was -75mV, my JH was -20mV new and stable now at -45mV - you can make small adjustments over time and the CPU seems to respond favorably by becoming stable at that new lower voltage.

Here is what XTU looks like, for settings, Profiles, and Benchmark, Stress testing:

44231


44232


G750JH max settings, -40mV, benchmark
44233


G750JW, default settings, -100mV, benchmark
44234


Settings for graph readings enable, click the little Wrench icon in each display pane to change pane settings. These are different and in addition to the program settings.
44235


5 minute stress test, simple CPU test for Power Limit Throttling, temperature throttling
44236


Intel Extreme Tuning Utility Download - XTU currently at 5.1.1.25
https://downloadcenter.intel.com/SearchResult.aspx?lang=eng&keyword=xtu

For border line thermal throttling, for long batch jobs, you can make a special profile that uses maximum multipliers for cores 1/2 and reduce the core multi's for cores 3/4 to just under thermal throttle point on extended runs - to maximize throughput/performance and keep the temperature under Thermal Throttle point. As an example, this worked on a G750JS setting to 36x,35x,30x,28x cores and 36x cache for rendering work.

You should be able to set max multipliers for cores/cache for normal operation on the G751, but always test to make sure.

Have fun :)

rmuniak
12-13-2014, 06:28 PM
Thank you very much mister! I must say i underestimated your knowledge in the topic and own you some sincere apologies :) I actually find out that i have this software already installed by Asus.
That option for setting particular maximum multiplier value would be really valuable to prevent thermal throthling.

Isn't messing with voltages dangerous? Will it affect the CPU performance negatively? I understand higher voltage = higher temperature and better performance, lower voltage=lower temperature and a bit worse performance?

hmscott
12-13-2014, 10:06 PM
Thank you very much mister! I must say i underestimated your knowledge in the topic and own you some sincere apologies :) I actually find out that i have this software already installed by Asus.
That option for setting particular maximum multiplier value would be really valuable to prevent thermal throthling.

Isn't messing with voltages dangerous? Will it affect the CPU performance negatively? I understand higher voltage = higher temperature and better performance, lower voltage=lower temperature and a bit worse performance?

rmuniak, the voltage drop when stable is ok. You try in small increments, do a stress test, and if it is stable, drop it a little more, when it finally BSOD's then back off 1/2 your increment and stress test again.

Eventually you will arrive at a stress test stable undervolt and then you reboot and let it sit idle for a few hours and if it doesn't BSOD, you have a stable undervolt. If it does crash during the idle test, back off 1/2 increment and let it idle again.

I usually can't do that much idle testing in a row, so I will save the test profile and go back to a stable or default profile to do work - important, don't have a test setting in play and try to do work.

Asus started including XTU with the G751, we needed to download it from Intel for the G750. What version is Asus shipping? The latest from Intel is 5.1.1.25.

I have been undervolting and overvolting for a long time. All the CPU's are still alive :)

insiZion1
12-15-2014, 10:04 AM
Rmuniak.....i too have been suffering high temperatures peaking at 92 especially with the house heating on! I wish to prolong the life of my g751and done a little searching.
Goto your power settings and on processor state change from 100% to 99%. Immediatley my temperatures fell into the 60s to low 70s. Without any noticeable drop in performance. I then set my fps in evga precision to 60 max. Im now gaming bf4 on ultra with average temperatures mid 50s for cpu. An incredible drop of over 30%s.
Have a play around and let us no.

hmscott
12-15-2014, 11:42 AM
...power settings and on processor state change from 100% to 99%. Immediatley my temperatures fell into the 60s to low 70s. Without any noticeable drop in performance. I then set my fps in evga precision to 60 max. Im now gaming bf4 on ultra with average temperatures mid 50s for cpu. An incredible drop of over 30%s.
Have a play around and let us no.

insiZion1, you don't need to give up that much performance to get the temperatures down. In fact the G751 cooling system will keep the CPU/GPU cool enough even at full performance, and as the temperatures touch Thermal Throttling at highest performance and maximum cooling point - the CPU/GPU will throttle gradually.

There is no need to give up all that performance from 2.5ghz to 3.6ghz (4860HQ). You say you don't notice, but you will :)

Your temperature drop of 30% is about the same as the performance drop, CPU performance. If you starve the GPU for CPU cycles by reducing the CPU performance you will be giving up GPU performance as well.

For some games, that isn't an issue, but the easier way to solve it is to leave the CPU/GPU settings as they are and use the FPS control features of the game and Nvidia to restrict runaway FPS, whatever is 10% above your refresh rate is wasted, some say 220%, either way use the demand reduction tools, not the source generation tools to limit power usage.

But, you have the right idea, just the wrong end of the stick.

Why spend all that money on high performance CPU/GPU and then not use it? Might as well get a 960m to begin with.

If you want to tune the multipliers so as to not limit performance - set cores 3 and 4 just under Thermal Throttle point, and increase cores 1 and 2 to max multipliers, that is a good compromise.

With this fine tuning you are only loosing performance in the situation where you at at 100% utilization for sustained periods of time, but all normal usage happens at full performance + extended multiplier for single/dual threaded demand.

Have fun :)

insiZion1
12-15-2014, 01:23 PM
Thx for pointing that out. Im new to laptops and never had any temperature issues with my built desktops.
Using hwmonitor and seeing such high temperatures was worrying me. I dont want to fry the laptop to quick. Trying to get my head around all the core adjusting is quite mind boggling. This machine runs everything i throw at it with ease even with the adjustment i made with the processor state. Is there a tutorial on tuning the multipliers as it sounds quite difficult. Many thx.

hmscott
12-15-2014, 02:38 PM
Thx for pointing that out. Im new to laptops and never had any temperature issues with my built desktops.
Using hwmonitor and seeing such high temperatures was worrying me. I dont want to fry the laptop to quick. Trying to get my head around all the core adjusting is quite mind boggling. This machine runs everything i throw at it with ease even with the adjustment i made with the processor state. Is there a tutorial on tuning the multipliers as it sounds quite difficult. Many thx.

insiZion1, it's really not tough, check out the images and text in post #65 above, and use common sense. You've already done plenty of the same kind of stuff with hwinfo and the Power Plan CPU Min/Max, use the sliders and look at the options - there aren't many. The CPU sliders can all go to Max, same with CPU cache.

Then read post #67 on setting a negative voltage offset for the CPU/CPU Cache to reduce power / heat without losing performance.

There are lots of posts in the forum about settings, testing, results, and please post your tuning results too :)

insiZion1
12-15-2014, 05:53 PM
Confused regarding setting cores 3 and 4 under thermal throttling point. When im stressing and benchmarking throttling remains on 0%.
With all the cores maxed my cpu temp is around 78-80 is this exceptable? Im now using msi and rivatuner and have set 60fps limit. My gpu is maxing at 62.
When lowering the cpu voltage offsrt..will tbis reduce cpu temps much? What would ypu suggest as a starting point? Maybe half -50 mv ? Is this likely to balls anytbing up thx.

hmscott
12-15-2014, 07:01 PM
Confused regarding setting cores 3 and 4 under thermal throttling point. When im stressing and benchmarking throttling remains on 0%.
With all the cores maxed my cpu temp is around 78-80 is this exceptable? Im now using msi and rivatuner and have set 60fps limit. My gpu is maxing at 62.
When lowering the cpu voltage offsrt..will tbis reduce cpu temps much? What would ypu suggest as a starting point? Maybe half -50 mv ? Is this likely to balls anytbing up thx.

insiZion1, you don't need to decrease the multiplier on cores 3 and 4 unless your task is heating up the CPU past the Thermal Throttling point, which is about 92c/93c - duration is important too, peaks into 94c/95c can occur without Thermal Throttling - hwinfo has those 4 lines of Thermal Throttling info and turning on logging to find the values between the peaks can help too.

But, if you aren't Thermal Throttling, then there is nothing to worry about, and you can leave all 4 cores and cache at max multipliers.

The voltage offset is different for all CPU's, I would start at -20mV, and work up in -10mV increments until you BSOD, then back off -5mV and try again. When you have a good run on stress testing with a new "high" negative offset then let it sit idle for a while - the last stability test is during idle. As above, my G750JW did -125mv stable, JX -75mV stable, and my JH did -20mV to start, and is not running at -45mV after months of slight increases - it still BSOD's at -50mV.

It sounds like you have it under control, please post your final stable results - and which CPU you have, I have a 4700HQ.

rmuniak
12-16-2014, 10:04 AM
Confused regarding setting cores 3 and 4 under thermal throttling point. When im stressing and benchmarking throttling remains on 0%.
With all the cores maxed my cpu temp is around 78-80 is this exceptable? Im now using msi and rivatuner and have set 60fps limit. My gpu is maxing at 62.
When lowering the cpu voltage offsrt..will tbis reduce cpu temps much? What would ypu suggest as a starting point? Maybe half -50 mv ? Is this likely to balls anytbing up thx.

You're setting fps limit which reduces the load on the gpu - thus you get less heat. This is not the proper stress test. I'm not sure which benchmark you're using.

Btw have you noticed much more bluring when setting V-Sync in Nvidia Control Panel?

insiZion1
12-16-2014, 10:41 AM
When i use the vsync i did notice some blurring on the middle of the screen in bf4. But setting the fps using msi and rivatuner i can leave vsync off and i get no blur.
Ive not really stressed with any program other than xtu last night, previous to me setting fps.. hwmonitor would show my temperatures ingame.
Regarding xtu ... the results seem to fluctuate. I can run a bench and get 715 , then run the same test again and the result can be 655:s

rmuniak
12-16-2014, 12:10 PM
Blur at the middle of the screen? I'm seeing far more visible motion blur - it all over the screen. Altough motion blur is inherent to LCD technology, it was far more visible with this VSync option. Is this what you talked about?

Thanks for the tip with FPS limit, need to try it out.

hmscott
12-16-2014, 01:40 PM
When i use the vsync i did notice some blurring on the middle of the screen in bf4. But setting the fps using msi and rivatuner i can leave vsync off and i get no blur.
Ive not really stressed with any program other than xtu last night, previous to me setting fps.. hwmonitor would show my temperatures ingame.
Regarding xtu ... the results seem to fluctuate. I can run a bench and get 715 , then run the same test again and the result can be 655:s

insiZion1, do you have AIDA64 installed? Sometimes that can interfere with XTU, but dont't worry about the benchmark, it has a weird 2 step result, not sure how it does the test runs, but if you look at the load curves you will see peaks and valleys, intersecting at different times, resulting in non-deterministic results. Some times I can get consistant runs, and some times not - so I stopped using the benchmark.

It depends on the app if their / Nvidia's vsync enable causes tearing, it would be nice if the new Nvidia Battery Boost could be used on AC power too, that would be another method of controling FPS.

You can install Nvidia Inspector to set custom settings for applications, like limit FPS (try different settings to avoid tearing) or limit by temperature.

Nvidia Inspector - FPS Limiter Tutorial
https://www.youtube.com/watch?v=pH9D__g1yqs

How To: Create Custom FPS Limits in Inspector
http://www.overclock.net/t/1368777/how-to-create-custom-fps-limits-in-inspector

Nvidia Inspector homepage, author site (German, not updated, don't download from here)
http://orbmu2k.de/tools/nvidia-inspector-tool

Guide:NVIDIA Inspector Wiki Updated Dec/2014
http://wiki.step-project.com/Guide:NVIDIA_Inspector

Nvidia inspector does lots of useful stuff, including forcing applications to run on the Nvidia Processor for Optimus laptops - not something a G751 user needs to worry about :)

Have fun!

insiZion1
12-16-2014, 04:38 PM
No aida64 installed.

Couple of things if you please...when setting cpu voltage offset is it right to set thecache voltage offset to the same value?
Im now at -50mv on both and passed the stress test. May i ask as to what benefit i gain in having it at -50 instead of the default.
Thx again

insiZion1
12-16-2014, 04:54 PM
Should I try and lower the voltage more? now at -60mV

rmuniak
12-16-2014, 05:36 PM
insiZion1 less voltage=less heat in stress

Just run Prime95 small FFT torture test with standard voltage and note the temperatures on the cores (average and maximum using hwinfo for example) after 10 mins. Then apply e.g. -50 mv, run Prime again to see if system is table watch the temperatures - they should be a bit lower! Every deegree less counts and increases lifespan of your computer :)

rmuniak
12-16-2014, 05:42 PM
rmuniak, the voltage drop when stable is ok. You try in small increments, do a stress test, and if it is stable, drop it a little more, when it finally BSOD's then back off 1/2 your increment and stress test again.

Eventually you will arrive at a stress test stable undervolt and then you reboot and let it sit idle for a few hours and if it doesn't BSOD, you have a stable undervolt. If it does crash during the idle test, back off 1/2 increment and let it idle again.

I usually can't do that much idle testing in a row, so I will save the test profile and go back to a stable or default profile to do work - important, don't have a test setting in play and try to do work.

Asus started including XTU with the G751, we needed to download it from Intel for the G750. What version is Asus shipping? The latest from Intel is 5.1.1.25.

I have been undervolting and overvolting for a long time. All the CPU's are still alive :)

hmscott, can i ask a small favor? Can you explain step by step what should be safe procedure to follow in XTU when undervolting? I know that when you save the profile and select it will load again with the next run of the computer - so i understood that when i save and load the nonstable profile and my computer will crash i will be not able to revert it? Is it true? I'm still not sure how this rebooting when not stable/profile thing works...
I would be very thankful if you could explain further on.

hmscott
12-17-2014, 02:11 PM
hmscott, can i ask a small favor? Can you explain step by step what should be safe procedure to follow in XTU when undervolting? I know that when you save the profile and select it will load again with the next run of the computer - so i understood that when i save and load the nonstable profile and my computer will crash i will be not able to revert it? Is it true? I'm still not sure how this rebooting when not stable/profile thing works...
I would be very thankful if you could explain further on.

rmuniak,

1) Start XTU, set negative offset, start with -20mV for both CPU/Cache. Apply. Save to a Profile name. Exit XTU.

2) Run a stress test like prime95 or some other CPU only stress tool.

3) You don't crash. Run with it at this setting for a while, leave the laptop on when you aren't using it to see if it crashes (and reboots). The idle test is the last to fail, so don't think you are stable until the laptop can sit idle for a couple of hours. Increase by -10mV, test again, increase by -10mV, continue till you crash.

4) You crash. When you reboot XTU will *NOT* load the last profile used, it will allow normal boot up with no changes to the CPU.

5) Reduce the setting of negative offset by half of what you increased it by. If you increased by -10mV then reduce to -5mV increment.

6) Keep repeating until you have a value that is stable.

Make small changes and test, then when you are sure it is stable, increase and try again. Stop and revert back when you crash.

What value are you at now?

rmuniak
12-18-2014, 09:24 PM
I haven't tested yet, since i'm having problems with XTU. It crashes with "perftune.exe has stopped working" message when i click on settings... I tried to reinstall with the latest version but it stays the same. I have no clue what to do to make it work as intended...
So frustrated because of this at the moment. No clue and zero help over the internet about this...

hmscott
12-18-2014, 09:59 PM
I haven't tested yet, since i'm having problems with XTU. It crashes with "perftune.exe has stopped working" message when i click on settings... I tried to reinstall with the latest version but it stays the same. I have no clue what to do to make it work as intended...
So frustrated because of this at the moment. No clue and zero help over the internet about this...

rmuniak, it looks rare, not many hits and none from the version 5.x series, you are running 5.1.1.25?

Ahhh, try uninstalling, and then install this time As Administrator - right click and Run As Administrator. :)

insiZion1
12-18-2014, 10:20 PM
Im at -70mV now and it seems to be fine.. been idle for a couple of hours with a long stress test with prime. Did try -90 which is the max, it passed the stress tests but crashed when playing bf4.
Earlier I ran both prime and furmark on default settings together. Temps went to 94 and started throttling on core 2.
Hmscott, you said something about changing the values on core 3 and 4 if you get throttling. Anychance you could enlighten me a bit more as im a bit lost. Especially as it was core 2 that throttled.....the gpu and cpu have mostly been fine in game, but after I stressed with both programs today, I let it rest for a bit then played bf4 and throttled for the first time:/ maybe it was to soon after a big test and there was still a lot of heat.

rmuniak
12-18-2014, 10:32 PM
rmuniak, it looks rare, not many hits and none from the version 5.x series, you are running 5.1.1.25?

Ahhh, try uninstalling, and then install this time As Administrator - right click and Run As Administrator. :)

Yes, latest one... Tried your suggestion but it didn't help. Installed all the WIndows updates and it's the same.

Otherwise the program is running okay... What can i change in those settings? Is it safe to run the program without possibility to open that settings window? I'm already testing -40

UPDATE: -50 and still throttling at prime+fumark - where is the temperature drop?
UPDATE2:-70 and throttling on prime+fumark

Btw i reloaded the XTU with defualt profile, it rebooted the computer and then it said that "Processor Current Limit" can't be changed. It stays at 90 000 A, while default value is 85 000 A. Can you please check what value do you have guys there? Why i can't revert the change????

hmscott
12-19-2014, 04:06 AM
Yes, latest one... Tried your suggestion but it didn't help. Installed all the WIndows updates and it's the same.

Otherwise the program is running okay... What can i change in those settings? Is it safe to run the program without possibility to open that settings window? I'm already testing -40

UPDATE: -50 and still throttling at prime+fumark - where is the temperature drop?
UPDATE2:-70 and throttling on prime+fumark

Btw i reloaded the XTU with defualt profile, it rebooted the computer and then it said that "Processor Current Limit" can't be changed. It stays at 90 000 A, while default value is 85 000 A. Can you please check what value do you have guys there? Why i can't revert the change????

rmuniak, somehow your settings / profile got corrupted. I would uninstall again, and reboot, then delete the associated folders to start from scratch.

This is the settings folder, delete it.
C:\ProgramData\Intel\Intel Extreme Tuning Utility

Delete this folder if it still exists after uninstalling
C:\Program Files (x86)\Intel\Extreme Tuning Utility

And although there are Intel folders under AppData folders, there doesn't seem to be any directly related to XTU, so don't delete them (Intel stores other things there).

If you have ccleaner, Norton Utilities, or some other Registry cleaner, you can run that as well before rebooting and trying to install XTU again.

Install XTU by right clicking on the XTU installer and select Run As Administrator to do the install.

It is rare but this does happen, you aren't the first to report something like this, strangely it hasn't happened to me...

Before you re-install, try doing a disk check on your C drive - right click on the C drive icon, select Properties, and go to the Tools tab == Error Checking and if it wants to schedule a disk check at next boot let it and reboot to make it happen. If it says no need to run it, continue and run it anyway - it will complete without needed to reboot.

That should get you back on track :)

rmuniak
12-19-2014, 09:40 PM
That didn't help. BTW, i noticed that when switching from XTU to pulpit and going back to XTU - when screen is transitioning i can see some artifacts for a while (like a blue, white and black bold vertical bars) - it just a few ms, but it's visible. I was afraid that it's due to undervoltng, but after reseting to standard the same effect persists! Did anyone noticed something similar (just switch quickly between XTU and pulpit)? Should i start worry? I think i will drop this undervolting fun - getting ****pants when noticing something unusual.

ExpatGamer
12-23-2014, 05:37 PM
No pics but I've got similar results stress testing with FurMark and Prime95 small FFT:
CPU temps peaking at 95C but mostly hitting 93C
Thermal Throttling on cores #1 and #2

From everyone else's results, it doesn't seem like there's much point in returning it. Looks like its just a consequence of the G751's cooling design. Maybe I'll RMA it near the end of the warranty period, especially if people report that Asus has fixed the problem. I thought they didn't ship models that throttled.

One thing I'm not clear on is what's actually happening when the throttling starts. I've got a 4710HQ and, according to HWiNFO, the multiplier (x25) and voltage (~.95) are staying constant under this load even with throttling on one or more cores. So, as far as I can tell, I'm not really losing any performance when the throttling kicks in. Its just not activating the turbo boost, which I thought it wasn't supposed to under these kinds of sustained heavy loads. So what's it doing? Some subtle under-volting or something HWiNFO just doesn't detect?

kanecvr
12-26-2014, 09:14 PM
My machine was hitting 90-95 with 1.5v DDR3. I installed 1.35V memory and the overheating issue went away - I get 88-89 tops. Did you replace the memory in your machine?

ExpatGamer
12-27-2014, 03:54 AM
No new memory or overclocking, all stock settings.

Just did another test under same conditions except with laptop positioned farther from the wall. On my last test, the vents only had 1-1/2 feet clearance from the wall. That made a big difference, the CPU temp avg 89.5C. Still, the ambient temp here is about 21C but where live its about 30C so I'm not celebrating yet. I really need to test it out with a higher ambient temp.

One thing that I think is really clear from my tests is that the fan profile should be more aggressive. In my last test while the temp was hovering between 89-90C, my fan was running 3000-3100 RPM. I'm not sure what the max RPM is but I've seen it run 3500 RPM in AIDA64. My guess is that its set to target 90C max.

I have an inquiry in with Asus about the fan profile. I'll let everyone know if I learn anything.

Nobody's discovered any way to control the fans, right?

NitroX
12-27-2014, 08:01 AM
@ExpatGamer: This is also one thing that I stated in another post. I like that the fans are quiet in idle but when I'm gaming and stressing the hardware, I just want it to stay cool. The vents on the older models (G75VW) would reach 4000RPM when temps went over 85C.The problem is that the heat is not exhausted fast enough and it stores inside the case.
We're waiting for your response when you get an answer from ASUS. A new driver should do the trick in maximizing the vents rotations.

Qosmio
12-27-2014, 10:20 PM
Hello All! I have a ASUS G751JY-T7015D Notebook with Intel Core i7-4710HQ . I had make a benchmark with the Intel extreme tuning untility. I got 463 score. I thing something wrong with my laptop CPU. Because it is too low. I saw here 710score from the same CPU. Is it normal?
Thanks.

saiyadjin
02-11-2016, 11:21 AM
guys do you use win10 or some other windows?

warryabel
02-11-2016, 11:28 AM
Hello All! I have a ASUS G751JY-T7015D Notebook with Intel Core i7-4710HQ . I had make a benchmark with the Intel extreme tuning untility. I got 463 score. I thing something wrong with my laptop CPU. Because it is too low. I saw here 710score from the same CPU. Is it normal?
Thanks. you get much more with dualchannel ram

warryabel
02-11-2016, 11:31 AM
No new memory or overclocking, all stock settings.

Just did another test under same conditions except with laptop positioned farther from the wall. On my last test, the vents only had 1-1/2 feet clearance from the wall. That made a big difference, the CPU temp avg 89.5C. Still, the ambient temp here is about 21C but where live its about 30C so I'm not celebrating yet. I really need to test it out with a higher ambient temp.

One thing that I think is really clear from my tests is that the fan profile should be more aggressive. In my last test while the temp was hovering between 89-90C, my fan was running 3000-3100 RPM. I'm not sure what the max RPM is but I've seen it run 3500 RPM in AIDA64. My guess is that its set to target 90C max.

I have an inquiry in with Asus about the fan profile. I'll let everyone know if I learn anything.

Nobody's discovered any way to control the fans, right?
normally nbfc should work for fans.... the thing is discovered was after rma and update to new bios 211 the ec work not correct!
my fans are blowing whole time 100% i must use nbfc to fix this issue
g751 jt

h3ll4ngel
09-20-2016, 04:31 AM
My laptop is G551jw, can I use a stable profile (-30vm) to play game and render for a long time using ? I mean a normal use
i see a graphic option so Can i undervolt graphic ?

JustinThyme
09-20-2016, 12:10 PM
Generally people do not undervolt GPUs especially if you are trying to OC as it WILL spell certain catastrophe!

To OC a GPU it takes more voltage, not less. Please keep this question to one place instead of grave digging.



59513

St_METAL
09-24-2017, 01:29 PM
http://voltground.com/haven/threads/69/