*OFFICIAL* Alienware 17 R5 Owner's Lounge

75C is quite good on normal paste. I bet if you increase the fan speed the temps will drop by 10.

 

Sorry is there a way to change the GPU frequency? I do not see any such option even in Alienware FUSION ?

Vasudev if you plan to overclock you NEED to set the FAN performance to at least performance mode. But without overclock you can even keep it to balanced or even quiet mode. With oc2 and CPU speeds of 4.8GHz maximum (I have not seen it going over that) maximum temps are about 93 degrees maximum but with FAN performance mode.

I currently have no overclocking (CPU reaches 4.7GHz), fan to quiet and maximum temps of 77. Will test after gaming too to see if I need to change FAN performance to balanced.

 

Use MSI Afterburner or nvidia inspector to overclock the gpu. For GPU undervolting use MSI AB.

 

Thank you Vasudev! I did not have any idea that I can do that...

Is there anything I can use to downclock the CPU? I mean limit the maximum GHz? I currently have CPU performance disabled on BIOS and no overclock profile selected in Alien FUSION but somehow CPU still reaches 4.7GHz!

 

How did you cannot disable the OC? As far as I can remember you can disable the Speedstep from the BIOS, in which case you will be running at 2.9GHz ***BUT*** with the GTX 1080 at 150W-180W, the CPU almost cooks, and throttles, and drops well below the 2.9GHz, so you cannot even maintain base clocks. Unless mine was truly broken and yours is not?

 

CPU performance mode must be used in Summer for effective reduction in thermals by 5-10C.
4.7GHz is actually great and do note high MHz don't affect temps too much.
You need to optimise W10 services using phoenix's guide.

 

Doofus can you tell me where this speedstep is in bios? Do you mean the CPU performance mode? It is disabled in BIOS. And when it is disabled in BIOS, you also get a windows notification from Alienware command center that oc options have been reset to default. And how do you check that GTX 1080 reaches 150W-180W? But I have not seen the CPU running at 2.7GHz, minimum is usually about 3.6GHz (using HWINFO)

Vasudev can you give me a link for the phoenix's guide? Another thing I noticed is that BIOS CPU performance OC profiles do not really change the CPU maximum speed (OC2 does not change maximum speed to 4.7GHz unlike the OC2 profile in Fusion). Only cause my fan to spin all the time. So, what does the BIOS CPU performance actually do? And why do we still use the FUSION? If an overclock profile is enabled in FUSION, this automatically disables the BIOS CPU performance. This is getting highly complicated.

Regarding the temps, I may have to say I have the laptop always on a cryo NZXT (only passive, did not enable the base fans).

 

Its in Advanced Section in BIOS. Don't disable EIST aka Speedstep in BIOS unless you want cpu to be locked at 800MHz. Use TS to disable it at OS level since Dell BIOS don't have HWP state in their BIOS Menu.

 

This is the MAJOR problem in this picture here:

One game running nothing much else
CPU % around 20%
GPU at 180W and steady 71C
Fans SCREAMING loud
MasterCooler U3 Plus Base with 3 positioned fans on full max
CPU temps to 100C
CPU at 45W
CPU THROTTLING non stop
PCH at 92C and RISING
SSD also at 92C (not shown in this picture).

So we have the GPU at 180W cruising at 71C, albeit with loud fans and Coolermaster support with raised base and 3 extra fans.
The CPU and PCH and SSD are all COOKING with the CPU throttling like crazy dropping to 8x (forget the 29x base).

Can you please run same test and tell me how your laptop fares?

 

Not true, it drops you to 2900MHz which is "base". It is thermal throttling (and other throttles) that drop you below 29x. You cannot tell the BIOS to drop you below 29x.

 

On my AW it stayed at 800MHz.

 

Did you lap the heatsink? It can shave off some 5C or more. Turn ON perf mode in BIOS. Use TS and uninstall XTU/OC Controls.

 

Doofus99 we can test by running the same benchmark with same settings.
Do you want to try Unigine Superposition? Post your screenshot. Tell me the settings you choose and I will run with the same and post my screenshot.

 

I did not open the laptop as I was planning to return it, or at least not invalidate the warranty!

BIOS was in perf mode and fans were quite on loud (another very annoying feature).

I did not use any OC software, and I did do a proper uninstall of XTU to see if it were the culprit, but no.

Sorry, mine has been returned and the money has hit my account today! I am free to buy another one now. Not sure if I should risk it...


I believe that mine was more problematic than most. The PCH needs a heatsink of its own, and so does the SSD. The keyboard also reaches 52C and am not sure what cooler you'd need there, I mean how do you fix that.

If you have the 17 R4 or R5 look at that: fans full on, CPU at 100C, GPU at 70C, check your vents. You have a left vent REAR with CPU hot air, LOW/MEDIUM airflow. But you also have left SIDE air ent with STRONG COOL airflow !!!! I cannot possibly explain this. Can anyone please verify it is also happening to them?
​

 

Exactly same Cpu overheating problem on refreshed Alienware 15R4 Coffeee lake models as well. See after re-paste...

 

Is that with GPU on ?

 

Yees.

 

May I suggest you examine the levels/heights of the GPU/CPU and whether the heatsinks make perfect contact without force? maybe use some carbon paper?

What I mean is it would be great result if the heatsinks which are not produced with the mechanical tolerances and accuracy that the motherboard is, if they were misaligned! Because it would mean that we can repair it...

 

Fyi. I don't own any of the newer Alienware's. Take a look over here. Post your suggestions there.

 

I have run some tests that we can all test on R5 and compare our temps. I run Dell 15min stress test and extended system scan. I have fan mode on balanced!

Here are the temps with OC2 enabled in Fusion (CPU reached 4.9GHz!). I run Dell 15min stress test and extended system scan

https://imgur.com/a/dlZmE07

And here are the temps with OC profiles disabled in fusion (CPU reached 4.2GHz). I run Dell 15min stress test

https://imgur.com/a/1gNFI8e

Please tell me your impressions. As you can see, CPU reached 99 degrees maximum.

 

Nevermind, lol .

 

Its too high. You can turn on CPU performance mode with aggressive fans.

 

Awful temps. And this is done with the wimpy week Dell stress test. Your PCH chips is near boiling 100C and ssd's close to 80C without much stress. And I won't talk about the awful Cpu temp results, they can speak by themself. After acouple of weeks or a month or two dust will come visiting inside the chassis, and this will be even worse...

 

I agree. Very hot and I suppose the fans were screaming loud during the test? Look at your SSDs they are also hot, did you not say you had fitted heatsinks on to them?

 

SSDs were reaching 99 degrees without heatsink. And Dell extended scan (which I also run as I said) stresses every part of the laptop including the drives too.

And Dell stress test is not weak. In fact, I run another benchmark and temps were much lower. And with Dell stress test CPU reached 4.9GHz and maximum temps 99. I think the threshold is 100. At least it did not reach temps over 99 as I have seen others reporting. What temps would be good for you at 4.9GHz? And what is the way to reach such low temp? These are the temps during Dell stress test. They drop in secs after benchmark. Temps are usually at 64degrees and fans so quite I can hardly hear them.

https://imgur.com/a/VhKHkcZ

 

Dells 4.9GHz is on single core. And your test show wimpy 56W Cpu Package Power. Could you put up and run Aida64 + Hwinfo as the guide. Use High Prformance in Windows power plan (I want to see clocks never drop vs max clocks). Take screenshoot of test while it run (after 15 min)Thanks.

For SSD testing, run 3 or 4 rounds ATTO Disk Benchmark in a row. You can as well test it while you run Aida64 Stress test as the guide.

 

@Aristotelhs2060 Dell Stress test barely stresses the HW at all, you see more heat because the HSF is unified and fans respond very late sometimes.
Uninstall most Dell apps and use trial version of AIDA64 Stress test with CPU, FPU, GPU to get real world 100% cpu/gpu unified load balanced test.

 

On the Aida64 stress test I have noticed that to pull the most amps from the CPU you have to select FPU only, do not mix it with CPU. And I have also noticed that "GPU" does both VGAs (the Intel and the external one) there is no way to separate them.

 

I am waiting for a Dell technician to come and change the thermal paste and heatsink. He is going to use the conductonut thermal paste which I have bought..

Do you guys know what is going on with the April 2018 update now? It now says not compatible with your hardware. It previously said about graphics but now says not compatible hardware contact your manufacturer.

 

Dell offered me to change heatsinks too, and I declined, because I could not understand how it would make a difference with the machine being so brand new design. Now I think I understand, they must have had a batch of badly made heatsinks or they must have developed a new heatsink design.

 

I wouldn’t trust Alienware’s outsourced tech team use Conductonaut on my machines. Will they do pressure testing and do it correctly? Follow as a hawk (all steps). If You get any leakage and frying some components, this won’t be covered by Dell’s warranty although it was the Dell tech witch put it on...
Edit. Read also http://forum.notebookreview.com/thr...4-owners-lounge.797457/page-818#post-10729381

Regarding Alienware and compatibility with Win 10 1803 April Joke Update. Read this thread... http://forum.notebookreview.com/threads/windows-10-spring-update-buglist.816324/page-6#post-10728721

 

I would be very cautious of using Grizzly Conductonaut unless the technician really knows how to do it, and how to seal it so it doesn't leak. If done right, it is fine, but if not, it could leak and kill your motherboard.

Safer would be a Non-Liquid Metal...something like Gelid GC Extreme or other quality paste.

 

FWIW, after bashing the Alienware for hot PCH, I have run hwinfo on my Clevo and the PCH is up in the 80C-85C when I play a game. The CPU and GPU are cool at 69C and 66C therefore the PCH is heating up all on its own. Both CPU and GPU have separate fans and separate heatsinks straight out the back. Not sure where the PCH is located.

 

The Dell/3rd party technician just left. He did not even know how to use the Grizzly Conductonaut. In addition to wasting it (again I bought it) he used another thermal paste which I bought (kryonaut thermal grease). Regarding the temps I will have to test now but they dont seem much better from a first look. They told me to run tests again and feedback.


Just one question for now. If I see spikes (going to 99 degrees but the average or current temps are about 20 degrees lower) is that ok? I am asking because the CPU seems to go from 797Mhz to 4.7GHz so understanding the acceptable temperatures is confusing. It is not like the older CPUs which had lower frequency change range.


At least I have seen how he disassembled the laptop. I may have put thermal paste myself as I used to do in the past for my old Alienware M17X R2 which is still working flawlessly without any heating issues for almost 9 years

 

Important questions:

Did he replace the heatsink or any other components? If yes, what exactly did he replace ?

Did he make sure that the CPU and GPU are making good contact without paste? What if the heatsink is bent or out of true? Then it will not sit properly on the CPU/GPU and you will still be in trouble.

Why did he leave without first making sure there was a change to the temps?

If you had not provided him with the better quality thermal grease, what was he supposed to use?

 

Doofus99:

He replaced the whole heatsink. It came as a big part with the thermal pads.
I do not know if he made sure about good contact but he put all the screws and seemed to do it carefully. It will be seen if it is better.
We checked some temps before he goes and reported back to them. He told them that I will run more tests and they will call me back. The Dell technician who issued the service asked me to do that anyways after the service.
I do not know about the last question and in fact it seemed really bad to me that they told me to buy the high performance thermal paste. In fact I bought them before they tell me just in case they would be needed (seems like I knew it)

 

Be happy he didn't used the Conductonaut Liquid metal.

99C ain't good nomatter how you see it. Run tests I suggested in post #426. Follow the guide. Remember test with High Performance power plan so all core run max clocks. None of the cores should drop below fixed Turbo boost with correct applied power limits. Post screenshoots.

 

The past is already pasted on to the heatsink so when the heatsink is connected,so is the paste.


Sent from my iPhone using Tapatalk

 

Yes, he mentioned the pre-installed thermal pads.

​

One last question:

Did you think to ask him WHY he was replacing the heatsink? What was wrong with the "old" one?!

 

in addition to that - did it look the same as the old one?

If there is a different batch/part that may be a indicator of a bad batch as you previously mentioned however looking at other laptops with the 8th gen i9 / i7 they do seem to run hot and have momentary temp spikes of 95+. I just couldnt handle 101 as my peak !

 

I did not ask why he replaced the heatsink but I would know the answer. Something like testing if something is broken or not good contact or something. Usual things that Dell techs say. The heatsink was the same. I did not check the part numbers because he came with the new heatsink and left with the old one.

Dasachmo yes those over 90 degrees temps are spikes. In fact, they last for secs or sub secs. Even though you can see such temps even without doing anything, the fans are not even spinning faster (probably because there is no time for the fans to respond).

So my question comes back again. Spikes up to 99 degrees for 1 sec or less. What harm can they cause to the CPU? And if the heatsink/thermal paste is the problem why temps go down by 20 degrees in secs? This is a wholly confusing.

 

For such a short time it is no problem at all. Prologned exposure to those temperatures are though.

 

There are two possibilities.

Possibility 1
The heatsink is fitted badly, it does not make good contact with the CPU surface, you are seeing spikes but also higher consistent temperatures. In addition during some of these spikes the CPU throttles down to 800MHz...

Possibility 2
The heatsink is fitted perfectly and you still see spikes. It means that the CPU sensors do not record average heat over seconds, but rather instantaneous spikes. The heat has no time to travel to the surface of the IC and get picked up by the heatsink. It is just a spike and imagine a perfect sinusoidal (which this isn't, but close enough, I have the charts to prove it) the peaks are sqrt(2) higher than the mean, and it is the mean we care about mostly. Are these peaks dangerous above the max junction temperature? The answer is does the CPU throttle down to 800MHz? If yes then we have a problem and it is Intel related. Nothing to do with Dell !

https://communities.intel.com/thread/119878

https://www.eteknix.com/intel-core-i7-7700-suffering-high-temperature-spikes/

​

 

Under stress with the CPU at > 90C, do you still have cool/cold air coming forcefully out of the left side vent?

 

Doofus99. Thanks for your reply. The fact is the more opinions I get from people, the more confused I become. I am not sure how those new CPUs work and regarding all those opinions I got, I am not even sure if someone is such a specialist to know.

When CPU temps go over 90 and they are not spikes but more constant temps (eg. during benchmarks), then the fan are spinning faster yes. And yes more heat comes out as expected.

Regarding the CPU throttle, I am not sure if the low frequencies that we occasionally see (around 790Mhz) are a result of throttle or just the CPU going down because even without doing anything special, frequencies go from 790Mhz to 4.7GHz (no overclock).

And regarding what dasachmo said earlier, those spikes are CPU related and not Alienware/Dell related. Looks like this is just the way new CPUs work.

 

Like some people previously told you, if your CPU is at 4.7GHz with no overclock, it mean that you're only stressing a single core.

You need to use something like AIDA64 or even XTU.

You're confused about what exactly?

 

Ok, I run the first benchmark after applying the kryonaut thermal grease. Something has changed and it looks like it is to the better. I run the prime95 test with the in-place large fft options selected (this is what the Dell tech who issued the service told me to run before and after the heatsink/thermal paste replacement) which was stressing my CPU to the maximum. In fact this benchmark stresses only the CPU. CPU was previously reaching 103 degrees. It was the only benchmark that managed to reach that high temps. Now it reached 95 degrees maximum!

And fans went down so fast, in secs after stopping the benchmark. And now stays so quiet! Idle temps about 37 and lower degrees. I have fan performance mode on BIOS, no overclock and thermal performance on alienware command center. Windows on balanced power.

https://imgur.com/a/L4wtx8I

 

osChakal, as you can see all the cores reached 4.7 GHz. And I run the benchmark with exactly the same settings before and after applying the thermal paste/new heatsink


Which AIDA64 benchmark do you suggest to run? There are many of them

 

Can someone tell me the following:

1) which benchmark to run (if more than one options within the program which one to run?)
2) settings in BIOS to select:
- a)fan performance setting?
- b)CPU performance setting?
3) alienware command center settings:
- a) overclock profile?
- b) thermal plan?
4) windows power plan?

Thanks

 

Already posted Aida64 in previous post. FYI. If you run P95 with 4.7GHz all 6 cores you would see thermal shutdown or full throttling within a minute.

Said it before. Use High Performance power plan and none of the cores should drop below max turbo all 6 cores. Easiest way to see if you had throttling.