*OFFICIAL* Alienware 17 R5 Owner's Lounge

Far Cry 5 avg temp 70C with 37x. Can go a bit higher but then the temps will also go higher.

 

I get average 70C with 3.7GHz.

Obviously your repasting has made a difference. Is it enough difference to void the warranty though? What if we used a simpler thermal grease like Arctic MX-4 ?

Did you also have to replace all the thermal pads? Did you try to see if the heatsink balances OK with all the thermal pads in place?

 

i see 75 degree with 20w... Wheres your 42w? And compare at the same clocks or same wattage... everything else is senseless. And use hwinfo. Then we see exact values.

with LM we can get 1ghz more allcore clockspeed compared to Stock with same temps ....

Ans where is your proof that it will void warranty?

 

I modify and tweak undervoltage + clock speed to get the appropriate wattage on the CPU, running WoW in the background at 60 FPS and Prime 95 with 2-6 threads on the small FFT torture test. WoW on its own does not use a lot of CPU and Prime 95 is more stable so the graphs are easier to read. It does not really matter how I do it as long as I get a stable wattage for sufficient time to see if it's rising and to be able to take a reasonable average.​

These are my watts / average temps

WoW + Prime95 + some browsers etc

room 26C now, typical 24C (hot day today)

CPU power, CPU avg temp
25W, 68C
30W, 72C
35W, 78C
41W, 85C
45W, 90C, moderate thermal throttling

temps are +/- 8C on the higher settings

I am recording all of this so when I repaste I will be able to compare.

 

I said it will void the warranty because I am under the impression that it discolours and eats into the heatsink and CPU die so it is visible? In addition you will need to re-open it and clean it all up if you want to send it in for service to try and eliminate the LM showing plus all the sealing tapes? I am thinking if we use standard grease it will not stand out as much.

 

i told dell tech support i had mine LM'd, he said 'you will have to do it again after we replace the mobo', i said 'iv got sum left, he said 'give it to the tech guy who is doing the work he can do it for you' haha xD

 

look at the pictures of the stock paste i posted a page or two back, the stock paste is oxidising the heatsink!!!!

i sent the pics to tech support as proof of the sh@t they are using.

 

So I have gotten my brand new replacement Alienware 17 R5 with an i7-8750H, 1070 and 1440p 120Hz display. As I predicted, Dell still used the bad stock paste. So, I toredown the system and did a repaste on the same day I got my laptop. I just used Thermal Grizzly Kyronaut; didn't use LM due to warranty issues and my bad experiences with it.

After about 3-4 days of using this machine, pretty safe to say, this is a pretty stable unit.

My core differential from hottest to coldest gotten only to a max of 5C and not occasionally and I didn't do anything such as bending the tension arm and/or replacing the thermal pads.

Another thing I saw was, Dell used a different heatsink in this unit. Instead of the R4's CCI or SUNON heatsink, I got a heatsink from a manufacturer named "ROBIN" (whut? haha). Doesn't matter, but it seems they used a better heatsink, because the heatpipes are fatter, similar to the 1080 R4 heatsink.

I did a full furmark CPU and GPU stress test with CPU -100mv (reducing unnecessary heat) with my laptop sitting on top of a Cooler Master U3 (USB Fans, didn't do any modding) in a room at 25 C and with fans on full speed mode.

CPU: 76C @ 3.9Ghz on all cores
GPU: 65C

While gaming and fans in balanced mode.
CPU:60-65C
GPU: 55C - 60C

Does anyone here have similar results? Forwarded this to Alienware and they told me it is below average and impressive.

So anyone to Dell and Alienware, PLEASE STOP USING THE STOCK TOOTHPASTE, ITS BAD!!!!

 

It could be territory specific. Last time I looked, in the UK they are still selling the grey/gray lid.

Just skin it in matt/matte black for a few dollars. I did my R4, and it looks very good.
https://www.decalgirl.com/skins/391371/alienware-17r5-173in-skin-solid-state-black

 

yes correct the colour is region specific. in the UK where i am there is no option on the dell website to get a black unit, however, if call sales support and are willing to pay £40 more you can get it in black.

it shows up finger prints like a big red flashing light!

 

Absolutely. I'm in the UK...
Saying that my first unit was a 15r4. I9. I got a black one without asking. Was terrible for fingerprints and looks worse than the silver.
Returned it and got a i7 15r4. Came in silver... Much better and looks cleaner.
Thermals are great too. Peak at 72 whilst gaming .

 

New heatsink manufacturer? I wonder if this is only for R4.

Of course it is impressive because Dell average is calculated based on stock crap thermal stamp. Kryonaut improved temps even on i9 on R5 but not impressively. Probably because of bad application from Dell tech (heatsink was replaced but same thermal pads used).

 

Any pictures of the new heatsink?

 

as soon as the 1180ti drops i will use it in grfix amp = more heat reduction.

 

Hey guys, I’ve got a question.

I’m at the end of the line here, I just recently got the new r5: i9 8950HK, GTX1080.

Now my problem is the opposite of most people’s, my GPU is roasting and my CPU temps actually aren’t bad..

My cpu temps were reaching 98 max so I undervolted her by 0.170 which was nice and stable and now I get like 83 max during stress tests so the undervolt did wonders.

As for my GPU, it idles at 50-60 odd and if I do anything that requires GPU it flys up instantly to 95 at which point it throttles heavily and in game FPS goes out the window. The max on HWInfo I’ve seen the GPU consume watts wise is 110w (gaming/stress test) so it’s throtteling so hard it can’t get anywhere near 180w that it should be able to use?

I had the dell engineer come round they changed out the heatsink and it made 0 difference to the GPU temp, slight improvement to CPU... now they made me give them a 1 hour sample from HWInfo and are saying to me because my GPU is going 95c max but averaging 89c max that it’s within specs??? My fkin games are stuttering like hell the FPS is so inconsistent up and down and up and down and I’m stuck with it they won’t help me... it cost me near 3000 for this new r5 please can anyone help? Is the heatsink not touching the gpu or pads maybe, but the temps are so high on gpu i don’t even know if reposting/padding will get it down enough? I’m beyond belief right now...

 

Yes it is 180W and over sometimes.

The Dell technician (not an engineer by any stretch of the imagination) who came over - did he know what he was attempting to fix? Did he run simple tests before and after to see the differences?

I had this stuttering too, games became unplayable, and if you read the forums there is stuttering in any game with any graphics card and any system. The forums are full of people complaining. In my case the stuttering disappeared when I removed the NVidia latest driver and installed one from Windows itself (device manager, update driver) and it used some older version. Try de-installing the card and install the standard Windows drivers to see if it helps.

However in your case it may be a very bad case of the GPU throttling. My GPU also throttles, continuously almost, it is a power throttle, around the 180W. Yours may be a thermal throttle and may behave differently. You could install MSI Afterburner, set the speeds as down as they will go, and see if it stops the thermal throttling. You may get fewer FPS but stable. And you take it from there.

 

If you are seeing massive, and fast heat-spikes like that, I think the first thing that I would check is the seating of the heat sink. It could well be that even when you screw it down, it is making very little contact with the GPU die due to either a badly manufactured and QA'd cooling unit, or a pad issue stopping the heatsink above the die from making contact, or even something really weird like no paste on the GPU die.

Worth also doing, is checking that the fan on that side is actually doing anything at all. The whole system could be trying to cool on the CPU side...

Best of luck. Y

 

PS
As far as I know, (and I could be incorrect), the thermal throttling issues begin at 90C on these GPUs, so Dell are just about correct if your average is 89C. Although the unit should run at max below thermal throttle in my opinion, not average. But hey.

I can't even get 95C on the GPU gaming for a couple of hours in 4K and using a stock fan profile, so there is definitely something wrong here. It's time to crack her open old son, and see what is going on in there...

Conductonaut and new pads should get you -10C fairly easily, assuming that you do nothing else, and that the heatsink unit is making contact. Although I wouldn't be 'happy' with 85C, at least it would be usable.

Let us know how you get on.

PPS... (Sorry for the edits). Run an hour of 4K Far Cry (or any b*std GPU intense game) and then send Dell that test. Guarantee your unit will top 90C if it is this bad and need refund/replace.

 

Does anyone here have a problem with the WiFi not connecting automatically every time the PC boots up? latest version of windows and drivers and this issue comes up...

 

COD WW2 temps, 1440P, everything maxed. this is after LM repaste. game for me is very stuttery, lots of rubber banding, anyone else?

 

BF1 temps, 1440p, ultra. huge difference! no stuttering or banding, just silky smooth gameplay.anyone else get rubbish temps with WW2?

 

Hi Tommy. There is nothing wrong with the hardware in your machine. What you've described is exactly what's to be expected when there is a poor heatsink contact on the GPU die. Did you notice the stuttering in game was the same or even got worse after the engineer repasted the laptop? I bet you did since I doubt he would have replaced the pads and the stuttering is very likely caused by poor heatsink contact with the GPU VRMs (which usually gets worse if the rubbish stock pads aren't replaced) .

The good news is this can all be completely sorted and fixed. The bad news is you'll have to either do it yourself or find someone competent enough to do it for you.

 

Please post results using hwinfo. XTU gives you so little information it's criminal (plus it's buggy as hell and half the time doesn't work. I wouldn't use anything but throttlestop! ).
You won't get much feedback when people can't see actual running frequencies, power consumption, temperatures, voltages etc

 

I don't see a problem.
What is the GPU up to?

 

As I mentioned earlier, on my unmodified AW17R5 I am getting 45W, 90C, moderate thermal throttling - with the GPU running full tilt (180W). If your readings were also while the GPU was being maxed, then it appears you were using 45W with average temps of 75C, which is a -15C improvement over mine, and you should be happy ?

In terms of stutters, I had it for a while, I read 1,000s of similar messages on the forums, most for wildly different reasons, I re-installed my drivers and used a much older version and the stuttering went away completely. Rubber banding not sure, unless you mean screen tearing? Is vertical sync on?

 

Please go to the graph settings in XTU, and add the CPU power, "Package TDP" - since we are talking about thermals, it will show you what kind of power the CPU was expending during your testing (and the cooling system removing, which is the biggest discussion point). This is because a whole lot of parameters contribute to power and power converts to heat. The heatsink removes this heat. But the parameters are too many and it is easier to just read the power. Parameters are clock speeds, number of cores used, mix of instructions used as some instructions consume much more power than others for the same unit of time, voltage of course etc. Since we can read the TDP directly we do not care about clocks, voltages and what have you. Of course the aim is to minimise the power so we undervolt if we can or we run it slower or we run simpler instructions (eg NOT Prime95) and so on.

 

Is there a quantifiable way to measure the GPU VRMs before and after repasting/modding ?

 

Please... Stop using xtu for posting results... use hwinfo

 

Hmm tbh XTU is better, it shows the peaks but also the average more clearly.

I mean I always have a peak of around 83c and after the fans actually are turning on it lowers a few degrees. In HWinfo that image is skewed because it simply only shows the peak of 83c. Average in HWinfo also includes ofcourse the idle temps. So in the end seeing the actually temperature graphs is so much better.

 

Maybe i'm a total pleb.. I'm trying to find Iunlocks guide for LM for the alienware r5 1080 8950HK and I can't find it anywhere. I'm looking to do the full job, tapes on the chips, pads, etc. Any help would be greatly appreciated. I'll post my results upon completion.

Thanks,

Encrypt3d

 

Yes, read the GPU stuttering thread from Alexnvidia

It talks about using sensors in GPU-z to look for VRel limit reasons every time the GPU power and load drops to 0 in game. this is evidence of stuttering due to the GPU being starved of power because one or multiple chips used for power delivery overheats. In game you will notice a freeze for a few seconds until the chips cool and start to deliver power at a throttled rate. Over time this increases back up towards TDP until wham! The cycle repeats

Sent from my SM-G960F using Tapatalk

 

With XTU you just lie yourself and confuse others...best example is doofus

I edited your post and draw a line where the avg is RED LINE is 75-77 degree... how can your avg be 70 if there is barely an value under 70?



and then he write:

In his screen I see avg 75-77 degree with wimpy 20w. So actually he has the 75 degree with 20w but he claims he has this temps at 30-34w... That's 50%-75% difference between what he show and what he claim! That's why HWinfo is better. And that's why I don't believe this wattages without proof. 41w constant Load average and average 85 degree without throttling.
stop cheating results and use HWinfo
-start hw infio
-tab into the game
-play the game
-tab out
-make a screenshot of hwinfo



this is BF1 with 4.7ghz, avg 78 degree, 87w max and 74w avg. Nearly same temps as doofus screen but 3.5x power consumption (55w more possible) with LM

hwinfo records the cpu load (min max avg) and with this we can see if there is a long idle time and you can post an xtu screen to show that ther was no idle time

 

There isn't one yet. He's done a thorough guide for the 17 R4 which will be 90 to 95% completely the samefor the R5 with i9s. Only the heatsink images and disassembly will be slightly different but it should be obvious what those differences are when you do it

Sent from my SM-G960F using Tapatalk

 

I'm not sure if people realize that the hottest part on these laptops is the gpu's vrm. 1060 vrm on the 13 r3 can hit 90℃ with the bottom off and that's with an ir gun so it's even hotter underneath. Especially when the 1060 can pull a max of 80 watts. Can't imagine 180 watt 1080. Easily over 100c. Which can easily trigger prochot and cause stuttering.

 

Well maybe I eyeballed it wrong and made an honest mistake. If I did, which I did not, but say I did, it would not have been done on purpose I am not trying to sell anything to anyone.

Looking at my graph again, I believe your editing of it is wrong. Here is my editing where I have drawn my lines and in the beginning there is an average say of 73C say, but then, and for considerable time it drops significantly and just reaches the 70s. The average then drops more near the 70C and I eyeballed it as 70C.

Further, looking at the TDP, I have quoted it as 20W but it was way more, more like 23-24W ? At the end the TDP seems to drop towards the 20W at the end and during that time we barely hit the 70s.

I am not trying to sell or to deceive anyone and your post is highly accusatory without basis. I do not care if the temp is 80C or 50C, I do not get brownie points for it.

 

even if so...and I still think you twist facts like you want... there is still a difference of 50% between your claimed 30W, 72C and your actual 20w, 73C... That's what I mean. If we want to see the difference between stock and LM / repaste then start posting hard facts and stop your guessing games. This way the discussion is senseless.

You`re selling your ,,eyeball facts,, as real results...that's no base for a discussion...I still want to see 41w average load with 85 degree without throttling and 180w gtx1080

 

I am not selling anything. Look at the graph again, please look at the yellow line I drew. This is 20W. We are way over 20W for the most part when the avg temps are over 70C. Then we reduce close to 20W and the temps drop to well below 70C.

 

At this point i will stop discussing with you. You just play guessing games with us. This is just stupid. You just ignore every other argument and twisting facts so that they fit to your argumentation...

 

Thanks I found it and am reading...

 

Ok perfect, I can't find that one either. Would you mind sharing it? I was more so interested in knowing the pad thickness to make sure I order the right ones. I'll document my steps and share the results with the forum (I own a computer repair company in Toronto, Ontario and this was is hopefully going to be my primary work laptop.)

Thanks for the help Pete

Encrypt3d

 

I left the hwinfo running and was away for lunch, send a couple of emails and back to lunch. Minimal use, no games running. Yet the hwinfo recorded a lot VREL and Power thottles and utilisation throttles when it fact there should be none and the power graphs do not show anything either.

On the charts below you can see the GPU power quite low for the duration, temperatures in the 40C, fans almost silent, yet there are power throttles, utilisation throttles and earlier on there were also vrel throttles.

In addition when playing my games the GPU appears to be under an almost constant power throttle according to hwinfo, but I cannot see anything happening in the game, ie no pauses or stuttering or anything.

Shall I put this down to hwinfo being wrong?

 

Sure

http://forum.notebookreview.com/thr...5r3-disassembly-repaste-guide-results.797373/

Sent from my SM-G960F using Tapatalk

 

It's normal to have this when not under load. Not VRel when under 100% load though. that's the only one that matters for GPU stutter. Look for that when load drops to 0

Sent from my SM-G960F using Tapatalk

 

There isnt any, only for the previosu generation which is very similar.

 

Can you tell me what exactly do you do to "balance" the heatsink?
My heatsink, although has bad gaps, doesn't cause temperature differences. Is it "balance"?

iunlock wrote that instead of soldering I better "rebalance" the heatsink? What is it?
I don't understand why I need to "balance" my heatsink. The problem with my heatsink is the big gaps between its "tails" and the "peripheral" components it needs to cool. For those the best and easiest way is to just "thickening" them with good thermal conductivity metal, let them closer to the heatsink, by copper-soldering.

 

I THINK that you are referring to other components covered by the heatsink other than the actual CPU/GPU ? If so, and you have "big gaps" as you say, then I think that unless you have a really bent out of shape heatsink, you have either:
1. Thermal pads that are way to thin, or,
2. A machine where some doofus at the factory didn't put the thermal padding on at all.

Any chance of taking a photo to confirm? If I'm correct, you really do not need to go soldering stuff. You need pads.

 

it's not easy. It requires patience and a light touch. You need to work out what needs bending and do this very gently with the aid of a heat gun. Every heatsink is different so there's no method which fits all really.

Balancing directly refers to how well the heatsink is in contact with the CPU and GPU die. if it's balanced correctly then the heatsink will sit perfectly flat on both chips (at the same time) with no rocking or gaps visible as well as no rocking if you push the heatsink down on either the CPU or GPU. it will therefore have the best possible contact for thermal dissipation. This can also be confirmed by having an even spread of pressure using pressure paper or a thin 0.5mm thermal pad like Arctic

 

Can someone please help me? I have the new 17 R5 i7 1070 512 ssd that I just setup today and the laptop was performing fine at first. Now the fans are on all the time. Quietly, but annoyingly when doing simple tasks. Don't know if this is normal seeing as the blade 15 didn't have this problem. Secondly the wifi is super slow out of nowhere. Was fine when downloading a game. Now not so much. Lastly while playing games I think the gpu is causing a stutter. The game gets super laggy out of nowhere. Thanks to anyone that can help solve this. I really don't want to return this laptop, but may go back with the Razer or try the aero out at this point. I'm loving the 4k screen on this thing by the way. Except for when watching 1080p. Don't know if that can be resolved either

 

Wow.....Alienware customer wanting to go back to Razer...how low can the Alienware brand go, from the good old days of great performance and quality control to Dell who seems to care less what happens to the brand's reputation.

 

I still haven't figured it out and am really disappointed after spending a couple grand. If anyone can help me with this I'd really appreciate it. I literally had this same exact laptop with qhd and now I have UHD. I don't know why there is so much stuttering going on in games. Here is a video of it.