*OFFICIAL* Alienware 17 R4 Owner's Lounge

Hi guys so finally I was able to do my re-pasting. It was terrifying, there are so many things that can go wrong. I stripped two screws that were over tight. One of them was one of the ones that connect the MOBO to the heat-sink..... that sucked. I was able to remove it by using a mini drill and going from the opposive side of the MOBO basically disintegrating the mini screw. Luckily no damage to the MOBO .... it sucked. I feel angry and dread having to do another repaste in the future, customers should not be having to do this.

These are some screenshots of my temps are FIRE STRIKE score with no OC

 

This one is with CPU OC lvl 3

 

What paste did you use ans what were the temps on lvl3?

 

I have a question, CPU iddle temps hover around the 30's and 40's after repasting, when I run prime 95 the MAX core temp is 74. Is this good?
Also do I have to keep an OC level of 3? what is currently recomended?

 

I used the grizzly paste that someone suggested. I did not record the temps on Firestrike with the OC LVL 3. I only recorded the temps on Prime 95 with lvl 3 overcled and they stay at 74.

Im gonna run it right now to find out..

 

It's fine for temps and idling. Definitely keep level 3 or make your own OC profile as that's the ONLY reason to have gotten the CPU over the 6700hk. They will be the same power if you don't OC. What were your temps on your overclock run?

edit: One thing to consider is GPU temps are more important (in my opinion) to be concerned with and so far these GPUs are maxing at very low temps. Modern intel CPUs that run in high end desktops have very low temps. However there have been THOUSANDS of intel i7 CPUs being put into gaming laptops for years running constantly in the 90s and I haven't personally heard of any CPU failures. Other things may fail and I am sure some people may have fried a CPU, but I haven't heard of any which means we will be onto the next BGA laptop before our CPUs will ever die.

 

Kyronaut or Condactunout?

 

ok I just ran firestrike with OC LVL 3. here are the results. Are these temps good for a lvl 3 OC? BTW I also repalced 90% of the thermal pads with better quality ones.

 

Its the krionaut.... is that good?

 

Yeah thats the best thermal paste. Condactonout is liquid metal. I plan on using kyronaut as there are less risks involved and fujipoly thermal pads. Altough if I get good temps out of the box I wont dovthe repaste.

 

New to the forum here. Just placed my order for the Alienware 17 R4. $2,440 after tax using employer discount. expected delivery date Dec. 21st.
Cant wait!!
1 338-BKFU Intel(R) Core(TM) i7-6820HK (Quad-Core, 8MB Cache, Dynamically Overclocked up to 4.1GHz) $
1 619-AHCY Windows 10 Home-HE (64bit) English $
1 490-BDMP NVIDIA(R) GeForce(R) GTX 1080 with 8GB GDDR5X $
1 391-BCXE 17.3 inch QHD (2560 x 1440) 120Hz TN+WVA Anti-Glare 400-nits Display with Tobii IR Eye-tracking $
1 370-ADGP 32GB DDR4-2400MHz; up to 32GB (additional memory sold separately) $
1 400-ANVR 512GB PCIe SSD (Boot) + 1TB 7200RPM SATA 6Gb/s (Storage) $
1 555-BDFG Killer 1435 802.11ac 2x2 WiFi and Bluetooth 4.1 $

 

Are my temps and firestrike scores good anyone?

 

Yes they are very good for OC level 3 take care of that laptop mate youre a lucky guy.

 

cant open your image. Can you repost it or just list the 4 temps?

 

Nm It is showing up now. The temps are pretty close to in line with people running lvl 3 OC. According to the Intel guidelines, as long as the CPU doesn't "Sustain" 80s, then everything will be fine. Your average temps are fine. Good job doing your own repaste!

 

It sucked. I don't even wanna do that again. I was scared of breaking the ribbon plastic holders and stripping more screws.
How often are repastes supposed to be done?

 

If you guys/gals want to really see your temps I would recommend running Crysis 3 or CS:GO as both are very CPU intensive and my temps go higher in those games than they do in any of these benchmarks.

 

People will say never on a paste used like what you did or I did. However a good general rule is in 2 years. My only concern is that the heat sinks are flawed enough that the paste may not last or be sufficient. Some long Crysis 3 and CSGO gaming sessions have already spiked my temps pretty high. Previously, I ran a few bench marks back to back with lower temps. I ran the CPU benchmark OCCT for an hour never exceeding 80*c but my gaming sessions (which are the ones that count) have spiked it higher in under an hour.

 

Wow that doesn't make any sense. I'll try running crysis 3

 

How were your core differentials when you got higher temps? I noticed playing bf1 last night my cpu temps went way higher on a certain map. I never hit the 80s before playing but i was up at 85c last night and i noticed the higher it went the more core difference i got. This is with a 4.0 overclock on all cores. I noticed it throttled too once it hit 85c, my frame rate crapped out for a good minute or two down to the teens then went back up to the 90's-110's, if i keep it under 85c it never happens. Its aggravating since my lower cores are hanging around 66/67's max but my top two are hitting 85/83 at certain points. Im debating on tearing it down again, and bending the arm/putting a little more paste up top to see if it helps.

Im going to take your advice and try out crysis 3 myself tonight and see what happens.

 

I agree on Crysis 3 but CS GO? Hmm.. I didn't know that game was demanding CPU wise.

I use Witcher 3 as a benchmark since it is a high demanding game.

 

I agree the repaste sucked, it was alot more work than i thought it was going to be. I was so paranoid about the ribbon cables, the little white clip that held the motherboard ribben on came off when i did mine. I panicked for like a hour over it because i had no clue how it went back on and i was so afraid i was going to put it on wrong and bend the pins. Luckily i managed to put it on the right way, but it was really nerve racking. I also had issues with my screws, the ones on the back panel wouldnt come off for the life of me. I stripped the top of one pretty good, but managed to get it back on with the flat head. Using a phillips to begin with was a bad idea. I did the same thing to a heatsink screw too. Idk if it was just me but the screws seemed really cheap. Ive read its good to do a repast every two years myself.

 

my core temps are within a few degrees of each other outside of core 0 in the instances that one goes up really high. When the temps spiked it brings all my other cores to the very high 70s. I didn't experience any throttling when I hit 85 though. If core 0 doesn't spike very high then I usually am benching around 72-76 on all 4 cores.

I am also considering another tear down but am thinking I will do so in awhile. I worry more about stripping screws or an accidental mishap the most. Part of me doesn't think this CPU will be in any danger of serious overheating as it is right now. My cores are likely close together because my paste IC Diamond is very very thick. I just wish it held my temps a bit lower. As long as I don't hit 90 I don't think I will redo anything for awhile. I have run hard OC laptops in the high 80s for years without one ever dying out.

 

I don' have witcher 3 but yeah, that game is a beast for benching/monitoring. CSGO isnt anything graphically demanding but the source engine is actually known to be pretty CPU dependent. It is a little surprising that my temps are as high as they are playing a game that looks very dated. I wish my temps were lower because it is my favorite game and I do play for hours at a time. Skins=wins lol.

 

Hmm something today I learned about the Source Engine. I'm intrigued.





Sent from my E6853 using Tapatalk

 

Yeah it it wasnt for the thermal throttle i had, i wouldnt really care about hitting 85. The weird thing tho is even when my top cores warm up my bottoms stay the same. I had about a 6c difference between the top and bottom normally. But i noticed when the top ones heated up over 85c i ended up having 15c degree difference. I think it has to do with your paste, when i noticed your temps compared to mine, i had generally lower temps by like 3-4c, but i had bigger core differences. So that thicker paste your using might be helping balancing out the heatsink. I probably should have aimed my little thermal dot more to the top then in the dead center.

I need to check my bios version too, i know alienwares "thermal fix" was just to up the throttle temps so maybe i just have a older bios. I just figured it was up to date since i just got it on the 2nd.

 

Probably highly dependent upon on the particular paste and daily usage. I've been using Arctic 5 for years and use my laptop ~10 hrs/day. I can't go for more than a year without seeing my temps climb regardless of how often I clean everything. Just pay attention to your temps. When your paste is no longer effective you'll see higher temps for no other obvious reason.

 

The answer is: "When the temperatures start going up again, and it gets to the point where you can't comfortably use the laptop without a repaste"

Some people will repaste at the first sign of elevated temperatures.

I have seen people doing it every 3-6 months, every 9 months to a year, but never have I seen those people wait it out for 2 years

Really it's up to your own sensibilities for the increasing temps over time.

Air is what causes the oxidation, so if you can use paste that is "rubbery" or seals out air and moisture - condensation, you can extend the time the re-paste lasts.

The AW paste doesn't sound good short term, or long term, but only time will tell.

 

Glad to hear I wasn't the only one. the whole process was stressful. By the way for the ones that haven't done the repaste yet, the screws that keep the MOBO and heatsink togueter are smaller than the rest of the laptop screws. They require a smaller Phillips head..

 

What else would I use if not a phillips? I don't get it...

 

I My 4930mx in AW17 was re-pasted with Liquid ultra early April 2014. And the liquid metal is still just as good 2.5 years after Sorry bro, I needed to fix your post

 

I switch over to a flat head screw driver, after i realized how cheap the heads were and it did way less damage. Atleast with a flat head screw driver you get a second chance if it slips.

 

It's not only the point of damage but throttling. Since the temps are high you don't get full power of the CPU or GPU

Sent from my SM-G935W8 using Tapatalk

 

Yeah tbh I haven't had to repaste more than once most of the machines I have had but I do upgrade fairly often.

 

I went ahead and requested for a screen replacement hoping there are better ones in the woodwork. The one I got shows serious signs of screen banding issues . Ran Firestrike and this is what I got so far. You might even be able to tell the screen banding issue I got from the blue around the overall score. Everything else is running perfect and cool after a repaste.

 

What are your gpu core and memory clocks offset?

 

100/150. Just started playing with this thing and it looks like I'll break 23000 graphic score easily when I'm done.

 

Details please.
GPU clocks
CPU clocks
How did you set above clocks.?
What gpu drivers?
Which BIOS?
What are your temps?
Which paste did you use?

 

Be careful with the memory OC. It will run at much higher clocks than what is the most stable. I've been pushing the 1080 I have to find my max. It'll run at 300,400,500 even 600 offset. But the best FPS results running Firestrike loops is at 175MHz.

Was able to get the card to take 200MHz on the GPU clock. That pushes the max to about 2075 which is really close to a MAX OC on the 1080 cores from other reviews. Still doesn't run at that max clock speed, but the graph as a whole is pushed up by 200MHz. When the card is actually running, it's constantly hitting power limit and downclocks to a lower voltage point. So actual freq during a run sits around 1775-1925. Do a GPU-Z Log file to see how the card is actually acting or have an afterburning graph running you can analyze.

Also, power limit seems to sit around 166-168W during Firestrike. And about 155-158W during Firestrike Extreme. Need a VBios to uplock these GPU's to about 200W and it'll help ALOT! 775DM with the MXM cards has a TDP design of 200W. Base GTX 1080 like ours has a Nvidia spec of like 165W TDP. Restricts the voltage points and lowers the running Frequency.

Oh. I also tried to max out the area under the curve by modifying the curve to add more freq at the common voltage points commanded. Didn't really work. Crashed Firestrike. 0.943V was at 200MHz offset. Tried 227MHz and it crashed. So 200 is a lucky sweet spot it seems.

 

RESERVING for Benchmarks at 36F/2C testing outside. LOL. Should eliminate CPU Throttling I hope.

OK. Point of these test is to push the hardware were it's advertised to go BUT removing as much of the temperature from the equation as possible to eliminate the CPU throttling that occurs at room temperatures. There's still a Delta-T of about 40-55C.

Firestrike with 200MHz Core OC, 150MHz Memory. CPU @ 4.00GHz (AW lvl 3 OC essentially)

Timespy same as above.


Notes:
Re-pasted using Grizzly Kryonaut and Artic Thermal pads. GPU Driver is Nvidia 376.19, BIOS 1.0.2. OC set via Manual @ 40x, no volt offset, 240,000/240,000. Tweaked via XTU to 300,000 for power limit unlock.

CPU at 4.0GHz. Core temp difference of about 17C between Core 0 and Core 3. Delta of about 55C total. Increases to about 20C Delta at room temps or close to 70F/21C.

Power limits to about 155W-161W at colder temperatures. Room Temperatures it's about 161-168W. Hits power limits constantly and downclocks via selecting a lower voltage point.

GPU has a 200MHz GPU boost and 150MHz on the Memory. Noticed a drop in frames so seems 150MHz is the best the memory can do before ECC takes over and reduces timings and frames.

Typical Freq and Voltage during runs: 0.893-0.912 corresponds to Approx. 1873-1911MHz GPU. Timespy is close to the same GPU load, Voltage and Freq levels.

 

AW 17R4 - GTX 1080 [Fixed and ready for the track.]

• Grizzly Conductonaut
• Fujipoly Thermal Pads 17.0w m/k
• Heat Sink Mod
• Uneven Heat Sink Fix

Note: It would not run OCCT stock out of the box. Thermal Shut Down. 30C Core Differential .

1 HOUR OCCT (After Repaste)
* Core Differential ~3C Max between Core #0 and Core #3.

Max Temps: (~3C Differential.)

• 57C
• 56C
• 55C
• 54C

Current Temps: (~1C Differential.)

• 49C
• 50C
• 49C
• 49C

Average Temps: (~1C Differential.)

• 50C
• 49C
• 49C
• 48C

After:


Owner of this Ice Box: @GTO_PAO11

Before:

Uneven Heat Sink (CPU cold plate) still exists on the GTX 1080 variant...however, it can be fixed.

 

So you went from 90+ temps and 30+ differentials to mid 50s and 3 differentials? Are you making love to these machines? I don't get it...

 

How about 12 Hours straight with 1 Break. (Yes really, ask @GTO_PAO11)

I don't do just simple repastes as I've spent several hours alone working on and trouble shooting the uneven heat sink issue, which is now corrected as you can see from the screen shot.

Looking further into the uneven heat sink issue will help others as well. Two in one... I've discovered some new things yesterday, which I'll make available soon.

 

Very impressive!!! The heat sink mod that you speak of is this a new mod technique or similar to the arm bending technique?


Sent from my iPhone using Tapatalk

 

It includes the bending of the tension arm still along with.....(I'll make a video for you guys.

Hang tight...

 

Awesome! I can't say thank you enough man!


Sent from my iPhone using Tapatalk

 

Very good work. Temp are very impressive

 

That's my BABY!!!!!
iunlock is the business!!!!!

Sent from my E6853 using Tapatalk