*OFFICIAL* Alienware 17 R5 Owner's Lounge

Thank you all for your responses! I believe I've come to a decision. One last question however, on the chance I'd like to upgrade my display panel in the future is it possible to replace the current?

 

Had what I thought was an issue today, suddenly my game dropped to 7 FPS and checked everything to understand why, in the end it was the GPU memory, one bad application was using 5GB of GPU memory and it is a desktop application, you would never think it is the culprit. This is because on the G-Sync model the Intel GPU is not connected to the monitor, and the NVidia GPU has to do all desktop rendering. In that way a rogue desktop application can cause so much damage.

 

I watched two youtube videos showing how to dismantle the AW17 and how to take the panel out. On the one video he did it by taking almost everything apart, on the other he did it without touching the underside panel. Replacing the panel is definitely possible at home from what I saw.

However if you want to replace with a different panel model then I do not know, the connections etc. It may even need a different motherboard who knows.

 

Battlefield 5 Alpha:
1440p, i9 8950HK @4.5ghz, GTX 1080 @1900mhz

Ultra Details: 78FPS average
High Details: 88 FPS avergae
Medium Details: 104 FPS average

CPU Temps hovering in the 70s
GPU Temp 65 degree

 

Is it on your machine? How can I get it? I mean BF5 lol

 

Mine looks okay.. Not sure about the rest though.

 

You need an closed alpha invite ^^

 

Hey guys, I have read quite a bit of this post. I have decided to buy the R5 with the i9 and gtx1080, i am prepared to do a repaste with LM but im unsure if i will need to repad. Any insights??

 

definitive repad!!!

 

Will following iunlocks R4 repad be a good idea? from what i can tell the R5 has a different cooler

 

is a good idea... Used the same guide. I used the i7 7820HK/GTX1080 picture for repading

 

Did you happen to do the PCH mod? if so do you have the finned heatsink size?

 

A few questions for R5 owners - since I just bought a 15R4 assuming close parity on the two models in software/features.

Is the new Alienware Command Center a solid upgrade? Does it faciliate UNDERclocking? Has anyone tried the "quiet" fan mode I'm seeing in some video overviews (by umar on Twitch)? I'm curious if my 1070 equipped 15R4 can game quieter than normal gaming notebooks and still hit 45-60 FPS on the 60hz GSYNC panel I got with my unit.

I found Dell's marketing material to mix old with new features. For example, the keyboard can support 16+ Million colors now? Or is it only supporting 20 colors? The product details page says both things. How are the new lighting controls and color choices in general? Can you set the keyboard backlight to all white? I know it's not per key lighting - but the few videos I've seen are saying the AWCC and control module are a significant upgrade.

Is there any good video online of a user walking through the new software?

 

@Papusan since we talked about full load tests, i was testing a few ways to get a more realistic real world ,,worst case,, scenario.
Was running Firestrike CPU Test and Heaven at the same time in endless loop. I (in my own opinion) dont see any sense in stressing my CPU with useless pi calculations.

I think there will never be a game stressing my System more then what i did here.

 

What settings are you running to get these temps/clocks?

I assume LM and pads were done

soooo keen to get mine delivered, est 20/7 though

 

Just LM and Pads + undervolting.

Firestrike CPU test (Loop) and Heaven Bench (max Details, extreme tess, 8xAA, 1080p).
GPU at 1900 with slightly undervolt + Core and Mem overclock
CPU at 4500 with slightly undervolt

Temps are high, yes, but its worst case and summer

This was ingame BF4 yesterday after my test.

 

what are your ambient temperature's @26C ambient my 7820HK (4.4Ghz = around 70watts when gaming for me) my cpu still won't reach 85C (gpu is around 170/180watts) while your's has a undervolt ? and those temps @1080p with a 1080 says something is not quite right YET but it seems to hold to it's own. when i benchmark heaven in 2560 x 1440, my temps start to rise (on the gpu side) hovering around 75/80c. seems you need a pad before you can OC (without it throttling).

 

I had 24 degree ambient.

The 6 Core Coffe Lake runs hotter than the 4 core Kaby. 70w during gaming with a 4 core? ind this Video the 7828HK @4400mhz consumes 68w @100% in cinebench... How can you get 70w during gaming then?







pls do the same test: Fire strike cpu test Loop and Heaven Bench and run hwinfo, then we can compare and see the reals temps and wattages i tested 20 mins to heat up the System before takting the screenshot

higher resolution do Change the temps only a bit. But i will run it again with FS CPU Bench and Heaven @ 1440p. I dont see any problem with the GTX 1080 @ 1900mhz @ 165w @ 65 degree @ 99% load... where do you see a problem? Temps too low or too high (lol) ?

ähm...what? Only Heaven and 80 degree? In my test i had Heaven AND FS Cpu test running at the same time.

There is no throttling.

 

also cpu usages when benchmarking is not fixed. every cpu boosts, and pulls V different.

68Watt only ??? mine reaches up to 85 !!!!!

something is very wrong with that cpu, or it does not run fully unlocked.




just look at how much watts @4.3 on a fully unlocked 7820HK generates.


and a 4.4 again



for comparison, my 1080 does not go past 60C @26C ambient on 1080p. ONLY on 2560 x 1440 will my temperature's push past the 75C else it never will go above 70c (when gaming on 2560 x 1440) your temps could be lower especially when using LM (and a "vapor chamber").

fyi all my benchmarks are done after gaming for an half an hour. All my score's are based on benchmarking after some heavy usage so my temps stabilize faster.

Come have a look in the R4 owners lounge. plenty of screenshots up to 4.5 Ghz (oc) with a 200Mhz gpu OC and a 450mhz Memory OC.

i don't run undervolts, because i bought a K processor to OC, not undervolt.

and i didn't say you're throttling, i said you need a cooling pad before you can OC WITHOUT throttling. if you're running undervolts across the board, try adding 100Mv and bump your clocks to 4.4 on all cores and see how it goes.

again i don't want to brag, bash anything (we own the same laptop) it's just for a laptop with a six core AND a vapor chamber AND liquid metal, your temps just should not be that high. that's all i'm saying. but i do have some firestrike score's for you in the 18000 and almost 19000

 

my gpu never passes 70 degree when runnig 1440p usualley it sits at 62 degree...so What do you want to say to me? look:

BFV @1440p @ Ultra Details @ 4500ghz Sixcore

i dont think so.

big Mistake

what? I already run 4.5ghz on my Screens ???

Dell uses different LLC on Coffee Lake models and too much voltage... Not using UV is just stupid.


Again: please do the same test to proof your post FS cpu test and heaven at the same time with HWinfo.

 

come have a look in the r4 owners lounge around the 800 pages

if 1 core can hit up to 4.5 is different if all core's hit 4.5. you cannot run a 200mhz overclock with a undervolt. stop spreading bull

 

you're comparing a game with a benchmark ?? i said my gpu goes past 75C when benchmarking @2560x 1440 not gaming. then it would not go over the 65C

 

Oh Boy... 4500 is ALLCORE on mine
And YES with undervolt... Stop comparing kaby with coffelake.

THIS actually IS allcore while running Heaven + Firestrike CPU Test @4500


Come on...do the same test

Name the Benchmark where you get 75 degree GPU and i will run it

 

that could also be your max, and when your speeds shifts from #CORE 0 to 5, all of your core's will look like they've been running on 4.5 while it was only was your max for a microsecond.

i still haven't seen HWinfo where it shows your voltage and clockspeeds for the duration of the benchmark.

you can ask papusan how it performs, he helped me tune my system.

@1080 all gpu benchmarks are below 70C. @2560 everything goes above 75C (cause of the overclocked core and memory) but stock it would be a max of 73C.

i'm at work, but as soon as i get home, i'll post my fire-strike scores

 

its not an max its the clock the cpu runs all the time.

i still dont see the problem you have with my repaste

 

I'm waiting for your apology later...

 

I have read your previous posts as well. Can you check up my older posts and give answers on them? I’m on vacation with small phone. Especially about What I talked about Core voltage.

Edit. Also Re-read bro @Falkentyne ’s useful posts about the topic.

 

Show HWinfo or nothing. On this forum HWinfo is used for logging and providing numbers.

I haven't seen a single file of you that shows how your cpu behaves when benchmarking. Cause everyone with the I-9 has to Overvolt to get to 4.4 on six cores, let alone 4.5.

Again I'm not bashing hating etc, I'm just point out that something is not quite good and that you can get better results. I haven't said anything negative because we have the same laptop lol. Not because you have an 2 core's extra it's a different cpu lol ? Coffee-lake is built on the same 14nm Skylake and Kaby-Lake where built en is effectively called 14++because Intel cannot get 10nm off the ground. So I can surely compare Kaby-Lake with coffee-lake. The numbers just shifted cause of the 2 extra cores.

 

Aye aye cap

 

your should open your eyes when reading....realy... Ive posted so many Hwinfo Screens the last weeks...

examples:

Cinebench @ 4.5ghz = 1478 Points



Cinebench @ 4.7ghz = 1538 Points



Cinebench @ 4.8ghz = forgot the result...was above 1550, minor powerlimit throtlting



you can find all the pics and more in this thread...

the CPU architecture is the same...yes... but its a different bios with different LLC and different VID regulation. That's why I need an undervolt even with 4.7ghz. Every bench was -140mv. but when rising the clocks, the bios adds an voltage offset too. And then the voltage is to high again.

You have 82-85 degree with Cinebench @ 4,4ghz... with wimpy 85w. I have 78-80 degree with 4.8ghz and 110w...

again: my repaste is bad why?

 

Can you find my older posts about voltage. And the scores doesn’t match what I got with 6 core Coffee. Same architecture means equal scores. Ram speed below won’t do much. Most today’s is 2666MHz.

FYI. Stock means +1450

 

Cap. Those are pictures. I said that you should be able to prove how your cpu runs DURING THE DURATION OF THE BENCHMARK. A picture of HW info @23 seconds won't say ****e.

Why won't you post something like this

 

everytime we talk about this you say different numbers... when I score 1350 you say it should be above 1400

then I fix it and get 1397-1420
and then you say it should be 1450+. @4.3ghz my CB score hovers around 1390-1420..depends on backround tasks. faster Ram would give me 1-2% and as you already found out it seems like the mobile coffee Versions need 1 bin more for the same score... so 1400 stock is fine for me my 4700mhz score is similar to your 4.6ghz score with faster ram... so I think youre right that mobile coffe need 1 bin more...

my windows is normal, stock bloatware Dell-Win10... nothing optimized. AW Command center in the backround aso... I don't care about 1-2% higher cb scores... the range is the same, so I know my cpu is running fine.

and if you think iam lying... your problem. I don't care at this point.

 

bruh where just trying to help you, but if you want to get all defensive and stuff, be aware of who you ask for help when your system breaks down.... papusan isn't the one with a off the shelve BGA laptop, it's us. he doesn't need to buy a 400 dollar warranty (every year) due premature failure's of multiple laptops around the world. it's us... so i actually think that brother @Papusan does not care, and you should (cause you paid big bucks for a cpu that fluctuates with it's performance while it should stay level when you've tuned your system) on top of the fact that stock 4.3 usually score's around 1350/1400 (without Liquid Metal) and you have to get almost up to 4.5 to score that, something is wrong.

why buy a K cpu when you aren't going to use it's full potential ?

 

first of all... I didn't ask you for help. You claim my repaste is wrong, but your temps with 400mhz less are higher then mine with 2 cores less and wimpy 85w instead of 110w... so why is my repaste not ok?

second

I slowly think you have reading problems...

Papusan scores @4.3ghz 1450 with faster ram.
mine scores ~1400 with slower ram.

The ramspeed makes 1-2% -> 1414-1428 so my score with same ram speed would be 2% lower then his ( with optimized win and maybe other tweaks)... so the difference is 2% = 86mhz. Even if the ram speed only makes 0.5% difference we talk about 3%= 129mhz

In another thread there was a discussion that mobile BGA cpus are 1 bin slower than LGA...-> 86mhz - 129mhz is in 1 bin range -> so I think this could be true. But then my 1400 score with my slower Ram is OK as it is...

and its fluctuating about 1%... I don't care.

My 4.7ghz is similar to his 4.6ghz score which is again 1 bin. Where do you read that I need 4.7 to match his 4.3ghz score?....realy bro, you should read posts before answering

 

I expect you run 2666MHz as me y sticks. One my phone, not easy to post all info. Stock 4.3GHz 6 core clocks.

Edit. See post about voltage.

 

nope 2400mhz... which is 10% slower than your Ram ...

 

Then read about what I have said about ram speed and Cinebench All info can be found with a search.

 

Yes... you said 0.5-1% cinebench score...

My score (1400) +1% = 1414. Difference to your score 53 points = 3% = 129mhz= ~1Bin

All Info can be found...

 

Yees, sir. And all silicon is equal from Intel’ factory. My points stands. - 1 bin. And I haven’t talked about higher overclock.

 

And i said you could be right... i dont see why we need to discuss this again

Misscalculated: 43points=3%=129mhz

 

Where have I said that your repaste was wrong ?

then give me the post number where i say that your repaste is wrong. and i want to see that i LITERALLY SAY THAT YOUR REPASTE IS WRONG.

cause you're trying to make me come out like the bad guy, while i just pointed something out in comparison to my OC.

lol and no no no no. look at my 4.5 and look at your 4.5

@4.5 Ghz my max is 79C and i've accumelated around 90watts heat. you @4.5 with a six core only accumelated around 84 wats. lthat should've been around 100/110 wats, so it's not running full speed. i bet you're running even adaptive voltage ....

and another thing i noticed is that your @4.5 Ghz your current temp is the same as your maximum, same goes for your cpu package power. And your average is just 2 C higher than your "minimum" temperature. Meaning you cleared the log and then made a screenshot cutting out the part at the bottom where how much time passed is logged. that also makes your benchmark score's iffy cause it clearly shows it's a fresh log else your current temp could not be your max

i mean a minimum of 56 watts ? who are you trying to fool ? you where benching and made some screenshots and cherrypicked the one that looked the best, that's the only explanation i have for those values. else your minimum would look like >15 watts

have a look at my @4.5 Ghz



and it was not about you're speeds, it was about the temps you claim @ the speeds you claim !! where not saying you're lying or you're a fraud or anything. we just want to help you get REAL WORLD values not some HWinfo screenshot without cinebench in the back. cause with you it's or cinebench or HWinfo or some game with a screenshot made with a phone while HWinfo can record !!!! even the command center has a option to log your hardware usage (in the old cm center, do not no about the new one)

so why keep making those sketchy pictures while a log gives us all the info.
we're all notebook enthousiast, where not here to be salty. why would be lol @Papusan his machine is almost twice as cool as your system @the same speeds.

but hey, if you do not want to get the full potential out of your money be my guest when my system brakes down LGA is the way to go

the fact that your PCH runs hotter than mine, already dissaproves your point and your system will start throttling in the future.

 

There:

Just look at my pic and your pic and think just 2 seconds about why this could be true...
Maybe because your voltage is 0.1Xv higher? Higher voltage= higher Power consumption. Just maybe?

Your right, i cleared the log in the second i started the cb run and made the Screenshot a Moment before the run finished.

Guess why? Not to fool someone... i just dont wanted my idle values to impact my load values. My Screenshot is 100% load time while cb was running to get exact MIN MAX AVG LOAD values. Difficult to understand...i know


My sys is not throttling and my clocks and temps actually are real world values...you just dont want to believe because they are better then yours... its easier to claim that i lie right?

But when i have a bit time i will do a 4.5ghz ruun with log and graph.

 

That is the correct way (by resetting the values right after you started the test) to record the values with hwinfo, otherwise your min and max max values are less relevant (especially since those values are not in sync for eg. you could have max clocks when idle on desktop with the high perf power plan, but those clocks could drop once you put 100% load, but if you read the max values, you have the max clocks that were recorded before the test, right up to the max load and the max power draw and so on) and the average values are less accurate.

Also, in synthetic tests that put 100% load, that 100% load should appear under all categories: current, min, max and average.

 

Thats why i've done it... if i did it the other way the people would claim i cheat my AVG with idle time... now its the other way round. You cant do it right.

 

Yes but CB always takes a few seconds to build up (maybe two or three) that always translates to your cpu outputting a higher value until the load settled down. I'm still not seeing logs bit just pictures of one second while a log shows the whole situation.

Locking a cpu to a certain MHz will keep it locked up to that MHz until you give the permission for it to clock down that's how computers work they are programmed to do what we say. If he cannot do that it means power (or voltage) is insufficient. So that argument does not play here. Especially IF he locked his MHz, and not just let the cpu on high performance (cause speedshift would still throttle it down due not having enough load and windows overriding third party software) if speedshift is turned off or set to 0 in throttlestop, there is no way it could clock down only if he doesn't have enough juice for that MHz.

 

i mean seriously ? lol ? you serious right now ??

because of that you became defensive like that ?

I'ts a pity that your view is so narrow, with something i meant your settings.
your temps show no core differentials so why in the world would i assume that your repaste was bad ?
haven't you heard of throttlestop and all it's various settings ?? i mean something could be wrong over there right ??

that sentence is not even talking about a specific component or interface how the heck do you relate something to your repaste ?

so if someone IRL says to you that something does not seem quite right you automatically think back to your fresh LM repaste right ?

mann i'm not even going to lose my time with you anymore. help is too rare to be taken granted so i'm gonna spend my time on someone who can listen and discuss like an adult instead of going back and forth.

you really think 0.1 V is gonna make a 2C+ difference nahh man. happy gaming and enjoy your laptop brother be carefull with it.

 

could someone remind me what bios settings I have to change to boot from samsung nvme drive? I already copied the OS to it but it says no bootable device. I managed to sort it out on the previous machine but not now.

 

CB can't make the CPU output higher clocks than the OC multipliers that were set (if I didn't misunderstood what you were referring to)

After you load the CPU, you either have the same clocks if nothing prevents them from being maintained or you get lower clocks if you are thermal/power/current throttling.

I was speaking in general, because in that case, it was obvious that he wouldn't throttle (he has enough thermal headroom, the limits are raised and epp is set to prefer max performance with a sst value of 0), but if you're testing a locked H CPU for eg. , then you would get the higher TB clocks before the test (because you had 0% load), then your clocks will go down in steps (firstly because of the way TB works when you have load on all cores, secondly because at least one limit will be triggered if the load is high enough) and you could end up with a lower nr. of relevant values.

In the previous case, the min and avr values would have been more or less affected (cpu load, temps, power draw) if he didn't reset the values during the test.

The part that I agree with, is that a pic of the sensors from hwinfo doesn't always provide sufficient data and a hwinfo log would be much better, but when correctly used, a sensor pic provides sufficient data most of the times.

 

my CPU can do what it want during idle... idealy clock down for better temps and lower power, as long as it holds his clocks during load

4.5ghz



4.7ghz



ok now?