*OFFICIAL* Alienware m15 Owner's Lounge

This is not true. At least for models like Dell. Look at this video, look at the minimum FPS difference in games. Those are laptops with same exact CPU and same exact cooling system:

 

Uh, that video seems to prove that the gap is only around 10%....most of the time even less.

Going by minimum framerate and not average as your reason to say it's not seems a bit disingenuous.

 

Well, the minimum FPS is what you will actually notice when your FPS will, for example, drop to 20 fps, unlike the time when your FPS will spike to 100 (which will inflate the average FPS number) ;-) But if you consider average FPS more important to you - hey, that's your preference ;-)

 

Well, he said exactly what I was saying, 10% boost in performance as for average FPS performance. Of course you are right as well as for minimum fps, its 19% by ownordisown. Anyway, my point was there isnt such a big gap between these cards, so being afraid of 6gb of vram is pointless, both cards can offer good and similar performance.

 

For those willing to get 90% of the performance in real world and draw way less heat consider:
970 EVO Plus 1TB
Pwr consumption
Idle: 30mW
Load: 5.5-6.0W (full watts please note)

760p (similar HP and Adata models exist)
Idle: 25mW
Load: 60mW

That is 1/100th of the power consumption in use, which leads to a lot lower temps in a confined space. There is a benchmark difference, but in real world load times, startup and overall use the difference is very tiny. I realize for most people here performance is the only goal. I just thought it'd be worth mentioning that for a small reduction in performance you can get way lower temps.

 

I think it's 80 or 90W, I'll confirm next time I see my friend.

 

Any idea if I should do these Alienware updates? I feel like windows update is more recent no?

 

That's the video that convinced me that the 2070 upgrade is worth it. For £150 extra I got 10% higher avg framerate, 20% higher min framerate, 33% more VRAM and 33% more powerful PSU (I didn't know the latter is probably unnecessary given how well the CPU undervolts). The £350 step up from the 2070 to the 2080 was looking much less compelling, even more so that I was worried about thermals (largely unnecessarily, as it turned out that, following standard system optimisation, the 2070 stays in low 60Cs under max load).

BTW for the sake of completeness: @ssj92 clocked 6070 in TimeSpy on the 2060. My 2070 pulls 7150 (+18%) on stock settings, 7500 (+24%) under mild overclock/undervolt, and 7800 (+29%) under max overclock. Stock 2080 clocks in at 8000 (+31%).

 

That's incredible. Does anyone own this and can confirm the power draw/temps?

 

I just recently started using time spy but for the sake of comparison, what do you get in firestrike stock & oc?

Stock 2060 (m15): 15,915 GPU https://www.3dmark.com/fs/19247376
Mild OC 2060 (m15): 16,817 https://www.3dmark.com/fs/19247445

Full fat 2080 (stock) [A51m] : 26,915 GPU
Full fat 2080 (Mild OC) [A51m]: 28,016 GPU

 

I don't rely much on FireStrike for 3 related reasons: 1. It's an old test 2. It's a DX11 rather than DX12 test and therefore unable to fully utilise current hardware 3. My system is notably cooler while running it (compared with TimeSpy), therefore the test is unable to push the system to its limits.

I did run it once or twice, from memory I got somewhere between 17k and 18k OCed (the result was above 4K Gaming PC). Will post the actual results later, including temp data to illustrate that FireStrike is mostly of historical and sentimental value these days.

 

I would go with these rather than rely on Windows update.

 

You will be OK most of the time with 6GB, since evidently you don't need 8GB per se (for some specific reason), just be aware that some games today use over 6GB on max settings, so you will need to dial textures down from time to time.

 

BTW obviously m15 2080 Max-Q scores would be more relevant here, otherwise I guess we can start having fun with some proper desktop results

https://www.3dmark.com/fs/17521021

 

So I am confused now. 970 Evo has termal problems but you are saying it consume way less power than adata? Something is not correct here. And to make things more complicated lets look at temperature test of adata M.2 here. It has some good numbers.

 

3 things: 1 - m15 with 2060 is now selling with 240W psu
2 - You are comparing overclocked 2070MQ and 2080MQ with stock 2060. If you overclock 2060 the gap is again close.
3 - 2060 in m15 is only card labelled as OC ready apart from 70 and 80. Just saying...

 

1. That's fine, 240W is only needed if one doesn't undervolt
2. I'm not sure if the 2060 TimeSpy number posted by @ssj92 was OCed or not, since the previous result in the video was OCed. If it wasn't OCed, what is the OCed score for 2060?
3. Fine, 2060 wins the OC Ready label contest

 

Exactly what I’m saying on the heat part! This was with a thermal pad before I got the heatsink

 

OK, so now we know:

https://www.google.com/amp/s/amp.re...ware_m15_rtx_2060_i78750h_90w_battery_review/

The 2060 pulls 6500 in TimeSpy heavily overclocked - 800Mhz memory OC sounds like a lot. BTW the review author's comments about overlooking the 2070MQ (and his overall conclusion regarding that variant) are incorrect. So, basically OCed 2070MQ (7800) is 20% faster than OCed 2060, which is about right looking at the difference in the number of shaders (+25%).

 

With heatsink ( i suppose you get original one from delll) what temperatures you get now?

 

16261 (stock), 16719 (mild OC), 16899 (max OC).

Looking at temps, this test doesn't stress the system as well as TimeSpy, therefore doesn't really respond that well to OC either (as opposed to TimeSpy and DX12 games), and my conclusion is "meh, DX11 games, whatevs" .

 

http://www.thessdreview.com/our-reviews/nvme/intel-ssd-760p-review-512gb/5/

 

Have you guys seen this? https://www.guru3d.com/files-details/reflections-rtx-technology-demo-download.html

 

I got it from Amazon actually. I posted the link one or two pages back. This was in a desktop just so you know

 

Hello guys,
I just received my m15 Oled, i7-8750H, RTX 2060. I started to benchmark it out of the box without any tweekings. To be honest it is the first time I do this so I'm not sure if I'm doing it right or how to interpret the results but here they are:

The battery had a 9% wear out of the box which seems unacceptable (Full charge capacity 82580 mWh instead of 90000).

The Unigine Heaven benchmarks were:
39.4 FPS with a score of 992 at 4k High Quality
138.2 FPS with a score of 3481 at 1080 Ultra Quality with Tesselation at extreme (whatever that is)
The card temperature was 86 degrees

With the prime95 stress tests , the max temperature for the cores were: 100, 100, 88, 100, 96, 95, 97.
They seemed to be staying in the high 80s to mid 90s.
I stopped the test after a couple of minutes.

What do you think?
Thanks!
I have to say that the Oled screen is gorgeous!
and that the 240w power brick is huge!

 

Well I may have to withdraw that 760p recommendation. It's been cool for me but under a heatsink and I only measured it with the onboard not a camera. However in doing more research to answer the question I came across this site which doesn't test a ton of drives, they do test temperatures both from SMART sensors as well as actual FLIR camera imaging while the drive is in use. The core basis of the Samsung getting hotter than the Intel drive is true, just not as big a degree from their measurements. I suggest you read up on several drives just to get a sense on what is "normal behavior" and then form some thoughts from there on what you might prefer. Apparently write temperatures are where things get hot. All drives throttle as some point set by the manufacturer, and the SMART sensors are low compared to what the camera was reading.
https://www.techpowerup.com/reviews/?category=SSD&manufacturer=&pp=25&order=date

Even if their methods aren't perfect, it's probably good information to have and keep in your "resources" list going forward. Enclosed drives in high write modes must give a lot of heat.

 

9% wear on arrival seems huge. Mine and a few other people's came 5% worn and Dell claims this is normal, but 9%?
As for the heat, you need to go through 3 steps:
1. Install Throttle Stop and undvervolt starting at -125mV
2. 65cfm or better cooling pad
3. Optionally repaste using Phobya Nanogrease Extreme or Kryonaut (and repad using Fujipoly or Gelid GP Extreme)

My 2070 doesn't get warmer than 65-67C under load (22C ambient). If it's just the battery that ends up bothering you, persuade them to send you a new one - it's dead simple to replace, and in general the laptop is very well designed inside from maintenance point of view.

 

The review suggested by @Papusan makes it clear that 50mW draw under load is BS.

 

Adata consumes less power than 970 EVO but a large factor. Heat is lower too, just not by as huge a factor.

 

You'll want to read this page on just why things don't test well with the 760p. What state the drive will sit in and what power it will consume (and what penalty of latency in wake up) you are going to get will depend on your tolerance and use case. So it won't suit all folks. But not all reviews dug into the reasons for the power state and they only described it in the language here because they didn't go back and retest all the comparing drives in the graph.
https://www.anandtech.com/show/12349/the-intel-ssd-760p-512gb-review/8

But you'll see it may well be one of the lowest power consuming drives at least in that class.

 

I'll be joining you in the OLED club but not until we get the 9th gen cpu's

 

Well, I am worried about 9th gen as for laptops overall. If 6 cores are marginal as for heat and difficulties with cooling, than what about 8 cores inside small space?

Anyway, 9th gen is already offered in m15 but with the highest specs and way too expensive.

 

Probably go 8 cores and shut off HT when going mobile. As demonstrated by the 9700K 8 non HT cores equal or exceed 6/12 HT cores. And when running on the AW Amplifier where the cooling only has to keep the CPU running, turn on the HT.

 

Thank you! Which temperature is considered normal for the cpu?
Is "138.2 FPS with a score of 3481 at 1080 Ultra Quality with Tesselation at extreme" in Unigine Heaven considered good for a rtx 2060?

 

If TDP is 45W it should be fine, the question is: will such a CPU actually run any faster than i7-8750H on the m15? If the architecture is still Coffee Lake or similar then unlikely. Check out Ultra Male's review of m15 with i9-8950HK. Executive summary: FAIL.

 

Well, CPU temperature depends on the power it runs at, as well as the cooling system and obviously the ambient temp (I will assume 22C). In desktops with 2kg coolers we would have even beasts such as i9-9900K run around 65C max; in a laptop like m15, even a tiny i7-8750H with 45W runs into high 80Cs under load, and quickly hits 100C and throttles when Turbo Boost kicks in on all cores and temporarily raises power limit to 90W. Apparently, this is adequate thermal performance for Dell :/

All is not lost though. If you follow the recommended steps, you should see temps under full load at 45W around 70C in balanced fan mode. These will start SLOWLY going up if Turbo Boost kicks in, but for me the max is 90-92C which is still not amazing, but at least there is no thermal throttling. When running less CPU intensive tasks like games I see 55-65C for both CPU and GPU and the laptop is barely lukewarm to the touch near vents.

I'm not familiar with your benchmarks. Suggest you install Cinebench 20 from Windows Store or the website, and free 3D Mark demo from Steam then run TimeSpy benchmark. You should be getting around 2500 in CB20 and 6k in TimeSpy out of the box. HTH.

 

You're gonna make me buy a RTX Titan for my graphics AMP aren't you

My TS score was stock run. I didn't get a chance to run OC but maybe next weekend if I get to play with it again I can try and see where it can go.

 

Thanks, just ran the tests and CB20 was 2640 and Time Spy was 5767 (5846 graphics, 5360 cpu).

With a -125 undervolt on cache and cores I get CB20 2865 and Time Spy 6083 (6035 graphics, 6372 cpu). Temps still in the mid 90s during the stress test.

Guess I should increase the undervolt? Only the cores or the cache also?

 

Skolar may I ask you about m15 with oled? Are there specific apps/software which protects oled from image retention or burn in? Like pixel shift/moving, or regeneration or for dimming static bright logos/icons? Thank you

 

Cool. Yeah, dial up the undervolt for both in 5mV steps. Mine runs at - 160mV. You will start seeing weird bluescreens unrelated to the GPU when you cross the line. You can also set the Intel GPU to - 50mV in TS.

Are you talking about average temps as reported by HWINFO or max? Is that with a cooling pad yet?

Your benchmark results are spot on, CPU better than mine in fact (on balanced fans).

 

Wasn't the score from SLI Titans? But then returns clearly do diminish, the test was probably CPU bound.

Cool, see if you can beat the 6500 from that reddit review. The key question with any OCed results is: is the configuration stable over a few days at least, which I guess you won't be able to answer based on a quick session.

 

Was 6500 the GPU score or overall score?

Not bad. I was getting 3088 CB20 with -100mV & 6072 TS overall with 2060/i7-8750H config

 

Overall score I think. That's what I've been posting anyway.

I'm not sure how people get 3k plus on CB20 with i7-8750H. LM? Some TB tweaks? I tried extending Turbo Boost time in TS but it doesn't seem to do anything. Silicon lottery I guess.

 

Not that I have noticed. But if you tell me what to look for exactly I can search.
The theme and wallpaper are dark out of the box.

 

I dont have this laptop yet and never had oled so I dont know, but as for oled tv this things are in menu/settings

 

I was talking about the temps I was seeing in HWINFO while CB20 was running. In max temps one of the core hit 100.

I do not own a cooling pad yet but will look into it.

Thanks for your help! I'll try tinkering with the undervolt a bit more tonight.

The battery seems to be draining really fast when not plugged in. I spent 10% battery for a half hour of YouTube video watching, screen brightness low.

 

Cool, I just take the average over a period by using the reset measurements button right before the test. Yeah, I'm not super impressed by the battery life either, but it gets the job done. I've tweaked the balanced mode for max. savings when plugged in and adjusted the new W10 power balance slider. I haven't installed geforce Experience, but that also has a battery saving tool (a frame rate limiter essentially).

 

Is anyone else experiencing this issue: http://forum.notebookreview.com/thr...-from-sleep-m15-possibly-other-models.828552/
it's very simple to test, just sleep, wake up, run CB20 and check wattage/clocks and score.

 

That is 5 hours of battery life, that’s pretty good.

 

Guess I was expecting a bit more of the 90WHr battery. I will test it more thoroughly tomorrow.