*OFFICIAL* Alienware M18xR1/R2/18 Benchmark Thread - Part 3

I'd have to go looking for them with Google, but there are already examples in the benchmark thread. They should not be too hard to spot, because I usually fill out a description field for my benchmarks and that shows up in the Futuremark search results. So, you can find a few examples of my stock 3DMark runs that way as well. Maybe after work today I can post a coupe of examples. There are not a ton of examples available because I don't really care that much about stock performance. I normally run benchmarks to see how far I can push things beyond normal. I don't have 880M, so I cannot provide any examples of that.

Edit: By "don't really care" this is not to imply stock performance does not matter to me. It does matter, because I play games with the GPUs running stock. When I say "don't really care" it is relative to the benchmark score and not caring what the benchmark result looks like stock.

 

Dark9T thanks for showing us your results. Do you think you could do a 3DMark 11 benchmark for us when you have time ?

 

3dmark 11 1280x720

those benchmark scores are high or normal or low?

 

Well i was expecting more from the 290! 1000 points on Diver over my 7970.. i know they are just rebranded but still....

 

How many points on your other benchmark with stock clock?

 

That's pretty good, actually, assuming it's stock. Crossfire tends to scale much better than SLI because of AMD's compute technology. The results you posted above would suggest that you're seeing about a 40% increase in performance over a single GPU, which is impressive, considering it's a re-brand.

How are the temperatures and whatnot?

 

i got 22106 with Sky Diver

And 12067 with 3D Mark 11

 

Thanks for doing that bench. Seems about right. I guess the best way to test R9 m290x to the 7970m would be to see how much OC is left it in..
Whenever.. if ever you feel comfortable I'd be interested to see how it goes

 

It disappoints me that the name in gpu-z doesn't have the name of the card that there selling.. I mean awesome score for stock but if i'm buying some thing thats supposed to be a 290x id at least like em to take the time to have the name reflect the card..

Wait a minute.. why does the first one say 290x and the last bench say HD8970m?

the first bench i didnt look at the name in gpuz just the specs on the card on the last one i saw the HD8970m so i assumed it was the same card.. are these benches from the same card? and does that imply that you can flash a 8970 to a 290x?

 

This is a Windows and/or GPU driver misinformation glitch. It's probably something messed up in the Catalyst driver that is misidentifying the GPU based on hardware ID. Remember the "6900 Series" and "5800 series" showing in Device Manager and GPU-Z? For example, it did not matter if you had 6990M or 6970M, it generally showed the same and you had to look at things like shader count to determine which GPU is installed, so this is not that uncommon.

You can actually do this on purpose by modding the INF. Works for NVIDIA also... for example, you can trick Windows into thinking 780M is 680M or vice versa by modding the INF. But, it does not mean you can use the same vBIOS.

 

did you ever run the fire strike and Heaven Benchmark 4.0?

 

@Dark9T, do you have links to your online results?

Edit: never mind... found them... I know you are not planning to overclock, (which I hope you will reconsider,) but judging from your physics scores it looks like you need to do some adjusting to get your CPU to function at stock clock speeds.

Sky Diver: http://www.3dmark.com/3dm/3399206

Fire Strike: http://www.3dmark.com/3dm/3401779

3DMark11: http://www.3dmark.com/3dm11/8468246

Question: is your CPU a 4900MQ, 4810MQ or something else? Your signature shows a different CPU than CPU-Z or Futuremark.

 

41c to 77c max

 

i got the 4810MQ but not sure why it's show 4900MQ on futuremark, i wish i got the 4900mq! lol

img host

 

I imagine the CPUz misinformation is just another glitch - just like GPUz

 

Probably because stock speed of 4810MQ is equal to the 4900MQ. Although the 8MB L3 cache of 4900MQ should let CPUZ it's not a 4900MQ...

 

Well, two out of three for CPUz getting the L3 cache and spec right....even if it couldn't quite manage the name properly. Maybe they are playing catch up on the new hardware.....

As for wishing for a 4900mq over a 4810mq, the only difference is the L3 cache, which isn't going to impact gaming performance greatly, but would prove handy for video rendering and the like.

4900mq vs 4810mq via Intel

 

Don't forget the 4900MQ also has 6 OCable turbo bins vs only 4 available for 4810MQ.

Honestly 4900MQ is probably the best chip to get. 4930MX is hit and miss and runs like a blast furnance. 4900MQ will still allow you to OC to 4.2GHz; 4.3GHz if you go for 4910MQ. So unless you plan to go beyond those speeds and know how to manage thermals, I'd settle for a 4900/4910MQ.

 

Well, so much for all of the ludicrous chatter from know-it-all guys that like to whine about rebrands. Clearly, R9 M290X is better than a 7970M with a new name. That "just a rebrand" thought process didn't hold true for 880M either. This is why I am a "show me" person.

The good news it that it looks like the R9 M290X is actually fairly close to stock 780M SLI. It is clocked higher and still has a slightly lower graphics score in 3DMark11, but it's getting within striking distance. As you can see, I slowed down my CPU to match your CPU's physics test results, so this should be about a close as we can get comparing them.

So the single most important questions for me about the new AMD option will be: (a) how well does it overclock; and, (b) have AMD resolved their reliability problems. Looks like a good option right now based on stock performance, so I am really curious how well it will handle overclocking and that's going to be the make it or break it element for many enthusiasts. However, as a less expensive option for someone that only intends to play games and never plans to do any kind of overclocking, only (b) will matter.

The score for 3DMark11 is going to be lower than what should be expected because Windows 8 does a rather poor job of handling CPU performance. This OS shortcoming shows up very prominently in 3DMark11. I ran it in Windows 8 because Dark9T is running Windows 8.

3DMark11: Compare Stock 780M SLI versus Stock R9 M290X

Fire Strike: Compare Stock 780M SLI versus Stock R9 M290X

 

I am using win7, mr. Fox

 

Thank you so much for doing this for me!! Mr fox
I think your laptop is faster than others laptop with same the spec!

 

Oh yeah, I see that now. I thought looking at your screen shots that I saw some Windows 8 UI ugliness going on, but looking at it a second time I see that fantastic Windows 7 UI. Maybe it was the tacky 3DMark color scheme that caught my eye. It also shows Windows 7 it in the benchmark results, DOH!. Well, that helped level the playing field for the CPU differences pretty well.

It should not be too much different than any other Alienware 18 with the same specs since nothing was overclocked. It might be a little bit faster because I don't allow most of the garbage Windows Updates to infect my laptops.

 

I didn't know some window update can slow down my laptop, which window update are you talking about?

 

This one can, and it is required for newer versions of IE. [WARNING!] KB2670838 - You don't want this Windows Update! (I don't use IE for anything if I can help it, so I really don't care if it is updated or an "insecure" browser.)

Many Windows Updates are applied with zero discretion to all systems and this can contribute to instability.

 

Remarkably close results...

4930MX wPrime Benchmarks	Spoiler
3920XM wPrime Benchmarks	Spoiler

 

ROFL its so sad sometimes I still activate windows update to get some files and upgrade which is the easiest way... it's sad indeed but rest of them, ALL disabled. sometimes I even have to switch on manual in service to be able to check update but if not disable quick enough it'll reactivate itself and install a bunch of crap without consent, for sure.

 

I have Windows Update Automatic Updates disabled in Group Policy Editor, so Windows does not have permission to check or install any updates. I can check manually if/when I want to, but that's the only way.

 

heres one thing i discovered though, if Microsoft security essential is used, only way to update it seems to be windows update, service have to start. even if left in manual instead of automatic, after it update itself it'll then install bunch of other crap so need to disable quickly lol

 

Yes, that's correct. I used to use MSE on my M18xR2 and still do on other machines. But, I use nothing now... been many months now with no Windows 7 antivirus software.

 

I would at least get MalwareBytes Pro. Not a proper AV but good enough for people who know what they're doing. Completely unobtrusive except the weekly reminder to update definitions and even that can be turned off. Very light on resources as well. Saved me from god knows how many drive-by downloads when I ventured into the darker corners of the Internet.

 

I've been freeballing without antivirus for about 4 years and haven't had a single problem. Compared to when I actually used a antivirus diligently and seemed to always have system issues. If I were to put my tin foil hat on I'd say half of this viruses put out there are from the very same companies getting rid of them.

 

I just been using MSE + Malwarebytes + bit of common sense....no probs here....I stay away from 'dark corners of the Internet'

 

Well, let's just say my 300+GB worth of movies, TV shows, and music didn't just download themselves from nowhere.

 

Haha, I was wondering what was in those dark corners LOL

 

I have Malwarebytes Pro (lifetime subscription) but I have never actually needed to use it. But, because it is excellent when you need it, I have it. Maybe Murphy's Law is in play here. I have needed it and used the free version more than once in years past with excellent results. Now that I have actually spent money on it, I suspect there's a great chance I may never actually have any use for it going forward, LOL.

You might be right... one way of convincing consumers to spend money to protect their systems is to create a compelling reason to spend money. I realize this is a conspiracy theory, but my theory is panning out to be one that is similar to yours. I used to download stuff from questionable sources (mostly game NO-CD and Trainer exe files) and even strong protection solutions had glaring weaknesses. I think its probably accurate to say that a PC's worst enemy is its owner and their web surfing/email habits. Fast-forward a few years, based on my own PC use habits, the need for it is minimal and the risk appears to be very slight... a calculated risk I am taking without any adverse implications thus far.

 

I just got the GTX760 SC for my two years old desktop, looks like the GTX760 SC same as a laptop single 880M..

 

Haha, wow!

 

My son's friend has an HP desktop with a fairly stout AMD cpu and a crappy GT 530. Given his PSU and limited case space I steered him towards a EVGA GTX 750 Ti SC.

Now he plays Mechwarrior Online on high settings at 1080p. Before it wasn't playable at the lowest settings. He told me he was on the verge of throwing that system out! I would say that was $150 well spent.

 

Firestrike Stock - 7631

firestrike at 925/1400 - 8190

firestrike at 945/1445 - 8315

 

which driver and overclock softwares did you used?

 

14.6 i believe and MSI Afterburner.

 

These results are without a skewed view of SSD performance involving system memory caching...

Just 100% pure, unadulterated SSD read/write performance without fluff to make it look better.

 

Hello, I've been trying to look around, but haven't come to an answer yet.
Does anyone know what thickness of Thermal Pads to get for the video cards?
I remember talk of there being thermal pads on Clevos with the same cards but not Alienwares, so I was going to go ahead and add some in, and repaste while I was in there.
I was thinking 0.5mm thickness but I've seen others say 1.0mm

 

The 1.0mm thickness is correct. You can buy sheets of thermal pad from FrozenCPU.com if you don't already have a source selection.

There is one tiny square that is 3.0mm thick. You will see it when you take it apart. If that pad is still soft and moist, reuse it. If it is dried out and rough looking you can stack squares of 1.0mm pad to replace it.

 

Awesome results, Brother Severna!

One thing I really hate about Fire Strike is the fact it gives almost NO CREDIT for CPU performance. That really sucks.

With a MUCH HIGHER physics score your overall Fire Strike Extreme score is just barely better than mine. All of your graphics tests are within 0.2 and 0.6 FPS of matching my 780M SLI result, so it looks like your 880M SLI setup is finally doing a great job. Your physics score totally destroys mine (about 12 FPS more and 4,000 points higher) yet Fire Strikes gives credit for only a measly 141 points difference in overall score. This really negates any value of using Fire Strike as a reliable measurement of overall system performance in my mind. Look at this side-by-side comparison... Fire Strike is a joke.

Compare Fire Strike Extreme Results: Alienware M18xR2, 3920XM and 780M SLI versus Clevo P570WM, 4930MX and 880M SLI​

Same deal with Fire Strike... major blood bath for my XM Quad against that Hexa-Core (again 4000 3DMarks), but minimal difference (200 3DMarks) in overall score.

Compare Fire Strike Results: Alienware M18xR2, 3920XM and 780M SLI versus Clevo P570WM, 4930MX and 880M SLI​

 

Brother Serverna, how did you get your 880m's to behave?

 

Hey all!
Just played alittle with some OC on my new AW18 and the GTX 770m SLI

This is the current highest I´ve managed to go so far on the GPU and still be able to complete the whole firestrike run.



Is it the TDP value that is limiting me ?
in this test I´ve not OC´d the CPU at all, I´ve not been able to get Firestrike to run with both CPU and GPU OC´d but individually they run ok.

I did not think that a 4810MQ and 770m SLI should eat up all power ?

 

It is still a shame how they made the new 3DMark suite. Those new scores are better, but still not much higher than my M18xR2 because they more or less totally ignore your physics score when calculating the overall score... I guess maybe they did that to keep the AMD CPU fanboys and the low-budget QM CPU gamers from feeling bad about having comparatively poor CPU performance. Otherwise, I cannot think of any reason they would make a system benchmark so gimped and worthless as an indicator of overall system performance.

 

how long can you run Intel Stress Test with 4.4Ghz?

 

It's funny that the Asus "Republic of Gamers" won't let their competitors compare their systems with Asus on their leaderboard. Maybe they have something to hide.

Looks like their best recorded result from a notebook is "63754" on the RealBench V2 Notebook Leaderboard and here is a run of Asus RealBench on my M18xR2 with a score of 74830. This does not seem to do much with GPU... mostly a CPU test with boring things like encoding.