*OFFICIAL* Alienware M18x Benchmark Thread

so some one can beat me with my own setup? and then give some long in depth speech as to how they all of a sudden did it. with some detailed technical analysis (having nothing to do with nothing what so ever - ) / debate on amd white sheets and power constraints...
ROTFLMAO!!

let me think about.

 

Now there ^^^^^^ is dedication to benching/overclocking a laptop........you are indeed "as crazy as a box of frogs" (no offense meant btw) but hats off to you John....the rest of us are not worthy......

Side note: how come when I opened up my beast , I didnt get any of those fancy extras lol..........

 

none taken stevie.

well...because you didn't have them stashed away like i did...

 

Nice John, where is the insulation? What are you using? Or are you just using them to pull heat? No dice or ln2? Never seen anything like that on the left gpu and cpu? Mainly just pots

 

Haha...in all seriousness, WOW.....I am really impressed with how far you have taken it.....incredible stuff. Thanks for sharing your pictures....I am sure you will have just saved a few people from killing their gpu's in the hope to compete with your scores.....can I ask "why the HUGE pipe on the secondary and not the same as on the primary?"

 

Who am I kidding with the 4-piped HS....dammit, gotta cancel the projects after this, lol.

+1 rep

 

the 2 small copper blocks are from my k|ngp|in ln2 pot.(love those) the dice/water block pot is the tall one.
tried some dice, but you have to watch for condensation so i just sit them on the heat sinks and blow cold air. works alright i guess.

lol on the 4 piper...keep working that angle.

but like speedy and i were talking...dell went cheap on the gpu heat sinks..they should be all copper just like the cpu heat sink is.

side note:

didnt do insulation because i only tested some dice to see what was going to happen. you would need to insulate the card and under the card & cpu socket.not about to be doing all that there...lol

 

OMG! that is so amazing and don't worry i will never take the extra step to be a professional benchmarker

 

so it's the "meaningless" thread

 

Nope...its the "lets see if we can thrash these beasts to within an inch of their lives to get the best performance BENCHMARK results out of them" thread......hmmm, maybe I could PM a mod to ask for change of thread title? lololol - nothing "meaningless" about that......

 

Aiki.....there's another venture for you to go at.... - i'd probably be just as interested in those as I am the cpu heatsink....

 

No where near John's monster M18x but getting there... I can't crack 27k on Vantage yet but very close... the down throttle is really nasty...

 

Overpower ! lol

 

So john, those are just to add mass for absorbing heat away from the CPU/gpus? Just blowing AC on them?

Maybe I need to bolt my TRUE to my gpus haha.

 

Wow John you have gone way beyond me in modding to Bench ... Crazy. You must have repasted already yes ? I am wondering how I am getting the clocks I have been getting without any of that ... maybe they put actual 6870s in my system lol.

 

what is the gpu clocks?

 

Core: 721
Shader: 1442
Memory: 1587

Untouched voltage.
Both GPU maxe temp @74

Going to keep pushing but want to do a heavy stability test, so far perfect with Witcher 2, Crysis, Metro 2033.

 

yep. that's about it. and i was looking at bolting one on myself...funny you should mention that.

nope. i only pasted the cpu. i thought about doing them but...that was looking like a bunch of unneeded work.
and your clocks are only good for 3dmark11 or am i missing something?

 

Really nice work john! I'm amazed to see what you can get of this beast like that!
Do you think a better cooling really helps to push the clocks farther? I mean when you set the clocks the gpus/cpus aren't really hot so it shouldn't be the limiting factor. Then when running bechmark yes you need better cooling to keep your temperatures under the limit. But prior to that is it really what makes the difference (I can't get even close to your clocks).

 

???
Not sure I know what you mean, but yea I guess ?

 

Yup that pc cant be used for anything else than 3dmark11.
Are u using N2 for cooling the 2nd gpu??

 

You have 2x RAM, my test system only had 8GB. Second, you have a mild overclock and XMP Profile, - mine had none. Also, I have a 2630QM.

My point was to show the difference between the 1333 cas-9 vs 1866 cas-11 within the same system

 

Given all the difficulty in figuring out how to mount something like a TRUE over one of them hot parts, I figured why not....and have now built most of a water loop that will go on my laptop if all goes as planned. Since I had the majority of a desktop loop lying around.....

Will have a 120.3 BlackIce GTX Xtreme Radiator, 6x Gentle Typhoon 1850 rpm fans in push pull, and my MCP-355 with EK top/res. I need to figure out blocks and how I will attach them but that is all that is left! Will probably end up using chipset blocks or something similar and will probably mount them on top of the stock heatsink because otherwise I would likely chip the die and the vram needs some loving too.

Since I tore down my desktop recently....my laptop will have a better water loop than most desktops haha.

This all because I figured, as hard as mounting a TRUE will be....I could just as easily have waterblocks mounted (And I only have one TRUE.....)

 

cooling always helps when benching. sorry, but benching ais about the best tools for the job and if that means getting it cold...then that's what one does. outside of that. the lappy runs fine on it's own. i did 26 to 27k gpu right off the desk for vantage.

meaning....i didnt see you posting vantage or 3dmark06 scores.
win lose or draw...i post them all...even at 28.8k gpu, im still getting the stuffing beat out of my by a 580m...

no ln2, but did try some dice. running ln2 is far crazier than just running dry ice. it will freeze up the whole laptop in a matter of minutes. if your not careful.

just set the true on top is really all you need. and something heavy to hold it in place. those copper blocks you see...about 8 pounds each.

 

The TRUE is top heavy though, getting it to stay and be flat enough to make a difference will be a challenge. Any sort of interface material being used John or just sitting this massive copper chunks on the heatsink?

I am now on the hunt for used 2nd rate chipset heatsinks (need something small) lol and I do not predict the 1/2" ID 3/4" OD tubing I have being friendly to this project!

 

This talk of water loops and copper is making my mouth water. Now I want to get my Koolance V2 out and do some benching. I went dice cup straight on cpu die. Took a while to find the sweet spot since i can't bolt it down. Im working on that part. Subzero benching is shhaaweeet.

 

i got my water block sitting there as well...lol. no way to hold it down though.
i didnt touch the die, because i wasnt sure if i was going to upgrade my cpu to a 2.5 version. that was really the whole reason why i even tried this. the partial cold allowed me to gain 2k points on cpu and allowed me to run almost 4ghz across 4 cores.

 

What was your max temp on that run? keeping it under 50*C should allow 4Ghz+ on all four cores. At least with my 920 it allowed it for a full wprime 1024 run.

 

3982 was what it maxed out at 38.6 on the multiplier.

 

John, tried dropping that voltage offset (can go up to 25) in bios? May be giving more voltage than needed and could free up some TDP

I will play with it later.

 

that's close to my desktop 580's lol

 

holy cow man, wow!

 

that benchmark is flawed for ati cards.

 

Why? 10chars

Edit: what settings were you running during the bench? (gpu/cpu clocks & ccc config)

 

because forcing ccc tessellation can make or break that benchmark in terms of partial images showing up. making it run that much faster.

 

meaning that the bench itself should override the CCC setting for Tessellation accordingly with the test settings? (performance/extreme)

 

As far as tessellation goes, just leave it on to "use application settings" or the default value to 64x. I think they put that option to combat excessive tessellation in certain games.

 

that's cool for the truthful bencher..but not everyone is going to be doing that..lol

 

on this thread don't matter how we get the numbers just post some !

 

remember the cheating in benchmarks via LoD. disabling Tess just as bad : P. we're all guilty though. thanks amd .

 

Dude................

That is insanely sick. Ripping my desktops apart for years to install custom monster copper blocks and cryo drops is one thing. I don't think I'd have the heart to do that to my laptops....kudos.

 

thanks.

that stuff is just sitting there. none of it is bolted down.

 

Ya, I am still trying to figure out how I will hold down my waterblocks...since the tubing will make them want to spring up

Might be able to use zipties for the GPUs but at a loss on the CPU. Maybe I will just sit the TRUE on the CPU and only watercool the GPUs

 

you can run the zip ties under and around the card. then screw it back to board then tighten down the block. the cpu is another matter...

 

That is what I had in mind for the GPUs

 

i played crysis the other day...was doing about 50 to 70 fps constant. not to shabby i gotta say.

 

i hit 20fps around water areas

 

got those dips periodically as well. like right when it would save the game mostly.

 

What is going on with my memory... completely underperforming compared to yours: