*OFFICIAL* Alienware M17x Owners Lounge - Part 2

i'll have to look, but since it came out to now...these guys have been pushing these things to new heights. 30 to 35 fps @very high and the single isn't to far off the mark if im not mistaken..(don would have these numbers on the tip of his thumb)

but keep at them and see if your able to upgrade when they do come up for sale..(new nvidia's or the 5870's).the new cards seem to be using the same slot, but who knows....(speculations of course)

 

Yeah, the Crysis benchmark with a single on my old cards was literally half when in non-SLi.

9800m GTX - Stock Clock.
11 FPS average after the three runs.
23 FPS average with SLi.

Note: It did help when I got a faster CPU because SLi scaled better.

On the note of the possibility of 5870s on the M17x. There is speculation, so who knows. For now, I'd just be content maxing out the current hardware.

 

i got about 32 fps in crysis maxed out...i forget what i got on single card...9800m gtx's as well.
vantage gpu single was 5600 or so and dual was 11k

gpus over clocked and cpu stock.so i was pretty close in that aspect.

you running a dual core?

 

Yeah, the aforementioned benchmarks were on a dual core. I am still on a dualcore, looking to find a cheap QX9300.

 

a chip can seriously make or break a system!
im going through that now...trying to break open a low over clocking cpu.
i say grab an oem if you can. retail is just to damn high right now.

 

Good point, and I'll keep that in mind.

I am looking for OEM rather than a shoddy engineering sample. Thing is that with the i7 train taking full steam, there might be certain ex-R1 owners eager to get rid of their OEM QX9300s for modest price tags.

Edit:
I have to say, your Crysis score seems very high. Quite good and interesting. You're actually beating the 280/4870 dual cards when they're at stock by getting 32 average. What were your GPU clocks, exactly? And, which CPU?

 

i have that stuff on imageshacke.us...but they are acting funny right now and i can't see all my pictures...as soon as i can, i'll post a few runs i did...over clocked and stock. think i ran it also at dual core nd quad core settings. 2.4 ghz and 3.0 ghz as well as 1.6 ghz..can't be fore certain till i can see all my images.


veryhigh/stock 1080p

veryhigh over clocked 1080p

 

just a quick question, is it true that all 4870 m17x owners do not suffer dpc latency and stutters even if they have a qx9300?

 

Pretty much.. ^^^^ I never really experienced any stuttering every, despite high DPC latency.

 

so you still have high latency spikes that go into the red zone but not stutters?

 

Let me be clear about my system's performance.

Without Hybrid SLi, I am absolutely sutter-free in games. Even -with- Hybrid SLi, only a few odd games rarely sutter, so I can tolerate that much.

 

Hey guys, now ive had quite afew motherboard problems and as such dell will be replacing my m17x. they are replacing it with revision 2 of the m17x, because of that it will have dual 4870's compared to dual 260's, how much betetr are the 4870's? also it will have core i5 520 compared to qx9300. wil lthe hyperthreading in 520 make it comparable? or should i switch them upon arrival.
also is the qx9300 compatible with the new system?

 

Dam you are lucky....just tried to get my system replaced with R2 and it was a no go....I was citing DPC issues as the reason they should do it. Got 4870s coming to replace my GTX 280m's at least.

They will be a good upgrade for you

 

4870's are much better then the 260's. Now your CPU is a differant issue. the QX9300 cost you $1000. Thats an Extreme CPU. You should settle for nothing less then the i7 920xm which is an extreme CPU but also costs only $1050. The i5 520 is a base processor. It's as if your paying $950 for nothing since you already paid for the QX9300. You need to get on the phone. Afterall I'm sure you need extra CPU power for a reason since you paid for it. And no the QX9300 will not work in your i5 motherboard. Totally differant technology. I suggest getting on the phone before you let them sway you with the 4870's. If they say no escalate and ask for a manager. Be polite but be firm. emphasis the fact that you need this processor for work and if they wont budge, request a refund.

EDIT: The i7 920xm is the closest thing to your QX in specs. They should be matching your specs. The next lower priced processor is $450 less. You really do deserve the i7 920xm.

 

yea i had dpc issuse too, ontop of mobo issues. had a tech come replace mobo 2 times. so they just said **** it lol. now the i5 will be a downgrade i think form qx9300. but i put it in manually before and if the two systems are compatible i will switch. How much of a performance is 4870's vs 260's?

 

ahh i bought the 9300 after all was said and done and put it in myself... in this case i will be selling my qx9300 processor if anyone is interested send me a message.

 

ahhh, didn't realize that. Thats what I did too.

 

$500 OBO is what i am asking for the qx9300.

 

Now that's a cheap way to upgrade. Congrats! BTW XTU 1.3 is compatible with PM55.

 

You don't even need XTU now haha, bios vCore is INSANE

I was hoping to get the same upgrade path Mandrake haha, had to try didn't I?

 

Worth a try but hard to be disappointed with the 4870 upgrade.

Regarding XTU, the system has a max overclock of 15%. XTU may be able to push that limit.

 

15% ?! What is THAT bs? lol

Guess we only got 25% though on our R1 M17x's haha, rest was due to multiplier

And I am in no way disappointed, quite happy actually lol

 

It's not over yet, you know it! With 4870s it might still stutter
So you didn't get your 4870s yet? I got 1, the second is coming tomorrow.

@Mandrake: What drivers do you recommend for the ATI GPUs?

 

I just got mine ordered today, so no, aint got them yet. And the stutter has not been an issue for me for some time

DPC was never an issue, but on paper it looked bad enough lol

 

Cat 9.12 mobility mod'ed. More info in the link in my sig.

 

I plan to go there to learn to properly OC them as well

 

so then i can switch the processors?

 

id like to request for an m17x r2 instead of getting 4870s to replace my gtx 280s but the thought of having only 1 hour battery life and no integrated gfx chip keeps me at bay.

 

the i5/7 has integrated graphics doesnt it?

 

the IGP will not be utilized as the chipset does not support it

and they will not swap it to an R2, I tried lol

 

someday they have no other choice then to swap our m17x to r2 systems.
Same story with the good old m1730s.

 

just wondering, where do the 4870's in crossfire compare to 280's in sli? also are the 4870's gddr5 or 3?

 

yea some day they will have no choice but to swap.. id really like the r2 cos of the rgb led screen and the i7 but low battery life is quite a turn off..

on my end i think getting an r2 would be quite possible once it launches here since the 4870s arent even available for the current m17x here so they have to solve the problem one way or another..

 

@cheeseborger

4870s are supposed to be at least 15% faster at stock.. they are ddr3 but according to the people here the 4870s dont cause any stutterings hence everyone wants them

 

sweet because i have major stuttering issues. and even betetr that they tossed me the upgrade free of charge up from my 260's

 

At least 15% faster? Where does this mystical information come from? I would like a link to a formal publication that verifies this claim, and forum benchmarks--especially synthetics--don't count. Because most studies and publications--which are far too few since these are laptop cards--say they are neck and neck, some say 280s are--slightly--better and some say the 4870s are--slightly--better. And the general view on a cuda core vs. a stream core is not 1:1, but rather, 1:4; again a hypothesis by fanboys and not an official number.

Theoretical teraflops, the 4870 in crossfire should break the 1 TFLOP barrier (Ref: Official - AMD.com). A single 280 does 570 (Ref: Official - Nvidia.com), and in the same theory, its SLi setup more than breaks the same 1 TFLOP barrier. (Of course, even official TerraFlops are theoretical thresholds that can sometimes mean very little with respect to real-world gaming.)

The only solid view: 4870s don't stutter as much, and will generally perform the same when you are actually playing games. They will likely overclock better and create fewer complications; don't get caught up in this numeric hype.

 

I think most people want to see what this particular system with these particular cards can do. Theoretical numbers won't help them. Thus the benchmarking thread would prove to be more useful.

 

Stock M17x with 4870's vs a stock M17x with 280's with the same CPU show the ATI solution beating the Nvidia solution every time.. There were a flurry of benchmark's to this effect late last year when we started getting 4870s. I don't know on what grounds you can say that forum benchmarks don't count, you have some of the best minds here and plenty of like for like configs to compare. I'd argue you won't find better comparisons anywhere else online due to the sheer number of benchmarks that have been run.

No stuttering is HUGE Kade.. and I might argue that stuttering is related to performance as well. Stuttering is poor performance. The Nvidia solution is also based off of G92b which at this point is just old. Are you arguing the Nvidia solution is better? I think between the DPC issues and lower benchmarks it's pretty clear that Nvidia is not the king in enthusiast mobile graphics right now.

 

I asked this explicitly and saved the conversation, they are supposed to be gddr5...

 

Prior to them coming out dell stated that they would be GDDR5, that is where the confusion came from. They are in fact DDR3.

 

They're GDDR3. If it was GDDR5 it would destroy a 5870 in performance. As it stands now, the 5870 should only be marginally faster.

 

But they tested the GDDR5 4870s with the i7 systems and worked perfectly!!

...one could hope right?

 

GDDR5 5870? Now that would be worth the wait.

 

5870 is GDDR5.

 

Ahh, good, havn't read anything on them yet. Thats good news.

 

Awwww man that just.... sucks

I am still learning ... Dellienware reps are ill-informed to say the least... or born liars

 

But this is the good news if you want GDDR5

 

5870 is GDDR 5 but only has a 128 bit bus That's why I said its only marginally faster than a GDDR 3 4870 with a 256 bit bus. If we had a 5870 with 256 bit bus, I'd be jumping up and down in joy.

 

Nvidia never was the king in enthusiast cards, in my opinion.

Stuttering is huge? So? Where did I say it is or it isn't? I don't get it, but if others are getting it, they should make the move. Was there another point to bringing that up, or were you just supplementing?

Anyway, what I am arguing, I won't explain to you. Just go back to my post and read. Maybe the point will eventually sink through when people aren't sensitive about brand loyalty. Just one request: Don't label me as a brand loyalist in the process. I am not interested in Nvidia, but I won't use forum experience as the official word. It will certainly hold weight and potential for further study, but it just won't hold the final word until someone declares that final word. That is all I am saying. When people start remarking specific numbers with reference to how much one thing is better than the other, then they have to have such a leg to stand on; you're educated, and you know it doesn't work that way, even if you had an army of bonafide engineers doing the testing.

Simple: Benchmarks say something, you say? So? They don't say 15% they don't say 20%. They say all of the aforementioned and more. So is it 20% better? AT LEAST? Or is it 5% better, at least? Just recently, in the benchmark threads, we had a 4870 barely beat out John's 9800m GTX. 1 FPS apart in a truly hungry benchmark like Crysis DX10 very high - something that determines the grunt in a card. Whether you care to admit this or not is irrelevant, but it's fact. What does that mean? 9800m GTX is 5% faster than 280m GTX? Can I go about saying that? Because if you can call me on what I have stated, then you validate my rights to claim that 9800m GTX is 5% faster than the 280m GTX, because quite a few benchmarks have accomplished 29-31 FPS average with Crysis set at Very High in DX10. And since Radeon's get about 31+, and 280ms are 15% less, then 9800m GTX is clearly at least 5% better than 280m. Do you see what I did right there? I used theoretical numbers to infer a very misleading contention.

Nick, I challenge people when they actually boast direct numbers to establish something that cannot be established in every practical scenario. You're an I.T. professional, so take it from the professional angle. This is why companies don't boast percentage figures with such specifics.

And Mandrake, that was exactly my point with the theoretical performance figures. Because they actually don't mean anything, yet they imply things that might not show in real world gaming. However, neither does 'atleast 15%' because then I will expect a 4870 to literally win out in every scenario by a very noticeable margin, which isn't always the case. One can't have one set of theoretical numbers to support their conclusion while denouncing the concept of theory altogether, I am sure you can appreciate this logic. I'd say that it is a safe bet that in a majority of apps., 4870 will do better, how much percentage is a grey field and subject to different levels of scrutiny.



Re: GDDR5.

Joker's got a point again. GDDR5 is nice, but the 128-bit bandwidth will hit you at those high resolutions like a brick wall. 1024p benchmarks will be less effected.

 

Kade you just love writing short stories.

I would agree with this. There are a lot of factors that play into it.