P370EM Xfire, or P170SM 880M GTX, which would you keep?

Hey guys, so i am in a lucky position to have two pretty good gaming rigs, but i am un decided on which to keep, specs are below, my gaming has picked up a level recently, mainly due to my two young children growing up a little so i have now more time for "me"

Specs below, now i am talking about longevity going forward, i don't want to be upgrading for at least another 12-18 months, i game mainly between 1600*900 to full 1080p,

P370EM

i7 3740QM 2.7ghz (full load temps of 86 c)
2 x 7970m's in crossfire (card 1 full load temps of 77 c, 2nd card 85 c)
16GB Corsair Vengeance (2x8gb)
Hard drives are the same in both 1 x 750GB SSHD, 1 x 750GB Scorpio Black, 1 x 60GB MSATA)

P170SM

i7 4700mq 2.4ghz (full load temps of 73 c)
8GB 880M GTX (full load temps of 84 c)
16GB Kingston hyper x
Hard drives as above

Opinions?

 

Thats a good situation to find yourself in

I'd go with the P370EM since 2x 7970M outperforms 1x 880M by quite a bit. It would be tough to see any difference in the CPU for gaming.

 

Yep its a pretty good scenario, the only thing that i lean towards the P170SM is the future upgrades if i chose to, such like 980m GTX,

think you are right, for some reason i tend to lean towards the chassis of the p370em, although the trackpad is terrrible

 

Sell both and get a batman with 980m. Lol.

 

+1000000.. Or sell both in Q3/Q4 and buy the new clevo with desktop CPU + single/dual GPUs

 

I woul say p170Sm, more compact, better battery life, no CrossFire hassle, 880m is good for another 18 month in my opinion.
in term of raw power, the 880 is 15% less powerfull than a Cross, but Cross is soooooooo unpredicatble.
So , i say P170sm

Elipsus

 

Both machines can take 980M (EM just needs BIOS/vbios mod, no big deal).

Personally I'd dump the P170SM. 880M is a terrible card. The list of people who have not had problems with them is incredibly small compared to the rest of us (I had two pairs fail on me in less than 6 months of ownership of my 9377).

 

P170SM, I had a m18xr2 with crossfire 7970M and while it was a good machine, at this point in time it's getting long in the tooth.

 

i would take the p370em. take both 7970m and sell em both then use the money to buy a 980m. that way you can buy another 980m when you get funds together.

 

Or two 970m cards would be a nice upgrade too.

 

Not entirely true. The largest issue with the 880M's are they are absolute crap for SLI mode. Far too many issues with overclocking, driver throttling and timing issues make SLI not play well with this, at all.

However, my 8258-S which has the 880M I've been able to crank just fine. Have had this laptop now for a year and 2 months, and the CPU still runs between 68-82c depending on the game (Bioshock Infinite pulls the 82c). I've not had any throttle issues that I've identified, and it's been performing remarkably nicely.

I think the majority of people who complain about the 880M fall into three categories.

1. Those who never updated their geforce drivers and have a broken/buggy one.
2. Those who are trying to tweak the hardware settings of the GeForce card itself (undervolt, overvolt, overclock, whatever)
3. Those who are unfortunate to have it in an SLI configuration.

Does this make the 880M a bad card? Potentially. It depends on your use-case.
SLI? Bad juju.
Overclocking? Bad juju.
Single GPU without modifying? (two thumbs up)

For single GPU configurations and if you don't tweak it, it seems to run just fine.

 

There have been more than enough complaints about single card systems like the Alienware 17 to say that isn't true. It sounds like you won the silicon lottery but you're the exception, not the rule.

 

Most people I've heard complain on the Alienware site were ones who were attempting to overclock their GPU, so they'd fall under #2.

I've not heard many complaints on Single GPU configurations with people who do not modify the hardware.

 

Really? Because Pidge from nVidia purchased an Alienware 17 and was hitting 94C out of the box with insane throttling as a result. Again, you're the exception not the rule.

 

Then that sounds like a bad paste job. I have the 880M, never get much above 82c under load. Wife has the 870M, she rarely hits over 79c.

If that was high thermals, I'd question the paste job and the type of paste more than anything else.

We could put the rumors to rest and ask XoticPC and other distributors what the average failure rate for 880M configurations were, if they were SLI mode, if they were related to overclocking, and how many occurred in single-GPU configurations.

That'd be better than conjecture or blogs where a small handful of people complain.

 

Is it really hard to accept that people who know video cards have said to avoid the 880M at all costs, even if you plan to keep it at stock?

In order to get my second pair working right, I had to undervolt and underclock them to 850MHz, effectively making them just slightly over a stock 780M and they still ran hotter than the 780M would have. Worked great like that for awhile and then started to artifact and crash which is when I threw my hands in the air and fought for a 9xx upgrade.

Anyway if you want to argue the point further, we have plenty of 880M threads around here. Let's stop derailing the OP's thread.

 

I'll throw my hat back in on this,

I decided to go with the 7970m p370em,

Mainly due to the insane what the 880m started to produce (obviously this is my own single experience and its the only 880m I have owned)

Very carefully pasted with gelid GC extreme, and making excellent contact to the heatsink on lesser games and benchmarks it never creeped above 80 ish

But before I decided to which one I was keeping I installed a bunch of benchmarks and games on both

Running 3Dmark after two runs through fire strike it hit 94,

Same on furrmark after 15 mins or so

No matter what I did on the 880m I never felt totally happy with those temps

The 7970's in xfire on 3dmark after two run through maxed out at 85 on the 2nd card and 76 on the first so I felt a hell of a lot happier

 

Well in future you have some nice upgrade options, but enjoy what should be some very nice performance for now

 

You keep saying that, but have they released any info about it?

 

I have to disagree. My card throttles so much now, to the point that I cannot run at full clock speed. My thermals are always at 93 degrees throttles nonstop, and no matter how much I undervolt and underclock, it always hits the same mark, and I have tried all drives from 330 to 350. I've listed my experiences on this card ever since I got it and it has been terrible. Good thing I'm getting rid of this thing very soon.

Perhaps you are one of the lucky ones.

I'm glad OP made the right choice.

 

That sounds like an issue with your heatsink/fan or paste job.

 

Check the 3Dmark11 in my sig of the 880m that's at room temp of 22c and the gpu never exceeds 69c on the benchmark

As for gaming temps it can range from 55 degrees playing LoL to 82 degrees playing Dying light on full settings while she is pulling 4.5 gigs of vRam

I have never experienced any sort of throttling on that card whatsoever.

As for OCing I have hynix vRam that reaches up to +750 MHz without sweating but the core is just too sensitive due to power consumption I can barely get stable +47 MHz on that

But in my opinion people exaggerate the whole 880m situation

 

Then again the core is already running pretty fast for it.

 

Yeah but the card runs just fine out of the box. On the contrary to common belief.

OCing will require a 330w power supply that's all because I am already giving my 240w a run for its money.

This ain't the card fault nonetheless

 

It could also be that nVidia fixed whatever the issue was finally. I had four of those problematic cards (two sets) and my issues were definitely not exaggerated, I had graphs and everything to back up what I was experiencing and I was far from the only one. It's great to hear that your experience has been positive but don't discount the experiences others have had.

 

Oh no I never discounted the experiences of others; I faced the furnace effect of that card too. It used to hit mid 90s c in dying light before the repaste "never throttled though to be frank" but really the repaste helped with the temps immensely.

It is indeed a very hot running power hungry chip but not the worst thing Nvidia done in the past decade after you help your system in accommodating it.

EDIT: I have to mention that I noticed my system came with a vBios different than the ones used in earlier AW17 machines. That combined with the machine not throttling even when it hit 90s made me sure than Nvidia/AW certainly changed something regarding the clocking policy of the chip.

If someone wants a copy of that vBios to check what have they done, I would be happy to provide.

 

You probably have the vbios that is in the 880M thread in the Alienware forum. But for nVidia to take almost a year to fix the issue is unacceptable. And it still makes no sense how much hotter the card gets than the 780M did, especially since the 780M can do 880M clocks with a slight voltage bump but the 880M still runs 10C or more hotter at the same clocks.

 

The default voltage of the 880M is higher which has an impact.

 

780M does 880M clocks at 1.025v, 880M does them at 1.018v, 780M runs 10C cooler.

 

Is that from personal experience?

 

Unless 780m can run Dying Light/Dragon Age Inquisition at full settings with 99% GPU load on 993 core clock & tons of vRam utilized @ 70c then it doesn't run 10c cooler than the 880m because I run these games sometime for 12 hrs non stop and never exceed 80-82 based on room temp.

 

I find that extremely hard to believe unless you're using liquid ultra and a modded vbios. Make a log for an hour with GPU-Z and post it here along with a GPU-Z screenshot please? I would like to see this.

 

I use Zalman STG-02 which made a load of difference on actually both my GPU & CPU & lol I most certianly don't have to lie about temps. So to avoid giving you a hard time believing as soon as I get home today & game I will post a 3-4 hrs log not even 1 hr log playing maxed Dragon Age Inquisition

 

You must have had no thermal paste if that junk brought your temps down from the 90s, its a mediocre thermal paste on desktops with high mounting pressure but its low viscosity makes it even worse for laptops with low pressure mounts and tends to pump out.

Anyway, you don't have to worry about the logs, it was just a request to see if your vbios has indeed fixed the throttling issue so I could make charts and put it in the 880M thread and see if others with buggy cards get an improvement with it or not.

 

I had the crappy dell adhesive thermal strips; for some reason they don't use paste at all ... Both my GPU & CPU used to hit 90s ... Now after using the STG-02 which is actually the only paste I could find in my country other than the generic 50 cent white non branded crap they use here ... my CPU never hits 90c on stock clocks no matter how hard I try unless I OC the 4 cores to 4.0 GHz & my GPU never goes past 82c ... On another note I see above that you mentioned that ur 880s ran @ 1.018v ... Mine actually peaks @ 1.0v and 993 core clock maybe that's another reason

 

Was not aware that Dell used thermal adhesive on high end cards. Yuck.

Nvidia's stock vbios ran the voltage that high (master card was 1.012v and slave was 1.018v), svl7's mod dropped it to 1v but his mod would have me hitting 94C within an hour with GC Extreme paste and liquid ultra would hit that point after three hours while the stock vbios dynamically throttled itself so it ran 80-82C with an average of 950MHz core which caused stuttering as it would drop into the 800s with some games (Watch Dogs, Sleeping Dogs) or any time I ran ultra high AA because it would hit the power limit and throttle. I actually used to run them at 850 @ 0.875v in order to get them to stop passing 80 under load with the modded vbios so they didn't throttle anymore. Four cards that all behaved the same.

 

Don't know why I suspect that they may "have taken a pointer or two" from svl7 vBios ... I found it very odd with my vBios that even at 95c the card didn't throttle at all which is very dangerous & now that you mentioned that his vBios ran @ 1.0V & now their bios runs the same voltage.

I think they may have stole his work

 

Except his does throttle at 95C lol.

Wouldn't surprise me if they used his modifications in their code but disabling the thermal throttle is reckless... But that shows how much of a problem it was for that GPU to hold its clocks. Default temp target was 87C I believe. Way too low, lots of throttle.

 

Maybe it did throttle at 95c I just wouldn't let it ... As soon as I saw my temps goin in the early 90s I used to stop whatever I was doing & let the chip cool ... The lack of safety measures dazzled me and I had to protect my investment .

What really pissed me off was the joke of paste that DELL used as well as the crazy fan profiles ... the fans used to stay @ 0 rpm till the temps reached 60ish where they go to 2k rpm & then go full speed at 80ish where already the laptop chassis gathered a serious amount of ... Which is simply unacceptable & stupid especially for a CPU & GPU that are known to be as hot as ovens. Even though it took quite sometime, effort and lots of tinkering; I am glad I sorted "their" mess out and finally able to enjoy my purchase.

 

Yeah you know what Clevo did with my P377SM-A? Fans used to kick up as soon as you start gaming then at 87C they hit as high as they are programmed to go (about 80% of max I'd say) but then with 9xxM they make it so the fans are dead silent until 90C now. You have to force max fans if you don't want your cards to heat up and max fans is like 15db louder than the 80% auto

 

Just when I thought DELL engineering ingenuity was unparalleled ... CLEVO closes in

 

Well to be fair, most systems don't have cards that get remotely that hot. It just seems to be my 980Ms...

 

Maybe prema could look into editing the EC for you to change the fan profile for some beer money. Worth an ask.