Secondary GPU (7970M) appears to be dying in my M18x R2.

My Crossfire 'problems' started a few months back, with a bizarre tearing-like problem that caused part of the frame to be shifted a few pixels right, an issue that varied from game to game and that did not occur with Crossfire disabled.

Well, more problems started a few days ago. I started to get bluescreens related to the GPUs. It was one a day for a couple of days, which I just passed off as random events, not indicative of a problem (neither occurred while under any GPU load)...the day after, there were two bluescreens, and after the second, more problems appeared. The audio output stutters at random moments along with a system-wide slowdown while the audio stutters (it usually lasts a second or two). Furthermore, games appear to be stuttering or running at a low framerate all the time, despite Fraps showing a constant 60fps. The audio stutters are sometimes happening in Windows with no GPU load, although seem to be a lot more common while playing games. Furthermore, ALL of these issues disappear if Crossfire is disabled in the CCC. I haven't had a bluescreen with it turned off, and it's been on near constantly for a few days now.

A fresh installation of the latest GPU drivers (uninstalling old drivers, booting into safe mode and clearing old driver files and registry keys with Driver Sweeper and CCleaner, rebooting, and then installing the new drivers) has done nothing to solve the problems.

I believe this all points towards a failure, or impending total failure, of the secondary GPU. (I experienced a similar, if more sudden, problem with my old M17 with crossfire 3870s: one day it randomly bluescreened, then it couldn't load a single fullscreen 3D game without crashing. It worked if the GPU drivers were not installed, so I also suspected a secondary GPU failure in that case.)

I'm on holiday for a week, so I can't call Dell and ask for repairs. But I want to be sure that this is the issue. Are there any further tests or diagnostics I can run to try and confirm that there's an issue?

I'm also concerned about technicians complaining about the fact that I've repasted the GPUs if one comes over to replace the secondary GPU. Technically it shouldn't matter, as even if I did cause damage, I have accidental damage coverage as part of the advanced warranty, but could they still refuse to perform the repair under warranty if the issue is brought up? (I repasted with Liquid Ultra and placed a ring of MX-4 around each GPU to act as a protective barrier just in case the conductive paste spilled out, so it will be blatantly obvious to any trained technician that I've worked on it myself. I am 99.99% certain that the repaste hasn't caused damage, though.)

As always, the help is appreciated, thanks guys.

With any luck, they'll be out of 7970Ms, and I can convince them to replace them with 680Ms/780Ms...I'm sick of the problems I've had with AMD cards.

 

Have you tried putting the secondary GPU on the primary slot? Does it work?

 

They don't care about repasting your GPUs and neither does Alienware. That is simple maintenance like cleaning the screen or blowing dirt and lint off the computer. As long as you cause no damage doing the repaste the warranty is totally intact.

Sorry to hear that this happened

 

I'm not able to open it up and try that right now, for various reasons. It'll be several weeks before I'm back at home, but when I am, I'll try it out then, before I get in touch with Dell.

I'm just worried about the technician complaining about it, particularly as I've used a highly conductive paste.

That said...I really doubt the paste could have damaged it, could it? It's been about a month since I've pasted, and I'm fairly sure that if the paste had caused a short circuit (or if I'd done any other damage in the process), then it would have just blown up immediately. Hopefully the technician will accept that, if they do try and claim that I damaged it myself.

Or better yet, maybe Dell will just send me the GPU to do the job myself, but I'm not sure they will given that it's an expensive part and they'll want to make sure I'm not being fraudulent and that the job's done properly.

Still hoping that they're out of 7970Ms so that I can talk my way into a 680M/780M upgrade.

 

If you're worried, take the machine apart before the technician arrives and clean off all the paste. I did that once and the guy actually appreciated the help. Saved both of us some time too as I do take these things apart fairly quickly by now Parts-only service is a bit of a myth though as there are clearly defined user-serviceable parts and GPUs are not among them. I know some people here have been able to request POS anyway but they won't do it for just anybody.

As for them being out of 7970M, they might not have any at hand but that could actually give you the shortest end of the stick--back order. It could take more than a few embarrassing delays before the rep responsible feels sorry enough to suggest an alternative.

 

Oh good. Now it's completely dead.

Was just watching a Twitch stream, had a game open but minimised, and suddenly the system hard locks with a black/yellow striped pattern on screen. After about 20 seconds the sound stopped playing too.

Try to reboot, nothing. No POST, no beeps. Tried a power drain, same thing. It lights up, fans come on, but is completely unresponsive beyond that.

I'll try removing both GPUs tomorrow to see if it will boot onto the IGP. I'm praying it will (or I would be if I was religious), because I have uni exams coming up in two months and I absolutely cannot be without my own computer...I'll be pretty screwed if Dell can't fix it without sending it to their depot.

 

If I were a betting man, I would bet: (a) your paste job has absolutely nothing whatsoever to do with the current problem; and, (b) the technician that services it would not care or complain about your choice in thermal paste. In fact, if he/she is anything like the guy in my area they will be intrigued at how excellent it performs and want to try it on their own.

It's not inherently dangerous, it merely requires proper handling. If your mom told you as a kid to look both ways before crossing the street it would not increase your chances of being hit by a car. It would only increase your awareness of the need to be careful and thereby decrease the likelihood. The chance exists every time you cross the road, but the odds are in your favor when you mind your mama. Same applies here. Ignorance of danger can be fatal, but doing things deliberately and with the right knowledge averts disaster.

 

I've been there before and it does run on integrated graphics with both GPUs removed. Some reassurance FWIW.

 

And now it boots and works fine again?

Part of me wishes that it would make its damn mind up.

The issue seems to have changed slightly as well. It no longer appears to be 'stuttering' in a 3D application, and the audio stuttering is gone too. FurMark, however, seems to show what's really happening right now...running FurMark on a single GPU is fine. Run it on both, and although the displayed framerate is twice that of the single GPU run, it looks the same on the screen. The frame counter shows the issue: only even numbered frames were being displayed.

So, although the secondary GPU appears to be functioning (the system detects it and even reports its temperature/frequencies fine, and it heats up under load as expected) its frames are not being displayed.

Hmm...could the Crossfire cable be failing? Could that be producing these symptoms?

I know, what I meant is that I was hoping the failure was the GPUs so that it could run using the IGP. Rather than, say, some other random part failing that wouldn't allow it to run on the IGP either.

 

A failing CrossFire or SLI bridge cable can cause a number of problems. I bet you can explain the situation to Alienware Support and have them dispatch a new cable to you. It might solve your woes.

 

Bluescreens becoming more frequent now, happening 2-3 times a day, and are occurring with Crossfire disabled. Seems like it was just luck that I didn't get bluescreens with Crossfire disabled earlier.

I'm starting to think that my entire graphics array might need replacing. The bluescreens are infrequent and random enough that I don't think it's reasonable to remove/replace the secondary GPU and just see if it doesn't bluescreen. I wouldn't be surprised if Dell suggested replacing all of it, they wanted to replace the motherboard as well as the screen assembly when I was having screen issues...only reason I don't have that new mobo today is that the new one was faulty. Not to mention, with exams coming up, I'd rather get this fixed and out of the way ASAP rather than trying to tackle it one part at a time...

In all seriousness, do you guys think it's possible for me to talk Dell into replacing them with nVidia 680Ms or 780Ms? Their performance isn't why I want the upgrade, but the durability of the AMD cards concerns me (the mobility 3870s failed in my old M17 after about two years, but maybe I'm just unlucky) and I'm absolutely sick of AMD's poor driver support. (I don't want these 7870Ms to be replaced just to fail again in two years so that Dell has to replace the entire laptop as they're out of stock of approved GPUs...I love my M18x and wouldn't be happy to see it replaced with an 18 today, for example) If I wanted to try to sweet-talk Dell into that, what's the best way to do that?

Thanks guys, I appreciate the suggestions.

 

I would ask... the worst than can happy is they say no. Unless you get a support person that is a newb the person that assists you will be all too familiar with the poor track record of 7970M and you won't be the first person he speaks with that wants something more reliable as a replacement part.

 

I'll try arguing that replacing these GPUs now with 680M or 780M cards will cost them a few hundred extra at most, whereas if the cards fail again within the next two years and they don't have any replacements in stock, they're paying out thousands for a replacement system. Hopefully I'll get someone who will give in. Hell, at this point, I'd even be willing to pay the difference between 7970M and 780M replacements for the better reliability and drivers.

Actually, now I think of it, is there any chance the AW reps on these forums would be able to help? I'd rather go through them to sort out the repair than call some call centre full of people with little actual tech knowledge. Though, I am in the UK (or, rather, will be back there in a week) and I know they're US-based, so I'm not sure if I could do that.

If I do manage to work out a replacement with 680M or 780M SLi, what else would I need? Presumably the heatsinks need replacing with nVidia compatible versions, and I believe the cards need a backplate as well if I remember correctly. Would that be right? (Oh, and the SLi cable, obviously)

 

Yes, GPU, heat sinks, support plates and SLI bridge cable.

The North American Social Media reps here are pretty much 100% hands tied with EMEA customers. Based on what they have told us, they have zero control over Dell Support for EMEA. I think, at best, they can make suggestions and that appears to be handled only on an exception basis.

 

Okay, thanks for confirming that. Hmm...I wonder, why do the nVidia GPUs need support plates when the AMD's don't?

Guess I'll have to ring up and hope, then. Perhaps I'll asked to be forwarded to a manager or higher tier tech support, I doubt the guys reading from a script can do much. I plan on pushing for the nVidia replacements because I'm so done with AMD's GPUs...maybe I'm just unlucky with having two laptops with AMD GPUs fail, but their track record definitely seems worse.

The laptop decided to die again about 15 minutes ago. Similar to last time, but with blue stripes lines this time...once again, won't POST. I left it for about 10 minutes and tried again, this time it gives a beep code but I've no idea which one, had to shut it off after the first beep because everyone else in the house is in bed and I don't particularly want to wake them up. I'll see what it does tomorrow morning.

 

AMD does need support plates. All MXM video cards do. That's what the heat sink screws attach to. Some people call it an "X" bracket, but that's a slang term for the same part. It's a different part number for AMD than NVIDIA and they are made slightly differently.

 

I stand corrected. But on the AMD cards, the bracket seems to be part of the heatsink. At least, I don't recall there being any separate support bracket on the occasions where I removed the GPU. Perhaps I just didn't notice it if there is a separate one.

 

For both brands it is on the installed on underside of the GPU facing the motherboard. They are held on by an adhesive-coated rubber insulator gasket. The threaded standoffs poke through holes in the PCB that the heat sinks screws thread into.

In this photo of 7970M parts, the support plates are in the upper right corner. They have a white 3M paper peel-off shield that exposes adhesive on a rubber gasket.

 

Okay, yeah, those brackets are on the back of my 7970Ms, I just never noticed them.

GPU failure is confirmed. Laptop boots and runs fine using the IGP without either GPU installed. It won't boot on the primary alone, so the primary has failed at the very least. I haven't tried switching them as I have no paste on hand for the heatsinks, and I suspect Dell would want to replace both to be sure anyway. (After all, they wanted to replace the mobo when I had a faulty display...)

Will the GPUs be fine if I just keep them in plastic zip bags? I'll probably reinstall them for the flight home so they don't get knocked, but I'll keep them outside at other times. Hope Dell's fine with that, because I need a working system for the next week before they can come and repair it.

EDIT: Called Dell, repair arranged for Monday, they're going to replace the GPUs and the mobo. The guy I talked to can't do anything about an upgrade, of course, but he gave me an email for AW management to try and arrange something. Time to send them an email and try to butter them up...I might try PMing the AW reps on here as well, see if they can at least try to recommend something.

 

No, unless you haven't been monitoring your temps and have been seeing sustained 90+ C on secondary GPU for hours a day over the last month there is likely zero correlation between your secondary GPU failure and the new TIM.

Only one person that we know of so far on this forum experienced what they believe to be Coolaboratory Liquid Ultra induced GPU failure, Unitoyle, who reported getting a code "revision F" which is indicative of the back-plate grounding out or what in his belief is the TIM contacting the protruding metal surrounding the GPU itself.

It is likely just an unfortunate coincidence that your secondary let go now. This is an extremely common issue with 7970's, I think we even touched on its frequent occurrence in the Coolaboratory Liquid Ultra thread with the discussion of the importance of mitigating heat.

Luckily for you, it happened while you still have warranty left and you might just get lucky enough to HAVE THEM REPLACE THOSE GPU'S WITH 780M SLI.

That's what happened to someone else on here recently who experienced a 7970 failure as Dell was completely out both 7970M and 680M so they were inclined to give that owner the latest equivalent in stock.

The technicians won't even be looking at your TIM, I have had them change out failed 580M multiple times on the refurbished M18x R1 that my current R2 replaced (after 3 unsuccessful attempts of replacing the GPU's Dell's policy is entire system replacement, the problem with purchasing a refurbished unit is that if you experience hardware failure under warranty, Dell replaces the failed hardware WITH REFURBISHED HARDWARE, I could totally imagine the burned 580M coming out of my R1 and being sent somewhere for "inspection" and then placed on a shelf to go into someone else's refurbished R1 as a "replacement" as all of the replacement GPU's swapped out of my R1 were bad).

Just be cordial and nice to the technicians, it is highly unlikely they will say anything about the TIM. But if youre totally paranoid about it, I believe your other option would be to hunt down and install the factory Nvidia thermal pads and do a thorough cleaning of the existing TIM.

Look at it this way, youll probably get lucky and get 780M SLI out of it, AT THE LEAST you will have the opportunity to switch to 680M SLI. Make sure you ask to see if Dell will let you do that, switch from 7970M X-Fire to 680M SLI. Cite knowledge of the high occurrence of 7970 failure and that you don't want to experience that again as your warranty will be ending soon.

Edit: Just read your update above, surprised they are going to replace with 7970M. If I were you I would call and demand 680M as replacement due to the high failure rate of 7970M.

 

I know the issue is almost certainly unrelated to the TIM, I was just worried that Dell would use it as an excuse not to perform the replacement under warranty.

Oh, and it seems that the primary is gone. It won't POST with the primary card installed. Perhaps the secondary card had gone as well, after all, its frames weren't being displayed in the week before the total failure.


Unfortunately, I'm not lucky in the slightest. I received an email back from AW management today. Sadly, they refuse to acknowledge the higher failure rate of 7970M cards. They also claim that the 780M cards aren't compatible...they're offering a paid upgrade to 680M cards for £400 (the difference in price between them and my current cards), and while that's tempting, I'm trying to escalate it further to get a 780M upgrade, even if I have to pay for it. All the contact I've had so far suggests that the people I've spoken to - even the AW management by email - are apparently not authorised to perform any sort of hardware swap without the customer paying to do so.

So, the technicians on the phone aren't able to do anything about it, the management team I emailed can't do anything about it either. As I said, I've asked them for an escalation of the issue...a very nice tech I spoke to on the phone has also given me the address of Dell's head office in the UK. (Believe it or not, you have to contact the head office by post to escalate an issue or file a complaint...I was quite surprised by that.) At this point, I'm not holding out much hope of 780Ms or even 680Ms...I don't think I can justify the £400 for 680Ms when they're barely any faster.

I also tried PMing Porras yesterday to see if there's anything he can do, even if he is US-based I thought it was worth a shot just to see if he can do anything to help. Unfortunately, he hasn't responded. I think I'll send another polite PM just to make sure he hasn't missed it.

Ugh...I just feel like there's nothing I can do at this point. The guys I spoke to on the phone apparently can't do anything beyond giving me the management email or the head office address...its horrible sitting here not being able to actually contact anyone higher up. There is, of course, no guarantee that they'll even read my letter if I were to send one to the head office, let alone respond to it or do anything about it.

And here I was, hoping I'd get some good news on my birthday, of all days...

 

Just to let you know that Alienware reps on this forum are not able to send PM's so you will have to email that Alienware resolutions department.

On another note, I have myself experienced revision FF error after using Liquid Ultra on my 780Ms. The manifestation of that damage in the 7970Ms could just be what you have been experiencing. Then again, since Dell is replacing your cards, it is none of our business to second-guess what the cause of the failure might be. Just be sure to take extra precautions when using Liquid Ultra again. I made a little fence using regular electrical tape.

 

Is there any way to point one of them towards this thread, then?

I did email their management department, and was told the same thing as the reps on the phone told me: The only way I can escalate the issue further, the only way I could have any sort of exception made for me, is to contact the Dell head office in the UK by post. Apparently they only handle outbound communication, so you can't phone or email them...

That's why I'm trying to get through to one of the AW reps here, to see if there's anything they can do to try and help. I'm not sure what else to try. As optimistic as the rep that I spoke to on the phone was, I can't help but be pessimistic and think that whoever ends up reading it will just throw it away with no further contact. It's so damn frustrating that there seems to be no way for me to actually phone and speak to someone higher up without trying to go through the head office first.

Seems that the EMEA support is lacking compared to the US support. It's looking like I'll have to 'make do' with another pair of 7970Ms...unless a reasonably priced pair of 780Ms comes up somewhere. I'm so damned sick of AMD at this point that I'm actually tempted to spend the 400 quid for the 680Ms, even though it's more of a sidegrade than an upgrade in performance terms.

I did take precautions. I placed a ring of MX-4 around the GPU die (not on it, around it) to protect the exposed components from any Liquid Ultra that might squeeze out from the contact area, or any that I hadn't applied properly.

Honestly, I don't plan on using Liquid Ultra on the GPUs again. It didn't seem to make a huge difference. The GPUs seemed to peak at 80C under 100% load whether I was using MX-4, ICD, or Liquid Ultra. Even the stock paste was barely any worse. I think I'll be satisfied with ICD on the GPUs.

 

Reading your thoughts, here are mine, yes 400 lbs. is quite a bit of money (what is that $600 US?), BUT if you can get a pair of 680M you might want to consider it and here is why:

ShadowPlay

PhysX

New API Mantle equivalent drivers

G-SYNC (if you plan on getting an external monitor, or have thought about it for a while, DO NOT GET ONE WITHOUT G-SYNC TECHNOLOGY. This is going to revolutionize gaming.

Do you enjoy Bethesda's games? SLI works 100% better than X-Fire in Fall Out 3, The Elder Scrolls etc.

Sure, speaking purely in terms of synthetic bench performance 7970M is on par with 680M but that's where the similarities end.

Oh and also:

Average of 10 C cooler temperatures, likely the factor behind hardly any or significantly less 680M failure as compared to 7970M.

But thinking about it further, you could just have them replace with 7970 for now, and if/when they fail again in two years time, 680M, 780M, and even 880M will be significantly cheaper (if you can still find them, or find them new, ebay etc.) at which point you could spend the 400 lbs. at that time to pick up 780M or 880M. But at that point, 2016, Pascal will be out and according to what we have been told by Nvidia, Pascal is going to make everything earlier completely obsolete.

 

Honestly, it is damned tempting, and I do actually have the money to spare. I was about to pick up a monitor as well - so GSync is extremely tempting.

But...400 quid is a lot of money to me. I'm a student, and while I'm well off, I'm still limited in the amount of spare money I have to throw at tech, games, and the such. Certainly more limited than many of the people here that replace their systems frequently and buy a new pair of GPUs with each new generation. And...400 quid stings when the performance difference isn't great.

But I'd still be going for it, if there weren't two other issues holding me back: the existence of the 780M (I don't want to be spending money on older, slower GPUs) and that I still feel I haven't exhausted every possible way to get in touch with someone high up in Dell. The tech rep I spoke with on the phone seemed very optimistic about Dell's main office offering to swap my cards for 780Ms if I politely wrote to them about it. And this is the issue - if I pay for the 680Ms now, I lose the opportunity to try for an upgrade to 780Ms, whereas if I don't pay for the 680Ms now, I lose the opportunity to go for them. It's a gamble, really. The only reason they're offering them for 400 quid is because that's the difference in cost between 7970Ms and 680Ms, and they're swapping out my broken cards. That repair happens on Monday, with the parts ordered on Friday (supposedly).

I don't know what to do at this point. I feel as if there must be some way for me to get in touch with someone higher up in Dell and get a definite yes or no, and that I just haven't found that way yet. This is part of the reason why I've been trying to contact the AW reps on this board, who don't seem to be noticing this thread. They might not be able to intervene directly, but at the very least, I was hoping they might be able to get someone high up in the EMEA region to contact me. I just don't know what I can do to get their attention besides posting a thread saying "AW Reps please help me" that links to this thread, and I get the feeling that such a thread might be pushing too far...I might try it, at least.

 

The performance gap between 680M and 780M without a dual PSU set up is negligible, as in maybe 15%. TBone is running 780M SLI and his recent 3DMark 11 GPU score was 19.5k on a single PSU. I'm at 16.5k with 680M SLI and am on a moderate OC (+250 core/+500 memory @ 1.025 V). I could up the voltage to 1.05 V or evey 1.1 V and push the clocks to +300 core/+600 memory to close that gap but I don't want to see sustained temps of 80 C in certain demanding games. For additional comparison 3DMarkVantage usually show 780M SLI around 55k GPU with a single PSU. I'm sitting at 48k right now again with a moderate OC and could probably hit 50k with an aggressive OC. That is barely a 15% difference.

The record holders here, Johnksss and Mr. Fox are running dual PSU set-ups, and are able to extract most of the power out of these cards.

The real enticing thing about 780M is the 4GB of VRAM, but even then there are only a handful of games that exceed the 2GB of VRAM of 680M, although there may be more on the horizon.

The point: there is isn't much of a performance gap between 680M and 780M when both cards are overclocked. It's the same architecture with a few subtle differences is CUDA core count and VRAM. It isn't the difference between 580M (Fermi) and 680M (Kepler). Even 880M is still Kepler with 8GB of VRAM and again the performance gap between 780M and 880M is next to non-existent.

If you will lose the opportunity to get 680M SLI new, right now for $600 while you attempt to send a physical letter to some high-up in Dell with the off-chance that they might grant your request youre really better off going with the sure bet as the performance difference is next to nil.

Oh and just remember that Dell is going to try to stick you the price difference between 680M and 780M, it could end up being double 400 British lbs.

 

Having owned both, I would say $600 to upgrade to 680M from 7970M is worth every penny. Sure, 780M is better, but 680M was truly excellent and I considered the difference from 7970M to be remarkable.

 

Yeah I agree with Mr. Fox, consider this your opportunity to switch to 680M SLI for the price of one 680M on ebay.

 

Is the gap really that large? It was from a while back, but a thread of yours showed that the two were very, very close in games - in your own words, the difference was within the margin of error in most cases - so has something changed by then to increase the 680m's performance?

At the moment, it just feels like I'd be upgrading for the drivers and the features like gsync, and not for added performance.

 

If you run them stock, yes... very similar performance. This is probably the thread you remember: Clash of the Titans: GTX 680M SLI and HD 7970M CF. 7970M has really good performance for a stock GPU, but that's not good enough to suit me and the poor reliability factor is impossible to ignore.

I don't run them stock, and that's where their similarities end. They run neck and neck to a point of modesty, then 680M rips the jugulars out of 7970M and leaves it for dead... literally a blood bath for AMD when you flash a vBIOS mod and start overclocking and bumping voltage. I'm not exaggerating, and if you look at my benchmarks with them, there is no contest whatsoever. And, 680M holds up really well to punishment compared to 7970M cards that seem to just roll over and die from normal use.

Beyond that, the improvement in graphics quality, additional features, and almost painless driver management, are compelling. Both times that I switched from AMD to NVIDIA (6990M CF to 580M SLI, and then 7970M CF to 680M SLI) the improvement in colors and sharpness on my display were unmistakable.

 

Fair enough. I knew the 680Ms had good overclock potential, I just didn't know it made that big of a difference.

Well, my plan at this point is to try Dell a couple more times and see if I can get hold of a rep that can actually escalate me to someone high up. (It must be possible to take this higher without sending post to Dell's head bloody office, surely?) If that fails...I'll email back the manager and see if I can negotiate the 400 quid price down a bit. If not...well, I'll think about it. 400 quid is a huge chunk out of my disposable income for the next half a year, plus I was about to blow 300 quid on a monitor as well.

If I had a stable job with a reasonable income, then I wouldn't be so hesitant, because I do believe that it's worth the upgrade.

EDIT: Great. Got an email back, turns out the manager had made a mistake and the 400 quid was for one card. They want 800 for the upgrade...I'm not paying that when I could get a pair of 780Ms from eBay for that money. Definitely doesn't sound like the difference between the 7970Ms and the 680Ms, either.

Honestly...this is a bit of a farce. How can the only way to escalate this issue be to post a complaint to the head office? Gotta love the lackluster EMEA support...

Well, this seals the deal then. I'll have a new pair of 7970Ms put in, and send a polite escalation/complaint to the head office when I'm back in the UK.

 

Firstly, I apologise to moderators about this thread. I'm aware that I am breaking the forum rules by posting another thread on the same issue in the same sub-forum, however this seems to be the only way for me to be certain that an AW rep is aware of my problems.

I would like to politely request that moderators leave this thread until an AW rep has seen it and responded to either this thread, or my other thread. After that time, I am happy for this thread to be deleted or merged with my other thread.


To AW reps: I'm having serious trouble trying to escalate an issue with Dell support in the EMEA region. I am aware that you all operate in the US region and therefore cannot intervene in issues in the EMEA region as you would with US-based issues, but I am hoping that you can do something about it - simply getting someone high up in the EMEA support region to contact me would be much appreciated.

My issue can be seen in this thread. As can be read in that thread, I have exhausted every option for trying to escalate this issue higher, other than posting a complaint to Dell's head office in the UK. I've been told by several people now that this is the sole way for me to escalate it further - apparently I can't phone or email anyone higher up in the company, and the manager I've been in email contact with does not have the permissions to create the exceptions I would need.

I have already sent two PMs to Porras about this issue, however it seems that they have not been read. This thread is to make sure AW reps are aware of my issue. Even if AW reps on these forums are not able to provide any assistance at all, at least I know that my issue has been looked at.

I am happy to PM my system tag and contact details on request to any AW rep that needs them.

Thanks for your help, it is much appreciated. And once again, I apologise to the moderators for posting this thread - I hope you understand my reasons for doing so and will allow it in this case. I just want to be certain that an AW rep has at least read my issue, rather than giving up on it.

 

One idea is for you to sell these brand new 7970m's and get a pair of 680m's. The 7970m is a very good card, IMO better than the 680m, when not in crossfire/sli. The crossfire kills it thou.

 

Not an option, sadly. There's just over two years of warranty yet, and Dell probably wouldn't be very happy if this needs another major repair and the graphics cards are gone. Worse yet, they could refuse to replace the system (if it needed replacing in its entirety) because I can't give them the original GPUs.

(Of course, if I'm wrong to say that, then feel free to correct me)

Otherwise, I'd be happy to do that.

I'll see if any AW reps can help (hopefully they will see the thread in the subforum asking for their help) and/or if whoever opens my letter at the Dell head office is having a good day...who knows, maybe the optimistic tech rep was right and they might replace my new 7970Ms with 780Ms if I ask nicely. (Hey, I can dream...)

EDIT: I'd hate to use a 7970m on its own because of Enduro...it's well know that Enduro is quite far behind Optimus, particularly in terms of performance. A 7970m with Enduro can produce significantly lower framerates than a 7970m without Enduro, unless AMD has actually done something to improve Enduro significantly since I've seen those tests.

 

Alienware has no prob with you replacing the GPU. Not sure what they would do if the whole system has to be changed.

I have read that Enduro is finally working well, don't know for sure though.

 

Oh, I know that they don't care about swapping parts out, provided you don't break it. I'm just worried about what Dell will say if the system needed to be replaced if I didn't have the original GPUs.

 

I think the problem is you want a different card then the one you already have. Also is the 780m originally sold with your system ? I'm on my phone and can't see system specs

Sent from my SM-N900T using Tapatalk

 

Basically, yes. I'm not really happy receiving another pair of 7970Ms because of their poor reliability, and I'm trying to see if it's possible to swap them out for 680Ms or 780Ms. I can't actually recall whether the 780M was ever an option on the M18x R2, something tells me it may have been an option in the US late in the R2's availability, but in the EMEA region at least, it's not an official upgrade path.

The problem, really, is that it seems to be very difficult for me to escalate the issue to people who actually have the authority to create an exception. What seems to be fairly trivial in America seems to be impossible in the EMEA region. No rep, not even the managers, can get me in touch with someone higher up by either phone or email, and my only option is to post a letter to their head office. Either that, or pay full price for a new pair of 680Ms (despite the fact that they're replacing dead 7970Ms anyway).

I was just hoping that the AW reps could get someone higher up with authority in the EMEA region to get in touch with me, or give me the details of someone I could contact, or just ask them to create an exception for me. Or something. At the very least, even if they say they can't do anything, I know that I've at least tried to get help through them.

I wouldn't still be trying this hard to get the replacement if I'd actually spoken to someone high up who's said 'no', but I haven't. The highest level I've spoken to so far is a low-level manager who doesn't have the ability to make any exceptions or offer any discounts. He didn't say that Dell won't do so, he said that he cannot, whether or not he wants to.

 

No 780m's were never an option. If you get one of the AW reps to work with you, you might be able to see a good change but no promises. They will either try the 7970 again or they will swap both cards out for 680m.

 

An AW rep has PM'ed me, so if moderators wish to do so, this thread can be deleted/locked/merged with my other thread/thrown into the abyss. Thanks again.

Well, I'd rather have the 780Ms, but either way I really want to have these 7970Ms replaced with nVidia GPUs. I won't complain - in fact, I'll be ecstatic - if I'm offered a pair of 680Ms for free, or for a price I can afford.

 

I'll send the OP's information to the team in charge of Europe.

 

Thanks again Porras, it's much appreciated. You guys are awesome!

Once again, if the guys need my contact details (or any other information), let me know and I can pass them on, but if you passed my service tag on, they should have all my details on file anyway.

EDIT: Thanks for merging the threads now, moderators. Once again I hope you understand why I created a separate thread in the first place, and thanks for letting it stay separate for long enough to serve its purpose.

 

Yeah this sounds like a total farce, tell them the most that youll go is whatever the difference was between 7970M and 680M at the time of original purchase. If memory serves me I believe it was only $400 or $500 US, or about 250 British Pounds.

Sounds like they are trying to make some money off of you. To me 800 British lbs. sounds like the cost of two new 680M's from Dell not taking into account the value of the 7970's. It's like they are just trying to get you to pay for the different GPU's and are only willing to foot the labor cost. I'm sure youre aware that you can change your GPU's yourself.

Here's what I would consider doing if they continue to jive you, tell them just send you the 7970's, turn around and put them up on ebay for a deal, say $350 a card. There are plenty of older M17x R2 owners that will jump on them at that price. Then turn around and pay the difference yourself for 680M, which can be found for $500-600 if youre patient.

Dell is literally shooting themselves in the foot by trying to nickel and dime you as youre exponentially less likely to purchase a replacement system from the in the future. If it were the other way, say they offered you the chance to switch to 680M SLI for free or 780M SLI for a small fee, you would be infinitely more likely to remain a loyal Dell/Alienware customer.

Right now the bean counters are eager to make any kind of money they can as the global economic recession is threatening their livelihoods. If they can point to you and tell whoever is above them "see this good work I am doing, we could have exchanged this customers hardware for free but I managed to get 800 lbs out of him!" with the expectation that this kind of predatory behavior is going to keep them with a job then they will do exactly that.

 

It's hard to tell whether the manager I emailed genuinely wasn't allowed to offer anything better, or just was not willing to do so. In any case, he's a dead end.

The difference between 7970M Crossfire and 680M SLi was around $500 when I ordered the machine (and this is when the R2 was not long on the market), so yeah, I shouldn't have even been asked to pay 400 GBP. Definitely sounds like they were trying to charge me full price for brand new 680Ms.

I'll see if I get contacted by the guys that Porras passed my details to. If I haven't by Monday, then I'll post a letter to Dell's head office. I'll be polite, of course, pointing out (as you quite rightly said) that I'd be far more likely to be a repeat customer if they replaced them for free, or even at a small cost, with 780Ms or 680Ms. Not to mention given the number of people that view this forum, it's not unreasonable to suggest that someone could view my good experience (getting better GPUs for free) and decide that they should also buy from Dell/AW. Whether or not that will get me anywhere...remains to be seen. (I also plan on complaining about how much of a farce it is to escalate the issue, and how much better the support is in the US region compared to EMEA)

Sadly, the behaviour you described is exactly how they make money. As recent polls show (most notably, Comcast just being voted as the Worst Company in America, and EA in the previous two years), the most hated companies are also some of the most successful. I suppose America's ISPs are a different case, where there is very little competition and people have no chance but to be shafted by Comcast, TWC or Verizon if they want internet access at all, but keep in mind that we are outliers when it comes to tech: there are plenty of average people (i.e. technophobes) that will simply never buy from a company again if a product goes wrong, unless they offer exceptional support within the first 3 minutes of a phone call. (Many companies could learn a few things from Apple in that regard.)

As a side note, as much as I envy the better support you guys over the ocean get from Dell, I certainly don't envy your ISPs. It's incredible to read how badly the big ISPs shaft all of you, when I'm paying about $60 a month for reliable 75 up, 20 down, along with the mandatory landline and free evening and weekend calls, on an ISP that doesn't shape traffic or throttle certain sites like Netflix. It really shows how much of an effect real competition has on the market.

 

Yup, there's really little competition here in the US. It's contrary to successful model of the American Way and free enterprise, but some elected morons thought it would be a nice idea for the government to heavily regulate the industry. That almost never turns out well for consumers or service providers. In many places, including where I live, you have a choice between 1 cable internet/VOIP provider or 1 DSL/phone service provider. I am very fortunate in that my cable provider (Charter) provides decent speed at a reasonable price, no bandwidth metering, along with NO SERVICE CONTRACT... easy come, easy go... but ironically, no place to go unless you want satellite TV and painfully slow DSL while being lassoed into a contract.

I have a good feeling they are going to come through for you on the GPU change. They have nothing to lose by honoring your request since they need to replace them under warranty anyhow, and a lot to lose by making it difficult for you. I hope I am correct.

 

Although it is somewhat unreasonable to demand replacement with 780M, it is fully reasonable to ask for them to give you 680M at the price difference between that and 7970M when you originally purchased your machine. It does sound exactly as though they want you to pay full price for a pair of 680M while simply offering to do the labor.

And youre right about us getting screwed here in Amerika. One relevant example related to the ISP monopoly would be the wireless telecom oligopoly. An oligopoly differs from a monopoly in that multiple parties convene and agree on a standard, usurious rate to charge customers. What has actually happened is that Verizon, AT&T and Sprint, the largest wireless telecoms here in the "Land of the Free" did exactly this. So now if you want a wireless plan you can expect to pay between $75 and $100 a month for talk, text and data. My response to these shenanigans is well, I just didn't have a phone for the past 6 months and now that I am getting my Samsung Galaxy S4's screen replaced I will be activating a pre-paid plan with T-Mobile, $100 for 1k minutes, that I have a year to use. I intend to only use the phone for employment and emergency purposes as you can talk and text over the web (Google Mail etc.) for free.

"It's called the American Dream because you have to be asleep to believe it" - George Carlin

I also have Charter, and the connection is INSANE (6 ms ping, 60 Mbps download speed). One way to offset the cost of their service even further is by sharing the wi-fi with either your neighbors or roommates. The speed is that good. You can never bog it down and I mean never.

 

It should be scary to most people that Comcast has enough money laying around to buy up Time Warner Cable. It's also worth noting that the big ISPs want to move towards wireless (huge profits, less to invest in) and refuse to upgrade their own infrastructure. Verizon and AT&T would rather stick an antenna on your roof and give you a nice data cap than a hard wired connection. The internet also competes with their other service (TV). Of course, if you would like a different ISP it seems they are conveniently located as to not compete with each other.. Unlike what the federal judges seem to think, you can not just unplug from one and plug in to another.

Of course, they are charging us (the consumer) and then want to blame services like Netflix for their bandwidth usage and have them pay for the infrastructure. It's a mess.

Edit: My connection is rated 30Mb down and 1.5Mb up with a 150GB cap. I really want something with at least 5Mb up, but that would run me around $150 a month with Suddenlink.

Sorry for the rant, EviLCorsaiR. I've been following the thread for awhile, hoping it gets resolved for you. Back on topic, I think it is completely reasonable to replace the 7970M set with 680Ms. Especially since you are willing to pay the difference, as long as it is at a reasonable price.

Sent from my SM-N900V using Tapatalk

 

Well, I would like a free pair of 780Ms (who wouldn't?) but I'm definitely not demanding for it. I'm asking politely for a change to either 680Ms or 780Ms, whether it's free or at a reasonable price.



And youre right about us getting screwed here in Amerika. One relevant example related to the ISP monopoly would be the wireless telecom oligopoly. An oligopoly differs from a monopoly in that multiple parties convene and agree on a standard, usurious rate to charge customers. What has actually happened is that Verizon, AT&T and Sprint, the largest wireless telecoms here in the "Land of the Free" did exactly this. So now if you want a wireless plan you can expect to pay between $75 and $100 a month for talk, text and data. My response to these shenanigans is well, I just didn't have a phone for the past 6 months and now that I am getting my Samsung Galaxy S4's screen replaced I will be activating a pre-paid plan with T-Mobile, $100 for 1k minutes, that I have a year to use. I intend to only use the phone for employment and emergency purposes as you can talk and text over the web (Google Mail etc.) for free. .[/QUOTE]

Damn, I'd forgotten that your mobile providers are just as bad.

For comparison, I pay $22 a month for 200 mins, 5k texts, and unlimited 4G data. Yeah, $22. It is sim-only so I provide my own phone at my own expense, of course, but it works out cheaper in total over two years than getting a contract that includes a phone. Even the ones with phones are far better than in America - I could have had the same contract with my iPhone 5s for about $60 a month.

 

time Warner offers me 50/5 speeds which are 99% of the time what they advertise. No limits no contract. They even gave me a static IP (not really but it hasn't changed for like a year xD) all for 80$ a month



Att gives me and my family (5 others) unlimited calling and text and then 10GBs of data and a constant 5 bat lte connection with an average of 50/10 speeds along with this Nokia which I've come to love for its responsiveness and cooperation as well as atts Cooperation. Unlike Verizon and my iPhone freaken Verizon lost signal everywhere gave me less then 1/1 and the iPhone crashed every time you typed a B in the messaging app (something simple like that) and Verizon charged us more for what the worst phone I've ever had and the worst internet speeds in El Paso even my brother joels iPhone 5 had only 5/1 on lte Verizon is sad over here.



So I got tired of my iPhone and Verizon and the Nokia attracted me to att and I convinced my whole family to swap. I'm serious about my iPhone if you wanted to tell mr fox something in a pm it would roll over and play dead. My iPhone was the main reason I swapped. Sadly joels iPhone was just unlocked and swapped over to att he can only pick up 4G non-lte however the 4G is faster then Verizon's lte which is really sad so Joel doesn't notice a difference. I'm done with Verizon iPhone and android im sticking with att and windows phone

Sadly idk what att charges us per month I know its less then Verizon charges us though I manage the house network more then the mobile xD btw this nokias battery lasts all day till 3 am no matter how hard you work it





Sent from my Nokia Lumia 1520 Black using Tapatalk

 

Alright, I've cleaned off the GPUs and heatsinks in preparation for the replacement. I know people have said that the tech won't care about the paste, but:
a) I was worried about the possibility of Liquid Ultra residue being transferred to other areas of the graphics card and causing physical damage to it if the tech didn't clean it off before taking them,
b) I wanted to make sure the heatsinks were clean for the replacement GPUs.

I haven't heard anything more yet, so I'm assuming that they're definitely sending out 7970Ms and will be using my current heatsinks. I'll take my IC Diamond with me and ask the tech if it can be used, as I'm assuming he (or she) would just have stock paste with them to apply anyway.

Or maybe the heads of Europe's support team have sent the tech a pair of replacement 780Ms and haven't told me so that it will be a nice surprise. Hey, I can dream, right?