Alienware-L_Porras Please Help..... (M17xR4 Discrete Graphic not detected on cold boot)

Hello,

Back in September, I had a faulty AMD 7970M graphics which I have detailed in an earlier post. The summary is that it took 2 phone calls, an ill-equipped in home technician, and finally a return to the repair depot to get fixed. The drama started as the technician on the phone refused to do anything until I removed Windows 8 and restored Windows 7 (which had NOTHING to do with the issue). It was way overly stressful, and I lost the use of my laptop for almost 2 weeks due to this. After it returned from the depot, everything was fine, until yesterday.....

When I turned the M17xR4 on yesterday morning, the first thing I noticed was that it was taking a long time to boot (I didn't notice at the time, but it got halfway into windows, and then restarted again), and when it came up, it was in VGA mode (resolution was 800 x 600). I logged, in and Catalyst Control Center complained that no AMD hardware was detected. I rebooted (soft reboot), and when it came back up, the card reappeared, but did not load correctly, and was still in VGA mode. So, I hit Fn-7 to switch to discrete only, and when it came up this time, it looked normal. So, I hit Fn-7 again, and when Windows came up again, it looked like the attached screenshot. I rebooted it again, by hitting Fn-7 to switch to discrete only, and when Windows booted (again looked normal), I removed the Intel HD4000 drivers, and spent the next several hours driver cleaning and doing the Enduro driver dance. By lunch time, It was working normally, and fixed (or so I thought). Note, that the only cold boot during this was the initial one in the morning.

This morning I turned it on, and again noticed that the boot was taking a long time, and I saw it reboot this time (I missed this yesterday), and when the BIOS flashed again, I hit F2 to go into setup, and saw that the Discrete Graphics Card was "Not Found.". So I turned it off, and then back on, and went into the BIOS again, in it was again detected. So I booted up normally, with no issues. The problem is, that something is wrong, and I am not sure what it is. Since it seems to only be a problem with a cold boot, it sounds like metal contraction/expansion on the contacts, of the socket for the video card, but I am not a computer engineer, so it is just a guess. The work around it seems, is to just reboot after the first cold boot BEFORE windows loads, so that the drivers don't get into the weird state from yesterday. Regardless, I have to contact Alienware again to get warranty work done, but I dread calling. I do NOT want to have to restore Windows 7 again, just to have them look at it, when the operating system has nothing to do with it. So I was kind of hoping that maybe Alienware-L_Porras could offer some advice or something to expedite the support process, so that I won't have to go through the issues from September again.....

:-(

Rich S.

 

I know I'm no porras but have you tried reseating the card? I would try doing that before contacting support. Sometimes taking it out and putting it back in solves bad connection problems.

 

This is a probably a catastrophic component failure, not likely a need to reseat the card unless they left screws loose at the depot. Considering the GPU in question is 7970M it is more likely the former than the latter.

Get it fixed and then pull the working AMD GPU and set it off to the side for warranty purposes. Buy a GTX 780M and NVIDIA 100W heat sink to replace it and enjoy.

 

Some times I've had this happen. I would pull the video card, turn on the computer and get into the BIOS, and reset default settings, turn off, reseat GPU, and voila everything works. Even when my 7970M would no longer display anything, it was still detected in the BIOS.

Either way, looks like you've got a failed 7970M. Get a warranty replacement on it. I wish you had done a vBIOS dump on it, curious as to which version it was. I had an early 7970M that was prone to trouble, but this new one with vBIOS 022 is amazing and beats my old card in every way possible. It overclocks better and holds cooler temps.

 

Hi technos,

That is what is strange about this issue. On Saturday (the first time it happened), I didn't know what happened, and my driver install got whacked (I am running with Enduro enabled), causing issues with Windows 8.1 (on top of the card not being detected). Today (Sunday), I was little more weary, and was paying attention, and noticed the "cold" double boot, that resulted in my issues on Saturday, and stopped before Windows could load (salvaging the drivers this time). Once, the computer "warmed up", a subsequent (soft boot), the card showed back up, and is currently working. So, how can I figure out which vBIOS the card is using? Thanks!

Rich S.

 

Open GPU-Z and it will show you. If technos has a different (better) version, maybe his would help yours behave better. You could flash it using ATI Flash for DOS (don't use the Windows version, as it is famous for soft-bricking 7970M cards).

You can dump the vBIOS to a ROM using the little icon off to the right of where the vBIOS version is displayed.

 

Get GPU-Z to dump the vBIOS. If you are running on dual graphics, you may have to click the second tab in GPU-Z, since it will default to intel hd 4000. If you can run on PEG only, you won't have to do this. Anyway, dump the vBIOS (there is a icon on the right side, with a green arrow on it (as Fox pointed out). Save it. Download svl7's vBIOS patcher, and see if you can get it to open your vBIOS. If you get proper clock speeds and voltages, you have an old card. If you do not (you get absurdly high values for everything), then you have a "new" generation card. You can grab the vBIOS patcher from here:
Vbios Patcher - releases & changelog

Note: the vBIOS had no effect on stability with my old 7970M. This test is just meant to determine if you have a new card or old card. The new card will soft brick if you flash an older vBIOS to it, and you would need a bootable DOS usb + vBIOS 022 +atiflash to get it operational again.

 

Hi technos and Mr. Fox,

I ran GPU-Z and have attached the screen shot. I tried to save the vBIOS, and it said "BIOS reading not supported on this device." Which I assume is probably because I am in Enduro mode. The next time I get a chance, I'll try to switch to PEG mode only, and try that again. Thanks!

 

You definitely need to be in PEG only mode

 

I don;t know if it will be helpful but I'll share an experience I had, though I have an nvidia gt555M GPU.

I overclock my GPU and disable my CPu turbo boost when gaming. I forgot once to disable the CPU turbo boost when I overclocked my GPU, and played a physX havy game and the GPU overheated and disabled itself. Not matter what I did it wouldn't show again as if it was plugged out.

I found out in the bios yo ucan disable the Graphcis adapter. I did that, restarted, and enabled it back again, it it was luckily all good.

So if possible you can try the same through the bios.

 

Howdy Everyone,

A quick update. Today I popped open the laptop and decided to take a look. The first thing I noticed, was the screws holding both the graphics card, and the heat-pipe were lose. So, I took everything out, and reseated the video card , and re-attached the heat-pipe. I did NOT clean off and re-apply new thermal paste (as I don't have any), but I was careful. Buttoned it back up, and when it booted, everything seems fine. I am going to have to wait a few days to see if it is just a "fluke" that it works, or if the issue was the card wasn't seated properly by the depot. I'll update soon

Rich S.

 

Hi Everyone,

Well, I spoke too soon. The situation has gotten worse. This morning I turned on the M17xR4 and during boot it locked up with a red bar across the machine. After I turned it off (by holding the power button down), it came up "normal". I then decided to test it using discrete graphics only, hit Fn -F7, and rebooted. When it restarted, the screen was completely black, I couldn't even see the POST screens. However, I could hear Windows loading, and got the familiar start-up chime. So I blindly logged-in and hit Fn-F7 again, and it restarted, back to "normal." Additionally when I start some high-performance applications, the computer hard locks. So, I broke down and called up support. Long story short, they are sending out a technician to replace the motherboard and video card (again). I hope this one knows what they are doing :-(

As an aside, I ended up PM'ing Alienware_L_Porras, and have not gotten a response. Is he still around? Thanks..


Rich S.

 

Is this the second or third 7970M that has failed on you? They should replace it with a 680M or 780M GPU and just put an end to the madness. If they put another 7970M in it there is a remote chance you might get lucky with a rare gem. More likely than not, it will fail prematurely as they commonly do.

 

I wonder what the true failure rate of 7970M vs 680M was.
I mean theres gotta be some big quality differences between those two cards judging from the threads about dying 7970Ms?!

The same can be said about 6970M and 580M too

 

I am not aware of any reliable stats. Any brand of anything can fail because nothing is perfect. But, yeah, it certainly feels like way too many 7970M cards have died. We see a few examples of NVIDIA MXM GPU failures now and then, but not nearly as frequent as AMD MXM failures.

 

There are way more 7970M failures than 680Ms, but I'd say the 6990M and 580M failure rates are about the same. As on as you're in warranty, of course I'd let them swap out the 7970Ms. If you were out of warranty, I wouldn't take a gamble with another 7970M. I'd go for a 780M (unless you use professional programs, then the GTX series would be useless).

Did you check any of the revisions of the 7970M like I had mentioned earlier? I'm curious if you had "early" 7970M, or the later locked vBIOS 7970M. I'm willing to bet they put higher quality silicon in the locked vBIOS cards, I can get way better clocks on my new card in comparison to my old card.

 

The difference is that 6990M and 580M have generally seen a lot more wear and tear. 7970M seem to die prematurely.

I think you meant way better "clocks" but forgot the "L" LOL. I hope that's what you meant.

 

Dear me. I fixed that lol

 

Hi Technos,

No, I am not able to boot into single graphics mode. When I switch, I have no display at all. The screen is blank, I don't even see the Intel RAID option ROM load. the BIOS version on the card was 2012, so I'm not sure if that is telling. The 7970M that shipped with my computer "lasted" for more than 2 years, and I would have been SOL if I had not purchased the extended warranty, the second one was either bad from the factory (very possible due to the tech) damaged when it was installed. The third was installed at the depot, and until last weekend was working fine. So this will be the fourth one, plus a new motherboard. The tech on the phone also indicated the next step would be to reinstall Windows. I don't really think that will help, but I need to play the game to get warranty work. This is the second Alienware laptop I have owned, and the second one where the extended warranty helped.

Rich S.

 

I never bother with phone service. Of course installing windows won't help....PM a rep on here, or list your complaints and mail [email protected], that'll get you in touch directly with Dell Executive Support, which is a cut above the telephone guys.

There were a few vBIOS updates, all of which happened in 2012. The collection is here:
Video Bios Collection | techPowerUp

 

Hello Everyone,

The technician came to my home today, with a new motherboard and video card. Incidentally, this was the same tech that came out in September that had no idea how to work on an M17xR4, and damaged the video card heat pipe radiator as a result. However, I was on hand to help him with the repair. He said that we would first just replace the video card and see if that fixes it before doing the motherboard. After he struggled a little figuring out how to open the computer and find the video card, he proceeded to remove the card. When it came time to install the new one, I stopped him and made sure he attached the heatsink bracket (it was included this time, the last time we had to pry the old one off as it was not provided). After attaching the bracket, he struggled to insert the card in the socket (not sure why), but finally got it in. I stepped away when he was dealing with the thermal paste, but I am not sure if he cleaned off the heatsink before installing. I do know he pulled a tube of cheap CPU thermal compound out of his bag. He then reassembled the laptop, and we tested it. The card was detected on the first cold boot (a victory), and I was able to switch to discrete mode only (another victory). We then ran the built-in diagnostics, which passed. He considered it fixed. So after he left, I did some of my own testing, several benchmark runs, and I even ran through a couple of sections of Diablo III that caused issues with some of the other bad cards I had. Everything seemed to be working. In fact, I shut it off, went out to do some shopping and performed another cold boot, and it is still working. It looks like it is fixed again 

In the future, I think it would be better to do as Technos suggested, and bypass telephone support. However, I PM'd Alienware-L_Porras, but never got a response. Should the problem resurface, perhaps I should drop [email protected]? Thanks!

Rich S.

Rich Selmon

 

No techie is ever touching my baby! My privilege alone haha

 

+++ 1 on that; I learned the hard way when a tech called three times and did more damage each time; I then had a 'mare experience getting a replacement unit as I had bought the original from a reseller.

The ultimate outcome was good though as the replacement machine I received about three weeks ago has a better CPU, a massively better GPU and a simply superb 120hz screen.... I've since installed my two 256gb SSDs (one 2.5" one mSATA) and added a 512gb Crucial as HDD2, and installed Win 7 64 as I just can't get into Win 8.1, so overall am a very happy bunny.

After what happened to my original machine I'll never let anyone get their paws on the new one - when you consider that the original issue was cosmetic degradation on the palmrest and the trail of disaster that followed I don't think anyone would blame me!

I'm UK-based and we are usually envious of the level of support you guys in the US get in comparison; I had Luis Porras doing his best from the US, but EU tech are a completely separate unit, and eventually went to the top which had the desired effect! I went directly to the President of Dell Europe who actually took the time to respond personally which surprised me, and the support manager he passed me on to has mailed regularly to see how it is going, so a good turnaround in terms of service there!

Loving my new machine which is working flawlessly!

@LinkRS; Is this the third instance of repair you have had? I may be wrong but I'd say that if your machine is still under warranty it is "Three strikes and out" time, and you request a replacement, and if you don't get satisfaction then you should drop Michael Dell an email.

I'd suggest you use a similar method that I used when contacting his European counterpart; don't make the mistake of coming across as angry, or aggressive, write as though YOU personally want/need HIS help to sort out problem you're having; not to punch him on the nose!

You are disappointed that the quality of the techs work has failed to remedy the fault (and the tech is probably a subcontractor paid per job, working for a company paid by Dell; that is how it is in the UK) and ask Michael what he thinks he may be able to do to assist you, and restore your faith in Dell/Alienware as a premium brand.......

It worked for me and I know it has work for others on both sides of the ocean,,,,,,,

Good luck

Dave

 

Hi Dave,

Yes, the Dell technicians are a sub-contract here in the US. It might be different if you live in Round Rock, Miami, or perhaps Austin, but elsewhere it is not Dell employees. Which is why I allowed my M17xR4 to be shipped back to the "depot" last time, as I assumed that they would at least be Dell employees. This is Radeon 7970M number 4 in this laptop since I purchased it in July of 2012. The one that shipped with it "lasted" just over 2 years before failing. The second one was either damaged by the tech (possible) or simply DOA, the third was installed by the depot and it went almost "2" months. So, now I have number 4, and it was installed by the same tech that damaged the first one. I still have warranty out till 2016, and if the card fails again, I am going to at least request they swap it (GPU) for a different model, as 4 replacements (I think) is excessive. My M17xR4 came with some options that do not seem to be available in the current versions such as a blu-ray drive, Sound Blaster audio (which on my M17xR4 has been rock-solid), and mSATA slot for SSD caching (preferable to me due to how small most SSDs are compared to spinning disks), so I would like to keep it. I will have to see how it goes, thanks for the advice !

Rich S.

 

No worries Rich,

FYI the Alienware 17 models can be fitted with a blu-ray drive, the Dell part number is 6M7XC, and the LG part number is a CA40N, the 2013/14 17s also have an mSATA slot - my system is now all-SSD; HDD1 is a 256gb Samsung 840 Pro, HDD2 a 512gb Crucial MX100 and the mSATA is a 256gb Crucial M4; I'm waiting for an external enclosure to reuse the 750gb hybrid that was the HDD2 and is now spare... The audio on the current model has Klipsch speakers and Realtek HD audio, with Dolby Home Theater V4, I've not heard the earlier Alienware audio so have nothing to compare with, but mine sounds pretty good!
The guy who 'did' my system (in the truest sense of the word!) managed to damage three palmrests, incorrectly fit a motherboard and lose several screws; I was picking up broken palmrest clips for a week afterwards!
All resolved satisfactorily now though

Hope you get a result with your repair/replacement

Regards

Dave