l had a Latitude E6540 with an AMD 8790 controller as well the Intel 4600 one. It got water damage although the hard disc was okay; so my PC repair guy sourced a machine with what he thought was the same spec (apart from the hard drive) so that he could put the old drive in the new machine.
However, it turned out that the machine he sourced didn't have the AMD controller, only the Intel one. But he advised me to go with it because he said he'd done some research into the E6540 and discovered that the models with the AMD controller were prone to overheating and that the AMD controllers were liable to eventually fail as a result - so he thought the new machine would be more reliable as a result of not having the AMD controller. He said that Dell fairly quickly discontinued the models that had the AMD controller built in, and latterly only supplied the models that only had the Intel controller on the motherboard.
At first it seemed that the new machine was still able to do everything I needed to do, so I went along with his advice - but now I've discovered that although it can record true HD (1080p) videos, which I need to be able to do, my video editing software really struggles on the new machine - whereas it performed fine on the old one. It's just about useable but it worked far more smoothly before. I had tested the ability to record in HD before agreeing to go ahead with it but hadn't thought of testing the ability to edit videos.
So my question is: is he right that the machine is likely to be more reliable without the AMD controller and that the AMD controller on the Latitude E6540 was prone to overheating and eventually to failing as a result? Or would I be better off trying to source a Latitude E6540 with an AMD controller, given that some of my important software struggles to run without it?
There are complex reasons why I chose the E6540, which I don't want to repeat now as I've explained them in previous threads - but that's definitely the model I need. I'm just not sure whether I should go back to the version with the AMD controller or not.
I've been running laptop repair business for 15+ years. That's why I know.
Statistically almost 100% of dead GPUs brought to me are discrete. In my practice I have had only 2 faulty Intel GPUs among thousands of AMD/NVidia. But it does not mean that 100% of AMD will fail, because my statistics is on already failed laptops.
The bottom line is that it's an 8 year old system that's exceeded its design life by at least three years -- any remaining lifespan likely has more to do with how it's been used and maintained, rather than which GPU it has -- just as a 10 year old, 150,000 mile car that has a maintenance history can wind up more reliable than a 5-year old car with 70K miles on it.
I only stopped using my 13 year old Latitude D520 two years ago - and not because of any faults but because I needed a faster machine in order to make videos - and there are still a lot of D520s for sale on eBay, fetching quite decent prices; and there are loads of E6540s for sale there. So my original question stands. Was the AMD controller a cause of overheating and serious unreliability in the E6540, and was it removed by Dell for that reason in later iterations of the E6540, as my PC repair guy believes, and as the ifixit article seems to indicate - or is he misinformed?
I'm not clear what you're saying in the third paragraph. You seem to be saying that there was no systemic defect with the models that had AMD controllers as well as Intel controllers but I'm not sure if that's what you meant?
When you say "we" I'm not clear who you are you referring to? Are you referring to Dell?
Also do you know for certain whether my PC engineer was miniformed when he was told that Dell withdrew the version of the E6540 that had both the AMD and Intel controllers on it long before they discontinued the version that had Intel only?
Nothing has been publicized about the reliability of the AMD GPUs in these systems, that is correct.
What I am saying is that any systems still out in the wild that are eight years old are survivors. No matter what you buy, there's no guarantee something won't fail in the future. Notebooks are designed around a 5 year life cycle -- anything older is essentially like a 15 year old automobile - it may be working now, but it may* need repair within the lifespan you have planned for it. There are simply no guarantees on any device that is beyond its expected lifespan.
I think the conclusion is being drawn based on suspect evidence.
First, the vast majority of these have Intel GPUs only - the sample size of AMD-equipped systems is smaller, which makes the variance in longevity much larger.
Second, as would be the case with a car model equipped with both a performance engine and a base engine, two units could be used quite differently -- both in intensity of work, and in the number of power-on hours.
Finally, my suspicion is that the age of these systems means than any systemic defect -- such as the nVIdia GPUs of about a decade ago -- would have surfaced and had to be addressed (as the class action that nVidia settled on their faulty GPU design forced them to address).
ilian359
1 Rookie
•
28 Posts
1784
0
Posted August 18th, 2021 14:00
I've been running laptop repair business for 15+ years. That's why I know.
Statistically almost 100% of dead GPUs brought to me are discrete. In my practice I have had only 2 faulty Intel GPUs among thousands of AMD/NVidia. But it does not mean that 100% of AMD will fail, because my statistics is on already failed laptops.