You’re lucky one 
well, it can be that I am unlucky one. Can be. But (besides this goes kinda off topic, but not really complete off topic since hardware was involved in initial question) let me explain of what’s happening in more general than just my own or your own personal consumer experience. When I want to produce electronics (let’s say I pretend to be a vendor), I start with chips. Let’s start with something simpler than motherboard - the principle is the same - let it be, well, a USB hub. I go on chip market and check for available chips. I pick something, let’s say Texas Instruments - great! That’s a name, and I can at least expect that it will be well documented and work 99% as documented. I download chip datasheets and USB Standard definitions and in principle I am ready to produce my own hubs. I wouldn’t even need a working kit for such a task. In the datasheets I see “recommended wiring” diagram. Now I have many options. I can follow the standard precisely and install power supply unit capable of providing 500 mA current per port (900mA for USB 3.0 afair). Or, I can decide - oh, well, it is unlikely that user will be charging devices on all the ports at once (besides formally speaking it CAN happen, actually), and install weaker (=cheaper) one. I can decide not to implement per-port switching (=less components onboard, whoa, I can exclude switches from part list!), I can exclude ESD protection circuits for ports (they ARE recommended in reference diagram, but we’re smarter engineers, than those in Texas Instruments, right? We’ll do quicker and cheaper), I can install electrolytes instead of expensive solid state capacitors on output, I can drop couple bucks here, few cents there… and formally it will still work. When you’ll use it for keyboard and mouse and never unplug them - it will work for years (remember, core chips are TI, so they WILL work), but it will be nowhere near really reliable and rock-steady design. Now, having those bucks and cents saved from “redundant” parts, I can install fancy radiators. In case of motherboard I (name me ASUS, if you like) can wire 6 ports instead of 4 on Intel board. 6 is bigger than 4, right? And I’m still even cheaper than those losers in Intel! People buy my shiny boards! PROFIT! Intel respects itself and can’t follow the same way, so just quits motherboard market. That’s what happening.
GTX 470. akdjshfdhjasflkhjasdflkhjasdflk hjalsdhfj la dhfjsasdf
I use Nvidia(750M) because I work on Linux systems. My experience of using ATI on Linux was terrible.
I am also using Lenovo Ideapad Y500 that comes with a GT 750m(2gb dedicated and 4gb shared memory)
Seems to explain all the crazy issues I see with some newer motherboards I have. This one, about 5-10% of the time just decides that one or all of my USB ports aren’t going to work until a reboot (preferably a cold shutdown). And it’s not an OS level problem–sometimes they won’t work in the BIOS settings, assuming I can get in at all. Once in awhile, even after a cold boot the BIOS screen will just freeze up for a few minutes before booting. Another motherboard I have seems to ignore the first keypress on GRUB’s OS selection menu (I guess that could be a bug, or “hardware specific”).
Even with expensive ones, motherboard advanced feature implementations can be half-assed apparently-- Ex. buggy or broken IOMMU, so good luck with high performance virtualization even though the MB’s specs claim otherwise.
Maybe I’m too demanding or unlucky, but from what I’ve learned about how low level electronics allow for “wiggle room” here and there and what you say above, I’m not too shocked. With the cheap stuff I can understand, but for very high end hardware, the damn things should WORK.
Yeah, that would well worth a thick book targeted at everyone who want to be “advanced user”. Unfortunately, English is not my native, so that would be too epic task for me, but if somebody would decide to write one, I would be happy to provide my humble observations and statistics… but that wouldn’t fit into this topic of course, nor it is a hardware forum in general
I would say your problem is exactly opposite.
If you buy a cheap crappy car, it will normally run quite okay. However a porsche is a toy, and will require way more work and money. So if you buy the best stuff, you get not stuff with a works usually target, but with a we need something cool, even if it hinders functionalty target
I usually only buy mass market stuff, eg. the hardware everyone and your supermarket sells, and it so far outlasted every thing of expensive stuff my friends bought.
It depends on what you spent the extra money on, I guess.
I quite frequently spend a little extra to buy motheerboards that were designed for servers. To a one they’ve been solid performers.
Also, the only systems I’ve ever had chronic hardware problems in were the ones I put cheap power supplies into. I learned that lesson a long time ago. The PSU is the only thing that will slowly take out the other components in your system. Spend a little extra on it and buy a quality brand… and buy 100 more watts than you think you will need.
Agreed, but is this a fair analogy? Motherboard inside your PC is not something you show to your friends every day, typically. And when you buy Brembo brakes for your car, you don’t expect them to work worse than stock ones you bought with your car ten years ago - and ten times cheaper? This is not that everyone just has to buy super high end stuff, this is a whole another thing - that every vendor wants to look like high end, and quality is sacrificed to that look. I.e. dropped even lower than that in even middle-priced things just few years ago. This is what I call “silly games”, and more and more manufacturers are noticeably playing it now. Others just quit sooner or later. This was all the way on the scene, 1st example I can remember was when AMD started marking its CPUs 5x86 - actually a 486, but with higher frequency. That was fast one, it was running at 133 MHz instead of 100, and I overclocked it even to 160 MHz, and it was showing comparable performance to Pentium-100… on some tests only. In reality it wasn’t that fast, not even close to that. But looking at the mark one could expect it being at the same level indeed. But when it is related to performance - there are tests and labs, and whatever you mark your chips the real values will appear soon, but when it comes to reliability - this is not that easy, no lab will run a continuous one-year-long uptime test to reveal whose board is more reliable, because people need boards now and not one year after. So it’s a question of trust now, mostly… if a brand started playing those silly games - it will play them no matter if you buy its middle-level stuff or whatever pretends to be cutting-edge. You still can get some crap embedded and more or less carefully masked from you.
Have you ever heard of the expression called “the silicon lottery”? It’s mentioned a lot by overclockers since not all components can be overclocked to the same degree, even if they are the same model. Why?
When chip manufacturers make wafers not all of them come out as good as they should according to spec. They are tested and then used for different models. Like the best possible chips are used to manufacture the flagships and the worse ones the consumer versions.
Like a perfect wafer piece will become the core of an 1080 and a majorly failed one will get locked down in some areas to fit the 1050 spec, and an i7 with screwed up hyperthreading will become an i5 or something (this part be totally wrong) so on.
But my main point is that all of these are (or at least should be) rigorously tested before being sent to packaging so you are getting at least the minimum required spec for that chip model and perhaps slightly better. So I’m not exactly sure what you guys are whining about.
And you definitely can’t compare this to cars since they’re way easier to manufacture and can be made nearly identical.
Disclaimer: I don’t actually know anything about anything.
I kind of agree with you. Even my cheap motherboards lasted for years. The only flakey behavior I’ve had could be traced back to something specific (like dodgy PSUs).
I never overclock anything, though. If you do that then you are definitely increasing the possible failure rate for all the reasons mentioned above.
My motherboard is no longer being produced and it still runs beautifully. It’s an M3N78-VM, and the only issues it has is sometimes my overclock profile will fail to load, or the system stalls upon boot up. But those are just minor annoyances that I learned to live with.
My very first PSU was some strange unheard of brand and was very cheap, it got blown capacitors over the first year so I had to replace it.
The CPU i have is a Phenom II X4 955 Black Edition, and I overclocked that thing to 4 GHz. I could push it harder but i’ll have to get a water cooler first so it doesn’t melt 
So far I’ve been very lucky when it comes to hardware, the PSU isn’t luck it was just not making a good decision. But then again I was rather lucky the rest of my system wasn’t taken down with the PSU sooooooo yeah!
Actually the high-end-low-volume argument makes sense to me, I hadn’t thought of it that way. I can’t definitively say that the fragmentation of the PC market into mobile crap (even though I own a small fleet of the crap) and “gaming consoles” has reduced the quality of the engineering and rate of performance improvement (I guess we’re hitting silicon’s limit now anyway), but it makes me feel better to blame these things. 
And it’s plausible - less of society’s resources going into that specific market now…
Years ago, I used to go cheap on PSUs - not anymore. But that didn’t stop me from having one blow up on me (high end EVGA!) while it was fairly new. That was surprisingly loud - like a big balloon popping. At least they have pretty decent RMA procedures.
I think I’m done with water cooling, one leak (even though it was an old thing) is enough to scare me off. Besides, those high-end Noctua air coolers beat some water cooling and they look bad ass.
I shouldn’t probably mention overclocking above at all… I don’t do it for many years, that was just emphasizing that AMD truly did a better CPU (of its generation) that time but later remarked it as a higher gen one which it was not. And yes, having a good PSU with some power room reduces half of potential problems, but what could be probably missed by some of you [quote=“pspeed, post:48, topic:37070”]
buy motheerboards that were designed for servers
[/quote]
Yeah, that’s what removes second half of reliability problems. But both points are not immediately obvious for most of home-users, and certainly they have some ground behind. Why would you pay extra money for server board if home ones were stable enough? Exactly, because in general they are not. It comes to lottery more and more, and board price level doesn’t correspond to the result, unless you do as @pspeed said.
I started to mount PC by componentes a few years ago for me and also for friends (you know, being the informatic guy of a group of friends, family, etc…) instead of buying the complete equip in electronic stores.
What I’ve observed is normally PCs are not well balanced. In stores you will see some flagships in the characteristics of the PC (processor, graphic card, amount of RAM…) and normally they don’t provide too much information about motherboard and PSU. Why? Because they use low price/bad quality components to be on budget. You can found the same on forums of people that build PCs, they expend most of the bugdet in the processor and graphic card, they use the most powerfull PSU but with the lowest possible price, cheap motherboard … and that’s a source of problems.
When I mount PCs, I select components from brands with good reputation in that component (ASUS’s motherboards, Corsair/Seasonic PSUs, Seagate’s HD, etc…), trying to balance the budget (I prefer to mount a i5 with good components surrounding it than an i7 surrounding by crap…). So far the only problems I’ve are related with Win10 updates and drivers issues.
By the way, only few brands of PSUs manufacture their own PSUs (Seasonic, Enermax…). I used to buy Corsair for PSU, but I was really disappointed when I found that they use designs form other brands (normally Seasonics) but in some models they use designs from who knows brand. In the case of that EVGA PSU will be interesting to know who is the actual (OEM) manufacturer…
Can only agree, I have my trusty old i5’s both still running. One of them was a server for 3 years running 24/7 with desktop hardware. (And a 10 year old server power supply). All still alive. But that is exactly what i meant by market size. While the i7 is the shiny pice, the i5 and i3 are sold million times more, (due to offices) so their platforms are better tested, as a (court forced) recall here would create an extremly high cost.