The First Skylake Motherboard To Fail Me: Goes Kaput After Just 4 Months

[duby229]

UEFI is a Intel idea and is a board firmware.
MS's idea is Secure Boot, a feature of UEFI board firmware.

UEFI is annoying mostly because those making the firmware are a bunch of morons, not because it is bad by itself.

Anyway, if you configure it right you can get neat boot managers like rEFInd instead of ugly Grub.

I agree 100%. I mean the problem with UEFI is that there is no consistency between implementations. Nobody ever cared enough to ensure sane and stable standards could be enforced. Too many boards are just simply unusable, especially true for OEM boards. You got SB enabled by default in some places, disabled by default in others and MS being the only one in a position to issue keys. You got 32bit firmwares on 64bit systems, and binary incompatibility basically across the entire line of available products. UEFI on arm will only make these types of problems with it even worse.

It's a terrible idea. If there needs to be a high level firmware UEFI isn't the right choice. I don't know what is, but that isn't it.

EDIT: At first glance I'd say a recent Open Firmware implementation is a far better option.

https://en.wikipedia.org/wiki/Open_Firmware

[kgardas]

Supermicro here, moved from Asus since Asus no longer provide feature they advertise, i.e. ECC. Buying always workstation or single-cpu server stuff thought...

[SystemCrasher]

Cheap boards usually come at crappy quality. Most notably, high-quality electrolytic caps are expensive and cheap boards are usually using crappy ones. They could survive for a while if using low power cpu idling most of time, but if you plug high-power cpu and run it under full load quite often, caps would not last long. Ever seen 105C mark? Its widely known el caps are supposed to survive 105C. But careful grinding through datasheet explains they are only rated for 1000 hours under these conditions. Which is about 1.5 months. Bad caps may or may not look poorly. Sometimes they blow up, but sometimes caps could expose high ESR even without looking bad. High ESR and reduced capacity warrants unstable operation and at some point system fails to start at all. It could also be crappy cooling of chipset, cheap boards save cents on adequate heatsinks, etc, so chipset and somesuch could run at insane themperature. Such boards are supposed to be used in low-end office computers, which are "idle" like 99% of time so overall life expectancy is quite okay due to low stress on components.

[SystemCrasher]

Supermicro here

Hell yeah, really great vendor, who is known by the fact they have put backdoor into their BMC. Oh, "engineering password" they call it. What a good company.

[Melcar]

I grew up in the days of Abit, DFI, and Epox. Back when instead of "gamer grade" you had "overclocking grade". At least these "overclocking" boards looked better.
These days I mostly buy Asus boards. Quality is decent and I like the UEFI implementation. Giga is alright, but their UEFI sucks a big one.

[starshipeleven]

It's a terrible idea. If there needs to be a high level firmware UEFI isn't the right choice. I don't know what is, but that isn't it.

EDIT: At first glance I'd say a recent Open Firmware implementation is a far better option.
https://en.wikipedia.org/wiki/Open_Firmware

Coreboot/libreboot will do fine on arm. The issue is that if customers (companies) need UEFI because secureboot or even just because they know the name already, we get UEFI on arm, and while on one hand it will solve a bunch of issues like idiotic stock bootloaders common in embedded land, it will bring its bag of fun with it.

[starshipeleven]

Hell yeah, really great vendor, who is known by the fact they have put backdoor into their BMC. Oh, "engineering password" they call it. What a good company.

Everyone and their dog knows that the BMC of ANY fucking board is full of holes and backdoors because the manufacturers are retards. (Hp and dell are well-known for their holes and backdoors too, dunno about tyan) There is a reason most decent admins never ever let a server on the internet on its own.

[zeroepoch]

It is handy feature for phoronix users that articles are categorizes, Michael summarizes news well, it is much nicer to read phoronix articles than some mailing list discussion or release notes. Distro pages are also boring, at least in debian:

https://www.debian.org/

Nothing happens there, more action here:
https://www.debian.org/devel/debian-installer/

I would agree with that if I was familiar with the topic. The search summary is a better way to find related news. In this case though I didn't know about "Clear Linux" and wanted to learn about it. The search results lead you to an endless maze that doesn't answer that question. I googled it in the end.

[darkcoder]

Unfortunately you're right about the older boards, great to know.

When I was installing Linux I was still not recognized as having a bit of an attention-problem, but even with legacy-boot option, didn't Microsoft/Windows 8 make UEFI built in to begin with? I don't see that, or nauseating companies like Microsoft doing anything wholesome for the most part.

Feel free to disagree.

If you have Windows 8 or 10, then you will need to disable Fast Boot feature in the BIOS, and probably in Windows too. That prevents Grub from working correctly.

Not idea when using Linux as the only OS if Fast Boot can be enable and Linux take advantage of it.

[boltronics]

I recently purchased an MSI Z170A Gaming M7 motherboard (had the features I wanted, was readily available, the price was right and I trust MSI over Gigabyte). The board has been pretty good, except the efibootmgr command didn't work correctly after updating to the latest UEFI version. It appears to work fine, but when I reboot, the changes did not persist and were lost. Fortunately, reflashing back to the second latest UEFI version fixed the problem.

Asus has let me down a few times by now. Back when DDR400 was new, I brought a mainboard and a few DDR400 sticks. The machine wouldn't post. The issue (according to online discussions) was that the board needed to be initialised with DDR333 or lower first, and only then would it work. So I took it back to the store, and when they looked at it they obviously just used cheap slow RAM and the thing posted fine (which in turn sorted my problem), but then they tried to charge me for it since they didn't see any problems!

More recently, Asus burned me when I brought the Asus G55VW "ROG" laptop. It came with Windows 8.0, but upgrading to 8.1 or anything else failed, as the screen would appear to switch off. Debian also had the same issue when Xorg was launched. What I eventually discovered was that the laptop's EDID didn't seem to reliably be returning anything immediately when queried the first time, resulting in the picture going to a video output that didn't even have anything connected.

I was able to sort that Asus issue out by dumping the EDID to a file and referencing that in Xorg, but the Nvidia drivers for Windows 8.1 had no such option. So I took the board back to Asus, and after weeks with it, they told me that they replaced the motherboard which had no effect, and said they don't consider it to be their problem. Asus just doesn't support running anything other than 8.0 on it. What a load of BS! And probably it was the screen that needed to be replaced anyway! And the worst part was that since getting it back, the battery LED now constantly blinks, indicating that power is about to run out - even on power. It never used to do that. I'd have taken it back again, but I don't want to wait weeks for them to fix it again (and it's out of warranty now anyway).

So naturally I'm giving Asus a miss for a while, hence giving MSI a try.