ModernFW Was An Exciting Announcement This Week That Went Largely Unnoticed

[ms178]

Not so much. AthlonXP that followed it was about as fast (usually faster, but not always, a few workload types worked better on intel) as Intel's best. An Athlon XP 3200+ performed about the same as a [email protected], but ran at ~1GHz less. Drew much lees power and was cheaper, too. And the cherry on top was that with a good mobo, you could get the much cheaper AthlonXP 2500+ and you were pretty much guaranteed an overclock to 3200+ levels, effectively getting the performance of Intel's fastest CPU. In turn, [email protected] could be overclocked to 3.2GHz, but the thing was still more expensive and hot as hell.

Long story short, not so much faster in absolute terms, but leagues better in IPC.

Well, even the K7 Athlon's (Slot A) provided better value / performance in many games than the Pentium III: https://www.youtube.com/watch?v=XfecL6lAq88

[DanL]

Athlon was fast?

It depends on which Athlon you (or the person you quoted) are talking about. Athlon64/Clawhammer/130nm took the performance crown without a doubt until Conroe took it back. And then there was the whole RDRAM thing..

[SystemCrasher]

When it comes to firmware, Intel is most famous for
1) Backdoor-like misfeatures, be it ME or BootGuard, it would backstab and put nasty surprises.
2) Fake opensource. Formally TianoCore is open. Now good luck to get the source of horrible blobs MB vendors throw on your head. ME, FSP and so on to make it more fun. Oh, seems that fake opensource isn't for YOU? So have fun doing unpaid job that only saves some penies to motherboard vendors & companies like AMI and somesuch. But I'm pretty sure MB vendors and companeies like this can pay their employees, so why bother at all?
3) Gross overengineering. Its hallmark of Intel corp.
4) Windows-only way of doing things. In case of UEFI it gone as badly as using MS file formats for firmware modules. Yes, with legacy MS-DOS HEADERS. Even on dunkin ARMs!!! That are inherently incapable of running that code!!! So wintel shit is wintel shit and when someone mumbles about getting rid of legacy they should ditch tiano core as unworkable overengineered wintel crap, that mostly only used by bunch of proprietary scums.
5) Good luck to find bios or uefi firmware on x86 that works without some bugs on e.g. Linux, not to mention others.
6) Even more luck to get source to try fix these bugs....

So uh yea, yet another intel firmware crap is surely exciting. Would it feature more backdoors, blob components, overengineering and wintel shit at expense of everything else, just as usually? How about truckload of preinstalled MS keys you HAVE to "trust"? Would all this trash be there, as usually? Oh maybe that's why it gone unnoticed?

[bug77]

Well, even the K7 Athlon's (Slot A) provided better value / performance in many games than the Pentium III: https://www.youtube.com/watch?v=XfecL6lAq88

Which is exactly what I said. Much better value, if you compare to the poor sobs that bought P4+RDRAM systems

Anyway, back on topic, where does this leave UEFI?

[Farmer]

Which is exactly what I said. Much better value, if you compare to the poor sobs that bought P4+RDRAM systems

Anyway, back on topic, where does this leave UEFI?

Disclaimer: I have no idea what I'm talking about on this.

In December of 2018 Microsoft announced "Project MU" - a fork of TianoCore. Which this is too apparently.

So UEFI apparently is shuffling off the mortal coil.

[c117152]

Disclaimer: I have no idea what I'm talking about on this.

In December of 2018 Microsoft announced "Project MU" - a fork of TianoCore. Which this is too apparently.

So UEFI apparently is shuffling off the mortal coil.

No. Even if Microsoft, Google and Apple all have their own reduced versions or none at all in their embedded products and servers, UEFI and TianoCore will stick around so long as Windows requires a UEFI implementation and the motherboard vendors don't see a reason to stop using it.

[R41N3R]

Uefi must go, this is overdue, it has been enforced on us with all its bugs, slowness, complexity, vendor control, security issues, ... Just I'm not sure how this could be fixed by using TianoCore as a base. If they want to get it right then it should be based on Linux Boot and CoreBoot.

[R41N3R]

Dropping legacy bios is like dropping X, not gonna happen because huge amount of applications that needs them. Intel does not dictate BIOS programming luckily. Nobody use intel motherboards.

I do have a Intel board and to be fair, it is one of the best ones I ever had, the system in total consumes only 10 Watt in idle. I don't like Uefi at all, but their implementation is much better than from the likes of Asus.

[Licaon]

BIOS? That old thing is bad... use our new EFI/UEFI...

Years of wasting time reverse engineering around UEFI bugs...

Here is ModernFW...trust us...it's better....this time...


Oh FFS

[starshipeleven]

I understand investing into some experimental stuff to look for potential improvements, but considering how versatile and flexible coreboot can be, I don't understand why Intel wouldn't build on top of it to create a successor to UEFI.

This is still a EFI-based firmware. The lower levels (base hardware init and all that jazz) are still exactly the same.

What changes is the UEFI "userspace" bloat, that is slimmed down considerably as it's not really initializing most of the hardware nor providing anything remotely complex or "smart", relying on the OS kernel to initialize non-critical things during boot if needed.

If they could accomplish something similar to rpi's firmware which can be configuring with a single text file, I'd be blown away (although I prefer booting straight into GNU GRUB).

For the love of god, no.