Discussion in 'Article Discussion' started by The_Pope, 17 Apr 2007.
Hm. BIOS as old as I am? Interesting.
Don't get rid of BIOS! You can fit an OS on it: http://linuxbios.org/Welcome_to_LinuxBIOS
os on a bios... first time i have heard about that, looks interesting
Wouldn't a higher level language and more options/configurability make the computer, or at least the starting of a computer more prone to errors? Even if it could work the other way in helping to diagnose errors?
UEFI has been in the pipeline for at least a couple of years now. Is it just that it's taken this long to develop, or is it that MS hasn't been interested until now, meaning that mobo manufacturers have steered clear?
btw, "on mass" should be "en masse"
Lets hope it allows overclocking and advanced options like the DFI board mentioned. I get a sinking feeling that with all the eye-candy and ease-of-use that there is a downside here (ala Vista - eye candy and not very useful)
as pointed out in the article... no current micorsoft os (including vista) supports it.....
and i think thats why it wont show up on the desktop market for a few more years.
I say give it a chance, if it sucks, then back to the BIOS we go; and AMI can try again.
The more errors you can report on, the more errors you have.
I'm not sure what footprint is being talked about here. If he is talking space in the flash part, then I think that number needs to be checked. But if we are talking about the services the firmware leaves behind in memory for the OS, then this number sounds right, but only if there is no legacy BIOS support in the firmware.
Phoenix is a totally different company from AMI. Someone a bit jet lagged?
Looking at the Intel server platforms, I wonder about all the claims of speed in UEFI, but that could be because it is a server platform. Where I think the real promise lies is in the ability for hardware on add ins to communicate to the firmware in a way to create a seamless user experience. There is a spec that would allow something like a RAID card to publish information that the setup utility can use to handle all the RAID settings from the main setup utility. I think we are still years away from this, but the idea is there.
As for an OS in your BIOS, I hope Rich or one of the other staff there can convince someone to let them boot to the EFI shell and play around.
As for Linux BIOS, I think it may be a solution in search of a problem. It also looks like the idea is to create a BIOS specifically for Linux, with the caveat that 'its open source so you can rewrite it to do what you want'. If MS did something like this, the community would probably be in an uproar. UEFI seems to be a pretty good middle ground. It provides preboot services that are OS agnostic (in terms of the MS/Linux/Apple organization), it is an open standard, and there is even source for one implementation at www.tianocore.org.
If you took the time to read the FAQ on the LinuxBIOS site it explained that it is essentially the same as UEFI but for current boards. the tiny OS just loads from NVRAM and unzips the main payload which could be just about anything; from Gentoo to Vista. It's original usage was to make the maintenance of clusters easier by allowing BIOS setting changes to be done over the network as opposed to having to do all the nodes manually (imagine doing a 256 node cluster like that!!)
According to the story, Microsoft is supposed to release support for UEFI in the 4th Quarter. Of course, It will probably not be for anything other than Vista. No XP OS I'd bet need apply, But that's M$, Not only do they want a lot of Money for this OS, But then they will probably only support UEFI with Vista too.
actually its not entirely their fault. XP is locked against such major changes by the nature of it's code, but Vista is more flexible and upgradeable, probably because Microsoft got smart and expected UEFI to be ready during it's lifetime.
Physical footprint on the motherboard. Flash is flash and is disproportionate the size of the package. I wasn't referring to the software side at all.
Changed to competitor like it should have been It's not jetlag, it's a few straight 15 hour days.
Sorry for the bash of the Linux BIOS, I was thinking of a project where the payload was a kernel in the flash and only intended to run Linux.
Now I'm totally confused. From what I have seen, the flash parts used for a board with EFI are essentially the same as those used for regular BIOS (with perhaps a few more megabits on them). The only thing I can guess is that 'EFI is so reliable, we can solder the flash parts to the board so we don't need the big, bulky, socketed parts.'
I'm curious as to the specs of what AMI was running their UEFI BIOS on, is that a production board or something that is not released yet?
As for booting XP and any other non EFI aware OS, there should be a CSM included which gives you the old BIOS interfaces for the current software. The CSM will be needed for a while as there are few (if any) EFI compatible PCI, PCIe, or PCIX cards on the market.
One more addendum to the information you presented. You stated that you were stuck in VGA land, but I'll bet that is more a quirk of the video card than the firmware. The graphics output protocol spec in the UEFI 2.1 specs make no mention of resolution limitations and puts it all on the video card.
So it looks like we can get better eye candy as time goes on
The plural of BIOS is BIOSes, not BIOS' which isn't even the correct way to make it possessive, which is BIOS's.
Finally byebye BIOS. There have been companies that ware practically
screaming for some EFI-luvin', mainly because it's a lót better then BIOS
and easier to code for.
OK you're thinking about it too technical. I was literally referring to the fact that 99% of motherboards use the socketed style which is a few square cm in size. The EFI shown is the size of a little pinky fingernail.
I was referring to the physical package of plastic which holds the central flash, not the physical flash itself.
All I can say is great. But, then there is more chance for bugs, and what happens if it crashes? Not good. But, I think that it's ease of use will shadow that.
Separate names with a comma.