this is big news
|
| Author | Content |
|---|---|
| tuxchick Feb 26, 2007 10:14 AM EDT |
Most proprietary vendors don't know squat about innovation. Phoenix and American Megatrends have owned the PC BIOS market for years, and what do we have? Buggy, redundant, inflexible BIOSes that think we're still running MS-DOS. And even worse, the DRM BIOS, which has already been installed on a number of PCs since 2003. This is a very good-quality mobo with open sound and networking chips, and not some moldy leftover that no one wants. Hurrah for Gigabyte and AMD! Hurrah for the OpenBIOS and LinuxBIOS folkz! This is way more important and newsworthy than Dell's announced "support" of desktop Linux. |
| Bob_Robertson Feb 26, 2007 10:59 AM EDT |
What I want is at least a 2 CPU NUMA mobo, preferably 4 socket, so I can put in dual-core chips for 8-core SMP. Ok, I'm dreaming, I'm not wealthy enough for that. $115 for this LinuxBIOS mobo is very, very reasonable. Single socket AMD 64 x2 it will have to be. Someday. I would have liked two IDE channels rather than just one, but people are so into SATA. I guess my IDE boot HD and a DVD writer will have to be all the IDE there is. I still trust boot IDE, until I see boot SATA working consistently. |
| TPuffin Feb 26, 2007 11:16 AM EDT |
Bob, my SATA has booted consistently for about three years now. (Gentoo, nforce3-250 chipset). Admittedly, one reason I ended up with Gentoo was the _working_ 64-bit SATA support... 8^) For my next system build, Gigabyte goes to the top of my list of contenders. Tuxchick's right, this is indeed a big deal. Rejoice! |
| tuxchick Feb 26, 2007 11:24 AM EDT |
What SATA booting problem? I've been using SATA drives for a couple years on all kinds of systems, and haven't had problems. I would love them for the skinny cables if nothing else. |
| jimf Feb 26, 2007 12:28 PM EDT |
> Tuxchick's right, this is indeed a big deal. Rejoice! Well, it's a start. One down and 'only' 999 to go. |
| Bob_Robertson Feb 26, 2007 12:45 PM EDT |
Ok, so I'm woefully out of date. Sue me. :^) The last "up to date" hardware I bought was this laptop, Sony Vaio, in 2003. It took a year before the built-in wireless card was supported. I still like a relatively small boot drive with a couple (or 9, if I'm feeling really wealthy) RAID-connected disks. I figure that's a perfect way to use both IDE and SATA ports on such a mobo. The reliability comment was just because the original "it won't boot from SATA" comments that I read, and the fact that I have yet to get a mobo that has SATA so I have no experience with it. It just hit me that I was thinking of a 40GB HD as "small". Oh, how the march of technology moves on! Indeed, Gigabyte is now top of my list too. Nothing would please me more than a system that is F/OSS from the bottom up, and as customized as the maker wants to make it. What really sucks is that I have no reason to build a good tower system. I have the server with a 1GHz Athalon that is all I need, and my laptop will serve me for the forseeable future. My wife likes Windows and XP is by her choice , so her system is all she wants. The only thing that would be an improvement to the home is a moderate laptop set up (with Windows, damn them) for my daughters games and so she can watch her Thomas the Tank Engine shows on something other than the TV or my machine. How arbitrary the universe is that her games won't run under WINE. |
| Aladdin_Sane Feb 26, 2007 12:46 PM EDT |
Um, the 2.6.17 and early .18 kernels from Redhat crash and burn on a (newer) Promise SATA card. This was oh, July-December 2006, if I recall. The RH .16 kernel did not do this; the driver "just worked." See, for example, http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=199216 I'm privy to other annoying SATA bugs that I really wish I didn't know about. I can find 176 bugs involving SATA on RH's Bugzilla without trying real hard. Unlike IDE, SATA has yet to get out of "1st generation" mode. The Debian based dists tend to have fewer of these types of problems because they're not trying to satisfy every corporate interest in existence. Like Bob, I use IDE for booting; I may build a complex sophisticated storage interface for the OS and the data (just for fun), but /boot goes on a throwaway 40GB IDE drive. |
| jimf Feb 26, 2007 12:58 PM EDT |
> What SATA booting problem? I used to see a lot of problems with sata on IRC support. Fewer as time goes on. My understanding is that the specifications were very loosely written to begin with... Not a very reliable standard. I tend to believe this, as, it's taken until the last year or so for sata boards to be reasonably consistent, and, the Linux support to cover all variants. |
| hkwint Feb 26, 2007 12:59 PM EDT |
Well, about the IDE/SATA stuff: There just was a major cleanup in the Linux kernel, and if I'm correct, for the Linux kernel IDE (PATA), ATAPI and SATA are all the same now, all implemented in libata, and all handled in the same way. If I remember correct they did this because the IDE API was rather messy, but I can't find that at the moment. There are also IDE>SATA converters (sockets): http://www.picco.nl/popup_image.php?pID=83 And, there are PCI cards with two IDE connectors. [url=http://www.aria.co.uk/Products/Components/Controller Cards/IDE/ IDE 2 Port PCI Controller Card?productId=20137]http://www.aria.co.uk/Products/Components/Controller Cards/I...[/url] No reason to throw away your IDE disks yet. |
| jimf Feb 26, 2007 1:03 PM EDT |
> No reason to throw away your IDE disks yet. Not likely for me anyway :) |
| Bob_Robertson Feb 26, 2007 1:27 PM EDT |
hkwint, yes, I remember reading about that. The core IDE code was really old (and reliable), but a god-awful mess and very different from the later more "standardized" interfaces. I'm sure that the standardization was a really good idea, and maybe linked to the async I/O push. |
| jimf Feb 26, 2007 1:43 PM EDT |
> The core IDE code was really old (and reliable), but a god-awful mess and very different from the later more "standardized" interfaces. The 2.6.19 and later kernels use the new IDE/DMA (v2) rewrite. Performance of IDE (and I believe sata) is noticeably improved. Mounts are nearly instantaneous. |
| Aladdin_Sane Feb 26, 2007 1:54 PM EDT |
> No reason to throw away your IDE disks yet. What? MFM/RLL doesn't work for you? (ooooh, ST-506 with 74xx series logic chips...) |
| jimf Feb 26, 2007 2:11 PM EDT |
> What? MFM/RLL doesn't work for you? /me Hears Simpy singing 'memories' :D |
| swbrown Feb 27, 2007 6:19 PM EDT |
We'll still likely be screwed for ACPI power management because AFAIK the bytecode is on the cards themselves, and they act as if the system is running the corresponding Windows driver when they screw with the hardware. Net result is that there are tons of workarounds per card to get it to suspend/hibernate correctly. |
Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]
Becoming a member of LXer is easy and free. Join Us!