Slashdot Mirror
Mandrake 9.0 for AMD 64-bit Technology
Wister285 writes "Mandrake Linux has released a version of their operating system that is compatible with AMD's 64-bit x86 architecture. This version is based upon Mandrake 9.0. In addition to this, Mandrake announced Corporate Server 2.1 for AMD64 to be released in April 2003 and MandrakeClustering for Opteron in June 2003. Although they say that you can download the operating system now, I cannot find any FTP servers. The press release is located on Mandrake's website."
Will it be enough to keep them afloat?
Is anyone really running Mandrake on a business server? I thought their target market was educational users and the desktop...
- Despite popular opinion, I am not perfect.
Most Linux distros announced their 64-bit OS last fall when the Hammer was scheduled for Q2. Last Fall at Comdex Epic was running the 64-bit version of Unreal Tournament 2003 on the AMD Hammers while running a 64-bit version of SuSE. Maybe Microsoft was the first to announce it, but Linux was the first to do it.
And yes, Intel plans to ride out the Itanium...they see no value in a 64-bit x86 processor
Lets be honest, it doesn't really matter who announced their support for Hammer first. It was inevitable that that major OSes that run on consumer hardware would support the latest processor series from the world's second largest CPU manufacturer (who are in a great position to topple Intel from the #1 spot).
Pro-MS, Pro-Linux, it doesn't matter - putting a slant of this kind on this is pretty silly. Somebody has to announce support first.
Alpha
IA64
Um, at least until recently MS supported Windows on Alpha, but since Alpha looks like a dying processor (market-wise), it doesn't make sense for Windows to continue to support it. And Windows will run on IA64 (Itanium, right?). Guaranteed. I think it already DOES run on that. Also, Windows used to run on PPC so it could conceivably do so again, if the processor market changed somewhat (doubtful).
I have been a long time Mandrake user (for the last 3+ years, I think) but wanted to try the new RedHat 8. So (as I have /home as a separate partition) I wiped the root and reinstalled. I had a comple of immediate gripes with RH8. First of all, both my partitions have always been resier since 2.4.1. The fact that I couldn't (even under the "expert" mode) install a fresh copy to an already-formatted reiser partition I thought was silly. But I was willing to bend a bit and made the root /ext3.
But I came to find later that the ntfs.o module was no where to be found and I couldn't write (ro) my win2k partiiton. Which was a must. I tried compiling the included source but someone got all these errors just for the ntfs module. Very odd - I've been compiling my own kernels since 1.2.13 and never found these errors before (don't remember what they said now).
Finally, though my harddisk had DMA successfully enabled, I just couldn't convince RH8 to use DMA on my DVD drive - the absence of which made everything choppy. hdparm just told me that was not possible.
So I'm back with Mandrake 9.0. Which I'm generally happy with SAVE FOR ONE BIG HEADACHE. I installed the "dev workstation" setup. But I still find I must keep installing -devel.rpm's left and right. O.k., this isn't a real problem, but I've found that these -devel.rpm's and their dependencies are quite equally distributed across ALL 3 DARN CDs!! I normally have to put in 2 of the CDs if not 3 to install any one devel package. This is infuriating!! Why?
Windows NT 3/4 ran on MIPS, Alpha, and PowerPC. They were all eventually dropped as they never caught on.
[Sorry, I hit Submit instead of Preview]
;-)
According to Netcraft, there are almost 100,000 web servers running Mandrake (look for Apache-AdvancedExtranetServer).
I thought their target market was educational users and the desktop...
Think again
I do not think that any current or near (5 years) future version of Windows will be ported to MIPS, PPC, or Alpha.
Well, certainly not Alpha, since that's officially a dead architecture.
MIPS and PPC...well, probably not, but the history of the previous ports of NT to those architectures suggests that MS will be happy to update those ports the moment someone shows them the money for it. This isn't as far-fetched as it might sound: remember that they're also pitching "Embedded NT" along with WinCE, and there are a lot of PPC devices in the embedded market.
Windows Server for Itanium is a lock. It might not ever sell more copies than the Alpha and MIPS versions, but it will happen, even if HP and Intel have to underwrite the development costs themselves. They have too much riding on their investment in the ia64 platform to do anything else.
News for Nerds. Stuff that Matters? Like hell.
I actually just got back from a presentation by AMD here at UIUC (and I won a free t-shirt, too). They oulined the whole hammer architecture and how it's going to be a good thing. By putting the north bridge/memory controller on the CPU die, they're able to cut the DRAM latency by 20% over Athlon! Anyone who's designed computer architectures knows that 20% is HUGE! It only takes 54 clock cycles to complete an instruction cycle, including memory access; if there's a cache hit it was around 30. Actually, the memory read process is started in parallel with the cache hit/miss test and then canceled if there's a hit. Memory bandwidth is also going to get pretty ahead of Intel. AMD is really going to step ahead of Intel with the new hammer architecture. In the future...multiple cores on a single die. That means a single chip, multi-processor system. That'll be huge for the server market! Tech talks are fun!
"The irony is that IBM once, rather cavalierly, dismissed the PC, they learned the hard way, Intel seems bent on making a similar choice."
Not really the same situation. IBM thought something like "People will never want their own computer- they're for businesses need to crunch lots of numbers". They dismissed a whole market segment, because they didn't believe it would catch on.
Intel is thinking something more like "Software companies aren't going to be realising much 64 bit software for PC users, so it makes more sense to spend cash on improving 32 bit CPUs then to try to convert them over to 64 bits."
We'll have to see if there is a big movement to make 64 bit apps, to see if they're right. But don't presume they're overlooking anything- they're just predicting a different future for software, and, considering there marketshare, they might convince a lot of designers to not sweat 64 bit precision (though I don't know what they'll do about memory adressing).
On Friday, the first 9.1 CDs will be burned by Warly and the gang for internal testing over the weekend. The plan is to release to Club members on Monday. Free ISOs will not be available until after the boxes are in stores.
Yea no doubt. I think x86-64 is a good idea. I mean, the x86 technology isn't the fastest, it's no the most effecient, but it works and it's market share is ridiculously large.
x86-64 ia a logical step. It allows us to be 100% compatible with the hundreds of thousands of software packages that exist today, as well as 64-bit software.
As technologies progress, and the Mhz keep on getting pumped up, the performance hit by staying compatible becomes less of an issue. Obviously the Athlon64/Opteron is not meant to replace high-end SGI type supercomputer CPU's, it's meant to provide an upgrade path for x86 to 64-bit computing.
I just hope AMD can produce enough chips, and I hope they are affordable. Since it's AMD's intention to replace all their CPU products with the new 64 bit versions, my guess is that the prices will stay competitive with the 32-bit offerings by Intel. At least in the Desktop market.
Personally, I can't wait for a 64-bit desktop!
- It's not the Macs I hate. It's Digg users. -