Slashdot Mirror
Microsoft Commits to Using Opteron
the_1000th_Monkey writes "According these articles at The Inquirer, Infoworld, and The Register Windows XP and Windows Server 2003 will support AMD's 64-bit Opteron processor. Beta releases can be expected in the middle of this year. Here is MS's official press release."
Linux already supports it...has for over 6 months now
"Some things have to be believed to be seen." - Ralph Hodgson
Windows XP and Windows Server 2003 will support AMD's 64-bit Opteron processor. Beta releases can be expected in the middle of this year.
A slashdot story where Microsoft are the good guys! What have you done with the real Timothy?! Taco! Help, Taco!!
64-bit Blue Screen of Death!
Half the MS-Intel Duoply is broken!!!
Now, if only the first half would die off...
Suicide Booth: You are now dead! Thank you for using Stop and Drop, America's favorite since 2008.
And does this make AMD part of the Axis of Evil now?
There's a Mercedes gap too. I want one and can't afford one, but it's not government's job to do anything about it.
Was there any doubt that this would happen? Since MS is running on about a 3-5 year Server cycle, the next server release would happen around 2008. I would assume that most high end servers and many workstations would have 64 bit processors by this time. It just makes sense that MS would support the 64 bit processor being released by the 2nd largest processor company.
You have enemies? Good. That means you've stood up for something, sometime in your life. --Winston Churchill
They will suck in even more platforms.
Amazing...even will have a version for the full 64bit (not using the 32bit compatability...much)
I can't find any information if Win2k3 has support for Intels Itanium 64bit processor...You'd think it would considering MS and Intel spend every night in bed together
"Some things have to be believed to be seen." - Ralph Hodgson
About time that we hear this news from MS. Now, what about linux?
nah, I don't think MS is going to support Linux.
Best Windows Freeware
.... it will be a while before the software catches up ....
...
Just my $0.02 cents
HallmarkOrnaments.Com
No, it's a poorly implemented unholy hybrid of UNIX, DOS, OS/2, and VMS that has processor-specific cruft buried deep under the skin.
It will be annoying when they do release the opterons and there is no (64bit) software to run on them. Sort of buy a system, install a 32bit os and then a few months later reinstall it.
Also I think many people will be dissapointed with the 32bit performance and AMD might get a bad name for it.
Mouse powered Chips, Open source Processors and Lego
Beta-version of Windows 2003 is likely to support a developers-only version of Opteron in 32 bit mode, however, only in case M$ does not discontinue the whole Windows 2003 product line. Sorry.
with Sun now supporting the Opteron, lending more legitimacy, it was only a matter of time before Microsoft jumped on the bandwagon.
then again, Microsoft could have been holding on to their press release, and Sun could have jumped on the bandwagon, releasing their press release early in order to beat out Microsoft.
either way, it really should be a simple matter for Microsoft to support this chip. it is backwards compatible, and they have had 64 bit for quite a while, so the heavy work is already done.
Have they made any commitments to Intel's 64 bit processors?
SIG: HUP
Change the record, grandad zealot!
Windows _has_ moved on since that.
Yes, the Sun's rays also shine on Microsoft from time to time.
Heh.
Have you read the moderator guidelines? Well, have you, PUNK? (and I want a Karma: Gnarly option)
I know, stupid question, but an honest one. I'm curious WHY 64-bit is so damn important, and I'm sure others still haven't clued in. Anyone care to post a nice explanation or links or something. Thanks.
"I'm a leaf on the wind. Watch how I soar."
-Hoban Washburn
This is no news to me. I remember reading that AMD was delaying their 64-bit processors until next fall, the reason was apparently that they wanted to have a version of Windows to run on it.
It is therefore no surprise that Microsoft announces an appropriate version of Windows in the same time frame!
that they'll be providing an advanced version of Windows 2003 to work on the new Cyberdyne 1028-bit chip. Biological detection and engagement will be 30% faster, and there will be a whole new meaning to the phrase "blue screen of death".
Microsoft was the primary development partner with AMD on the x86-64 instruction set. What MS wanted, AMD delivered. And it's great! Not like that crappy HP/Intel Itanium fiasco.
Dang it. I like Win2k (I also use linux, but sometimes Win2k is just easier). Too bad they are not going to go back and release a 64bit Win2k. Not that I really expected them to, but that is just a bit more money that my 64bit athlon is going to cost me.
sweet now I can use my new cd key I got yesterday on that chip ;)
---
The Opteron has not been delayed, it's comiing out at the end of April. This is a server chip/chipset, and the motherboards are missing AGP.
The Athlon 64 is the consumer version of this. Both are 64 bit, basically the same chips, Opteron is for serveres, Athlon 64 is for desktops.
From a user's point of view, I wonder if in a couple of years users will have to decide if they want binaries for Intel's 64-bit architecture or AMD's. This as you all know is not a good thing, since it will bring market confussion to users (however, in the server space where these chips are first targeted this is not so big of an issue, specially with technologies like Java). A workaround is for companies to ship versions of their products for both architectures, thus at the very least this represents a burden on developers.
Another posibility I see is that AMD's choice of creating a backwards-compatible x86-64 instructions set will reign supreme over Intel's, and thus force Intel to adopt in AMD's x86-64.
Either way, I see turbulent times ahead...
Most likely, something in a future version that would make it partly incompatible with Linux. Maybe, some chipset feature remaining undocumented, or something that to write code to use would infringe a patent.
I wonder what it will turn out to be.
For most stuff, 64bits will provide negligible benefits. For now, the areas that I see it will shine:
1) Servers that need more then 4gb memory (large db / e-commerce / etc).
2) Programs that need high precision math (scientific / video and image processing)
What do you think the chances are that their main motivation here is that they don't want to be beat to the punch by Linux.
Just imagine if the only 64-Bit servers you could buy were non-MS based...
>Now if you have 64 bits you actually have 2^64 which is an outrageously huge number
2^32 used to be an outrageously huge number, as well. Remember the days when a top-of-the-line box had 640k of ram, and a 20 MB harddrive was considered gigantic? 4 gigabytes was an absolutely ridiculous number to throw around back then.
Just reminiscing, that's all.
I don't need no instructions to know how to rock!!!!
Why not?
64 bit instruction set for faster low level functions, faster 64 bit pipes. Good stuff, all around. Will you have an immediate use for it? No. It will eventually replace 32 bit, and you'll be happier. Just like the Pentium replacing the 486, the 386 replacing the 286, it's a move in the right direction.
There is no reasonable defense against an idiot with an agenda
:wq
If I read the interview at ZDNet correctly, the Opteron can due both 32-bit addressing or 64-bit addressing.
To quote directly: One product feature "is the integration of the memory controller onto the microprocessor silicon, which of course is a good thing for memory latency, and hence a very good thing for performance."
32-bit applications will also get this feature, meaning that they should see improvements as well. It's not exclusive to the 64-bit apps. Both should due well on this architecture.
Forgot a potential benefit in my last post...
3) High geek factor.
2) Most CPU's do IEEE floating point calculations in 80-bit, if you didn't know...
Ahhh...the great dumpster continuum. Many a free computer will be found there. -- sowth (748135)
If you do, you're wrong.
You're no different than the people who type "Lunix."
Learn to enjoy civil discourse.
I hate liberals. If you are a liberal, do not reply.
you mean Itanium 64 right?
There isn't a 64bit pentium processor.
"A plan fiendishly clever in its intricacies"- Homer Simpson
" Right now you are running 32 bits. That gives you the ability to access 4 gigabytes of addresses (2^32)"
Actually, since the pentium pro you have been able to access 2^36 addresses.
graspee the boringly informative
2^32 times the addressing space of 32 bit, so goodbye 4 gig limit. And greater speed / precision ratio. Those are the two biggest points.
-Reid
"Microsoft's 64-bit Windows operating systems represent an inflection point leading to higher performance and greater efficiency for businesses and consumers."
So, are they admitting their previous versions sucked?
By the time their crappy server OS does get launched, they will be facing an entrenched group of free OSs that have 100% market share.
"Only in their dreams can men truly be free 'twas always thus, and always thus will be."
--Tom Schulman
Windows _has_ moved on since that.
So, Windows is now processor-independant! Great! When can I load it on my Mac?
A firewall can not protect you from yourself. Turn off what you do not need. Do not use the firewall to do your work.
It's like adding on and making an 4 lane highway into an 8 lane one, it allows you to handle more traffic. For most drivers it doesn't mean they'll get where they're going faster as the same speed limits apply. But a few proffessional ones will find that the extra lanes mean they can utilize the extra space better and drive faster.
Wax-Museum Fire Results In Hundreds Of New Danny DeVito Statues
Although there are Xeon-based servers that support more than 4GB of RAM, it just doesn't handle it very efficiently by having to use windows and page addressing extension (I think that's what PAE expands to) to address anything beyond the 4GB 32-bit memory addressing limit. Xeon's support 36-bit memory addressing.
With the AMD Hammer's handling > 32-bit memory addressing natively and without hacks like PAE, it will definitely help improve high-memory use applications like databases, large rendering jobs (think Unreal II, future movies), or scientific crunch jobs.
Microsoft has been itching to get a piece of the Unix/Linux dominated server market (and while their share has been slowly growing, Linux has been thumping it for years), so it's not all that suprising to see them support the upstart. Every so often they do make a good choice. Every so often...
Beyond that, Microsoft has been slowly helping AMD over the years, if by just using the 3D-Now optimizations on the early K6-2 processors. Of course, you'll never get Intel and Microsoft out of bed together, but then again, you'll never see them be exclusive either.
It would be interesting to see, since the Beta for Windows 2003 (I actually have a legitimate copy, just no legitimate reason to use it) was kinda slick. Though i doubt if I was running a business I'd consider anything but Linux for base server operation.
64 bit application
-32 bit application
--------------------
32 bit application
There is your answer!
Two reasons. 4gig limit, only 2 can actually be used, this is starting to become a real problem blah de blah de blah.
The more subtle one is that the x86 instruction set is as broken as a broken thing, as we all know, and x86-64 goes some way to fixing that. Particularly in terms of having more registers.
Dave
I write a blog now, you should be afraid.
True, but because of how the memory is addressed by programs (32bits at a time) special considerations must be taken, and things are generally slower.
Maybe you know the answer to this? Do most modern CPUs have a full 80bit adder / multiplier pipeline, or are they handled as multiple instructions?
Anyway, the ability to address a 64bit programatically should have benefits in itself.
It's not like this is a poorly implemented Unix workalike that has processor-specific cruft buried deep under the skin.
No, apparently its much more difficult to port.
Doubtless poorly designed from the beginning. I'd point to the emphasis on "integration" rather than "modularity", just IMO.
Galileo: "The Earth revolves around the Sun!"
Score: -1 100% Flamebait
On the Windows side of things, because the Opteron processors utilize NUMA (Non-Unified Memory Architecture), Microsoft needs to make sure that the x86-64 kernel is stable with NUMA and make sure that 32-bit applications can indeed run while the OS is in 64-bit mode. They (Microsoft, and AMD for that matter) cannot afford a botched release of the operating system.
The one thing that got me curious is that it seems that only Enterprise Edition (and Datacenter Edition as well) of Windows Server 2003 will support NUMA and not the Standard Edition... I wonder if that will also be the case with the x86-64 Edition of Windows Server 2003 (ie: it will be more expensive than Standard Edition, which has a 4GB memory limit anyway).
Really, I think they are. Porting your software to run on an OS sounds like supporting that OS to me. Okay, so the porting isn't done in-house, but it is MS funded.
I am TheRaven on Soylent News
64 bit instruction set for faster low level functions, faster 64 bit pipes
This is just plain wrong. 64-bit words at the CPU level has no direct effect on instruction speed, unless you make tricky optimizations, like packing 32-bit variables into a single 64-bit register and doing operations on them simultaneously (which, in general, isn't that useful, BTW). Yes, there are a couple places where wider registers could be useful (bulk data transfers, etc) but there really aren't that many. Some people have mentioned higher-precision arithmetic, but IMHO, if you need that, you're using the FPU anyway, and thus have had 64-bit (80-bit internally) precision for some time now.
The main reason the Opteron is a good thing is because 1) it provides MORE registers, allowing the compiler to make smarter register allocations, which can provide drastic performance improvements, and 2) it provides access to a larger address space, meaning you can finally have >4GB of memory without nasty paging hacks. Of these, only the first is really that useful to your average Joe, which is why you're only going to see the Opteron in higher-end workstations and servers for the immediate future... at least, IMHO.
me: dud3 1 g0t 64 b1t CpU!@#!
friend: 0wnz0r!
But really, its all said above. Main thing being the memory access problems which current 32bit systems can have.
/* sig */
They're going to support Opteron/Athlon64. Nothing in the press release says they're going to be using it to run their sites
But a few proffessional ones will find that the extra lanes mean they can utilize the extra space better and drive faster.
Wow, I didn't realize you got that added bonus. Goodbye Speed Limit!
Dacels Jewelers can't be trusted.
Don't look now, but Apple's got you by the throat in that department. They won't even give the specs to allow a minor competitor like BeOS run on your Mac. Why would they let Microsoft on?
Remember the 68k->PPC changeover at Apple when they used to ship fat binaries, those with code for both PPC and 68K?
Why wouldn't this be an option? Or maybe that weird dynamic recompilation stuff that the Alphas had for running x86 stuff in emulation?
On the other hand, you could ship discs with multiple versions of the binaries on. This could be the thing that finally makes software distribution on DVDs catch on in a big way.
No, the Pentium series cannot address 2^36 bytes of memory.
The sequence is:
effective address -> segmentation -> linear address -> paging -> physical memory.
PAE uses the paging mechanism to allow banking of physical memory. The maximum size of the linear address space is still 2^32 bytes (4GB).
The bottleneck is the limited address SPACE, not the amount of physical memory. What 64 bit addressing does for you is remove the limit on allocating space for files & data.
Forget physical memory, its not important. Thanks to locality you can efficiently cover a lot of address space with 1GB or less of physical memory. You can see this if you increase your windows memory from 256->512->1024 MB. There is little benefit. Paging physical memory in is really quite efficient.
On the other hand the poor old VMM gets really slow if its trying to squeeze everything into 4GB of address space.
I would assume Microsoft is hoping that by the time this happens most developers will have jumped on the .NET bandwagon. I'm not totally certain about the details of .NET but I believe it allows for platform-independant code, in which case MS would only have to provide the runtime environment for each platform.
Got Proof?
Actually, I believe (2^32)^32 != 2^64
2^64 = 2^(32*2) = 2^32 * 2^32 = (2^32)^2
... but not (2^32)^32
Not quite right. Opterons are for workstations and servers. Opterons have up to two memory controllers, can have larger cache, and support SMP. Athlon64 will be single-CPU only (this wasn't the original plan, BTW).
I expect Opteron workstations running Linux64 and Win32 will be available sooner rather than later.
Check the bottom of this article for info on a BRCM chipset supporting AGP 8x (among other goodies).
Sweet!
Galileo: "The Earth revolves around the Sun!"
Score: -1 100% Flamebait
The hardware specific code is contained in the Hardware Abstraction Layer, which is layered under the kernel. Remember that NT has been available for several different architectures in the past. The fact that only IA-32 remains has to do with market realities, not with the design.
Yep, I knew all that...still why was it so much easier to port Linux?
It would still be pretty easy for MS to provide NT for other 32-bit platforms. Now, porting to a 64-bit platform isn't the same thing.
So your theory is that Microsoft has been sitting on it's hands all these years without porting to any of the readily available 64-bit platforms? Including Itanic, which has been around in beta incarnations for several years?
Sorry, I don't buy it.
Galileo: "The Earth revolves around the Sun!"
Score: -1 100% Flamebait
Umm, a 64 bit bus means a 64 bit integer word size.
Yeah, we know that if you use a high level language, that's lost in conformance to a lowest-common-denominator standard that applies to all processors.
Call that 'tricky optimizations' if you like. The world still runs on machine code, even filtered through your kludges.
Yeah but since win2k it's actually gotten harder to port. WinNT 3.5 or so was much easier to port. Why? because the display driver wasn't as integrated into the kernel. They started to move move of the drivers into kernel space with Win2k. Which means that it's harder to port.
Micro kernels are easier to port to different chips. Even the drivers are easier, but everything is slower.
A workaround is for companies to ship versions of their products for both architectures, thus at the very least this represents a burden on developers.
So M$ and Windows developers would be copying what Apple and Mac developers did 9 years ago...
Sounds like par for the course to me.
You're making the erroneous analogy that all the bits on a parallel bus carry different cargo. The wider bus means vehicles 'twice as wide' can travel on the road. That would mean, in two dimensions, four times as big. But your analogy was crap to start with, so my extension is wrong also.
Wow, how the hell did my post end up in this article? Malda, check your code or something :)
* 2 and 4 CPU models
;-) I WANT one of these babies!
* Up to 24Gb of RAM (6 * 4Gb modules)
* Built in boardcom TOE gigabit
* AGP 8X/Pro 110
* Optional SCSI or SATA adapters on board
* ZCRaid compatible for either SCSI or SATA RAID
Time to take out a large loan ?
*drool*
SB
It's old. The more humans I meet, the more I like my cats. At least they are honest.
One question though: Is the Opteron spec an addition to IA32, where winshit would already run, or a completely different chip, in which legacy winshit software won't run? I seem to remember that NT for the alpha architecture died off because of a lack of binaries.
;-)
Opteron is an extension of the IA32 spec for running 64-bit programs. It offers some additional nice features, but needs no emulation to run 32-bit software. I assume that 16-bit software would still run too. I wonder if I could get CPM to load
LedgerSMB: Open source Accounting/ERP
I think you'll find one of the things purged from the x86-64 is some of the worse legacy cruft, including some 16-bit compatiblility.
It's backwards compatible. Read all the Hammer articles from the past year for details on AMD's 64-bit plans.
If you have a real need to address that much memory, you'd be a fool to cheap out and buy inferior hardware like Intel or AMD.
Nonsense. Intel/AMD smoke memory I/O and SPECInt, and in terms of price performance do so in a manner which is breathtaking. Putting a DB on a x86-64 using quality system parts (like, say, a rackmount Compaq (move fast before HP fucks the rackmount Proliants up!)) with large memory makes a lot of sense. If you wanna start doing quad-bank interleaving, 64bit lets you do so with large memory quite nicely. Business computing for the most part only cares about stability, I/O and int performance.
I look forward to 4/8-way smp opteron rigs with quad-channel DDR400 support, featuring 4-16 DIMM slots and multiple 64bit/66mhz PCI, multiple gig-e on hypertransport.
2^32 times the addressing space of 32 bit, so goodbye 4 gig limit. And greater speed / precision ratio. Those are the two biggest points.
Alright! I can finally start using 64 bit floats for curency data types and not have to worry 'bout droping pennies until the numbers get *REALLY* huge.
Just kidding.
Moneyed corporations, non-working 'poor' and criminal prisoners are turning productive citizens into tax-slaves.
NT was available for DEC Alpha for years.
I remember reading an article in PC/Computing back in the days when the top of the line PC was a 486-DX/2 66mhz drooling over the speed of NT 3.51 on an Alpha at 300mhz.
More on 64-bit Gaming and AMD's Athlon-64 Benchmarked. There's probably some other info floating around too, but these sprang to mind.
Rock over London, Rock on Chicago. Wheaties: Breakfast of Champions.
The way I see it, there are two huge advantages to the Opteron - 2^64 bytes of addressable memory, and the new registers. I know that unless the OS is x86-64 targetted, applications won't be able to address more than 4 gigs of memory. That's a VM mapping issue the OS must deal with. So running a 32-bit OS on the Opteron negates that potential 64 bit advancement
However, could x86-64 applications that have no need for that much memory still run in the lower 4gigs of ram while using the new registers? I'm not familiar with the x86-64 specs at a low level but I do know that as Intel expanded x86 from 16 to 32 bits, flags had to be set in the CPU (and presumably supported by the o/s) for applications to use the 32 bit registers. Does anyone know if this is the case with x86-64?
"This is Zombo Com, and welcome to you who have come to Zombo Com" - www.zombo.com
Another chance for leaking Windows code...
could you name a source on that one? thanks...
Who is this Anonymous Coward character, how does he post so much, and why is he always such a whore?
Let's say that I have 64bit.. Two times 32bit... *grin* ;-) Just kidding.
Ahhh...the great dumpster continuum. Many a free computer will be found there. -- sowth (748135)
from the beginning it was designed for portability.
it was originally written on dec alphas to make sure that no ix86isms made it into the code.
Yes, but this was only a 32-bit implementation.
Galileo: "The Earth revolves around the Sun!"
Score: -1 100% Flamebait
on ix86, memory pulled 64 bits at a time. This is because MMX registers are 64 bits wide, and the FPU is capable of a precision of 80bits.
Doesn't mean the normal registers are wider than 32bits however. And there is only the one FPU.
though, you could(and probably should to have stability) use dosbox or some other virtualization program for them anyways..
world was created 5 seconds before this post as it is.
Its called segmented addressing. Its stupid, and I thought the computer industry learned from this already.
Basically the stance that microsoft has taken is that if you need to use PAE, you only need to run 1 application / server.
MSSQL has a PAE option, but it will consume all available memory and not release any of it.
Linux supports PAE too, but applications are still limited to a 3GB address space (when configured this way). Its only benefit is when you have more than 1 process that needs access to large amounts of memory, then the segemented addressing will still be faster than swapping.
NT has been running on AMD64 for more than a year now. There's a difference between porting the kernel and releasing a fully tested product with major application support (Exchange, SQL etc).
So your theory is that Microsoft has been sitting on it's hands all these years without porting to any of the readily available 64-bit platforms? Including Itanic, which has been around in beta incarnations for several years?
Of course not. In fact, there are released Itanium versions of XP and Windows 2000. I heard that before that, there was an internal 64-bit Alpha version of Win2K.
What did you think I ment, that the kernel had to be recompiled everytime the display driver changes?
Duh, that can't be true as how could the driver be plugable like it is without it being a loadable module.
Since the driver runs in kernel space instead of user space it does mean that the driver is less portable which was I tried to say (badly) before.
Remember how we used to say: "Windows is a 32-bit shell to a 16-bit operating system, originally designed for an 8-bit processor with a 4-bit bus by a two bit company that can't stand one bit of competition." Any ideas how to bring that up to date?
Escher was the first MC and Giger invented the HR department.
Does the NT port to Alpha count as a 64 bit BSoD?
Friends don't help friends install M$ junk.
Of course not. In fact, there are released Itanium versions of XP and Windows 2000. I heard that before that, there was an internal 64-bit Alpha version of Win2K.
So, I'll ask again - what is taking MS so long? Linux, with full application support, both 64 and 32 bit, will be ready at Opteron release... MS is shooting for a six month later release date, and knowing Microsoft's record on hitting such dates I'm not optimistic.
Having 64-bit versions of SQL Server etc. ready is an entirely separate issue....although those apps must have already been ported to Win64 for Itanic...right?
Galileo: "The Earth revolves around the Sun!"
Score: -1 100% Flamebait
The key to this whole issue is .Net, and the CLR.
.NET performance, but I think as an idea it's actually pretty good.
Microsoft is supposedly moving all their code to run on the CLR, which is similar in many ways to the Java runtime. Programs are compiled to processor independent byte code, and then before being run are put through a JIT compiler to produce PROCESSOR-SPECIFIC machine code.
Thus, all MS has to do is distribute a version of the CLR for AMD or Intel, and suddenly all applications have optimal support for those processors. Also, when a chip manufacterer releases a new chip, or adds to the instruction set, or your machine has some other hardware change, all you need is, at most, a new CLR, and suddenly all of you applications, from any developement house, have improved performance.
This is actually a neat idea, and allows MS to loosen its ties with any specific platform.
I can think of several difficulties with this idea (There must be some way of doing architecture specific P-Invokes (native calls) for performance reasons, but I don't know enough of the details.), and have only heard disappointing things about
Although its fun to kick MS around for licensing reasons, they have been pretty reasonable updating the software limitations of hardware - 2k pro supposed only 2 cpus, but I am very certain that it and XP pro will see and use all 4 logical cores of a 2 CPU hyperthreading xeon machine.
Therefore, I wouldn't be surprised to see the RAM limits increased when more ram becomes necessary. Right now, MS publically says due to lack of PAE support in Exchange 2000, there is no reason to put > 3 GB of ram in a E2k box. Once apps start pushing things on a more regular basis, MS will probably have to relent. Right now I think the people who really need 64 bits have it, and are using *nix. People who would like it, find ways to deal. AMD hopefully will commoditize 64 bit computing, and make it something we don't need to worry about.
ostiguy
Testing takes time. Does that sound like a reasonable explanation?
How many times are people going to confuse this?
From the story you link to - "Opteron, in keeping with the company's original launch date, is set to officially debut on April 22 in New York City."
The *Opterons* have not been delayed. It is the *desktop* 64bit procs that AMD will release later in the year.
It's not that tricky, people.
Everyone rails on the x86 instruction set. Yeah it's not pretty, it's not fun, hell it's downright ugly. But what are the top SPECint machines these days? Wanna guess? That means something is ok with x86. Yeah it might be hack-on-hack-on-hack but this collection of hacks seems to be working. (They'd be pretty near the top SPECfp's except for Itanium, everyone else's favorite Intel punching bag -- give me a break it has stellar FP, which is what it was made for!)
More seriously, there are some academic studies around that show that variable-length instructions of the x86 ISA actually are improving performance over fixed-length RISC-style ISAs. Why? Because the instruction density in the cache can be higher, and therefore the I-Cache fill rate doesn't need to be as high. Sure, the I-Decode is a b*tch to design and build, but apparently Intel and AMD are able to run it in about 500ps (~2GHz, or better) in 0.13u and below technology. Not bad, not bad.
What?! Oh, ok the silly closed source send your stuff out on CD model of selling software might suffer. Microsoft might have trouble porting the Universe all the time. I can't imagine how the average Visual Basic disaster shop will cope. What's new? This is just another indication of the superiority of the free software model. Free software has been ported to different platforms for years and it's not difficult for the various distributions or their users to keep up with all the archetectures. It always makes me laugh to this kind of excitment for M$ junk and I can't wait to hear excited anouncements like , "With more than 1,200 applications ported, NT^H^H Opeteron is really catching on!"
From the Register, "For instance, there are five varieties of Linux, three BSDs, Beowulf and Windows in the offing. Most of them have either already been released or are due to be released at the Opteron launch." I hope they are not waiting for M$ to finish putting together all it's little peices.
Friends don't help friends install M$ junk.
In the fourth quarter of 2002, Microsoft delivered to industry partners a developmental release of the 64-bit operating system with application development tools for the AMD 64-bit processor hardware.
Helps to read the article before posting a comment on it.
MS has been working on Win2003 which is a major release supporting two architectures (x86 and IA64). Adding AMD64 support would have delayed it. Looks like they felt it was more important to release Win2003 sooner than to have AMD64 support sooner.
Last Windows disk that said it'd work on PowerPC was NT4, IIRC. So you could try booting that. ^_^ You're on your own as far as support goes though.
So, when's lunch?
A 2 CPU hyperthreading xeon machine is still a 2 CPU machine, and Windows not recognizing that should be considered a bug (Now fixed, I believe).
w00tle
64-bit Solitare!
2^32 times the addressing space of 32 bit, so goodbye 4 gig limit.
16 exabytes ought to be enough for anybody.
"Microsoft's 64-bit Windows operating systems represent an inflection point leading to higher performance and greater efficiency for businesses and consumers."
Can someone please tell me what the fsck an "inflection point" is?
I'm a lawyer with excellent karma. Something's gotta be wrong.
Here are all the details from a Microsoft.
http://saveie6.com/
In a recent article in PC Magazine, Dvorak states his thoughts that the Mac may be moving into the Intel camp. The article is http://www.pcmag.com/article2/0,4149,939886,00.asp . He claims that the Mac maybe first at the plate when it comes to using the IA-64 platform. After reading the article, its possible Microsoft knows this and is changing it's 64-bit roll out strategy. Wasn't there also an article here on Slashdot saying Windows wouldn't be available for the Intel's 64-bit platform until 2006?
This is an odd question, but the answer can be split into 3 cases.
First, you have 64-bit apps running on a 64-bit OS. In order for AMD64 processes to run at all, they obviously need to run in a 64-bit OS. So AMD64 processes have available to them the large 64-bit virtual address space as well as the 8 new AMD64 general purpose registers.
Next you might have 32-bit processes (old code) running in a 64-bit OS in what is known as Compatibility Mode. These old apps will get true 4-gigabyte address spaces (no 3gig user/1gig system split like current systems), but they will NOT have access to the 8 new 64-bit GPRs. This makes sense, because the new GPRs can only be used if the applications were compiled for AMD64 to begin with (i.e., the compiler knew to use them). If you recompile these apps for AMD-64, they would automatically run in Native 64-bit mode discussed earlier.
The last situation that you might have is a 32-bit OS running 32-bit apps, in which case none of the 64-bit features apply at all. This is known as Legacy Mode, and is what you would get if you ran Windows 2000 on your Opteron.
Hope that helps clarify things.
I was getting tired of my old one.
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.
I know a few companies that are moving forward with plans to use opteron or release opteron-based systems that have been until now 100% intel camps. In one case, I know the company *tried* to embrace itanum first, but found to market rather cold to the thought. A few years ago, the market would have folowed intel anywhere with respect to the future/replacement of the x86 family. AMD has really done a top notch job here. For one, the price is such that system makers can enjoy a decent margin, something they haven't been able to do for a loong time with intel based systems. From a technical perspective, it is the logical next step, the power of 64 bit computing without the detriment of lack of legacy. Legacy has left us with some bad things, but it is vital for organizations and companies that cannot afford an intrusive migration. Plus, a lot of the legacy from 386 days no longer necessitates much of an impact to new development as it does with 32 bit systems. Intel dropped the ball. If the market wanted 64-bit computing without caring about compatibility, there is already Alpha, PA-RISC, Sparc, Power4, MIPS, and others. Windows was *not* the reason, the price was. Now with AMD maintaining compatibility and providing the product at a reasonable price target, they will be really hard to beat.
XML is like violence. If it doesn't solve the problem, use more.
The key to this whole issue is .Net, and the CLR.
.NET implementations such as Mono, in order to put a CLR on the Opteron, MS will need an operating system to support it. I'm using "operating system" in the pure sense, as in the stuff that provides hardware access (kernel and drivers), not the GUI with a web-browser definition they used with the DoJ.
That's fine for applications. However, the trick will lie with porting the O/S itself. You know, the stuff that the CLR depends on!
MS will NOT be distributing a version of the CLR for *BSD or Linux.. I guaran-damn-tee it! Excepting other non-MS
The nice thing is, just as Windows 3.1 on the i386 had all the 16-bit thunks for calling 32-bit DLLs in "enhanced mode," MS can take their time transitioning from 32 to 64-bit mode. Once the main kernel and the libraries it depends on are 64-bits, then the apps that NEED 64 bits will work. They can take their time porting the MMC, Notepad, and all the remaining utilities to the CLR. After all, why should notepad.exe be 64-bits?!
THAT is why the Opteron will be a smashing success. Backwards compatibility; just like the i386..
2^32 times the addressing space of 32 bit, so goodbye 4 gig limit.
Actually the Opteron uses 40 bit physical addressing and 48 bit virtual addressing (similar to the Alpha).
This will give you a cap of 1 TB of memory or 256 TB if you want to play funky software games.
And greater speed / precision ratio.
There is no inherent speed benefit to 64 bit computing (read: addressing). In fact there are some significant drawbacks when you are talking about applications that have complex data structures containing a lot of pointers (memory size increases).
Also, they have the exact same precision. You can do 64 bit integer math on a 32 bit CPU, it just takes a little longer if it doesn't have 64 bit registers.
The world is neither black nor white nor good nor evil, only many shades of CowboyNeal.
It's not a very well known fact but the company name 'MicroSoft' was originally coined by a girlfriend who was about to go down on Bill Gates when he was at Harvard....Micro-Soft became his nickname, and has stuck with him (and his company) ever since!
Just another little known fact!!
Its been known for about 4-5 months now that MS was supporting AMD's Opteron over Intel in its new OS's. Microsoft had to pick a side, only AMD had a tangeable candidate for incorporation into their new releases. It seems that Intel believes the 64-bit platform is useless to current users. While this may or may not be so right now, down the road the accuracy of the added address space will become etremely important. It will even affect the gaming industry, games that take advantage of the new address space will not be faster, they will be more precise and realistic. (graphics and physics)
-bb
PRINT "Signature line broken."
GOTO 1
That's similar what I said about my last computer: "256 Megs is tons, I won't have to buy more ram!" Then the next Windows came out.
I hate to pull a newbie here gents but can someone explain some stuff to me....
.. more "bits" at it's disposal to calculate and therefore some calculations will be much faster due to a single calculation or not having to "trick' it's way round a mathematical computation.....
...newbieness but i seek knowledge.....
I understand a 64bit cpu has
I understand it can address more memory also.
but what about the ways in which we speak to the cpu / entire system?
do we need a "64bit" pci slot? will that help - will it only help due to increased bandwidth on the PCI slot or will a 32bit pci slot (even if it was "hacked" to work very fast somehow,.. ) would it slow the machine down?
What about the AGP port- USB hub controller? - the file system itself - does a 64bit exe file contain more lines of code,....... does it contain more physical datA?
do we need a different file system (or rather SHOULD we have, rather than do we need?)
I can see the uses for a 64bit cpu,..... but what else needs to "move over" on the system to help it out?
(I guess I should remember that even 32bit boards had 8bit isa slots......... but they slowed the machine down in general when being accessed right?)
is there a general 64bit MACHINE faq rather than just a 64bit cpu faq?
p.s sorry for the
thank you.
..now with 16 character addresses!
BRAIN has performed an illegal operation in AMD64 at 0123456789ABCDEF. BRAIN will now terminate.
But you must remember, Microsoft is evil.
Who else here remembers the promise of Service Pack 7 for Windows NT 4.0, that NEVER HAPPENED?
I would treat MS announcements as if there was no announcement at all. At least then you won't be lulled into depending on them, for something that never comes.
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
They were soon abandoned once Intel put their foot down, PPC NT didn't even leave beta did it?
That's similar what I said about my last computer: "256 Megs is tons, I won't have to buy more ram!" Then the next Windows came out.
OTOH, there are only 2^265 subatomic particles in the Universe.
Mod parent up. They missed it - it's not on the processor, it's on the CHIPSET.
One word. PALLADIUM.
-----
Hiroshima '45, Chernobyl '86, Windows '98
It is a major blow to Intel, since this makes AMD a much stronger competitor in the server and workstation market. The fact the Intel did not succeed in preventing this from happening indicated that Opteron must be very competitive.
Comment removed based on user account deletion
The main reason the Opteron is a good thing is because 1) it provides MORE registers, allowing the compiler to make smarter register allocations, which can provide drastic performance improvements,
Having more logical registers in the ISA will not make much of a difference. 32-bit x86 processors from Intel and AMD already provide dozens of internal physical registers that are dynamically renamed on the fly.
" No, the Pentium series cannot address 2^36 bytes of memory."
...which is from the pentium pro developer's manual.
I never said the pentium could address 2^36 bytes of memory- I said the pentium pro could.
Your argument seems to be based on out-of-date web pages regarding the pentium series which was to have PAE but the feature was dropped, and later introduced in the p pro.
I blame the fact that the first few hits google gives you are really really old web pages.
Try this:
"The new PAE (physical address extension) flag in control register CR4, bit 5, enables four addi-
tional address lines on the processor, allowing 36-bit physical addresses. This option can only
be used when paging is enabled, using an advance page-table mechanism provided to support
the larger physical address range."
graspee
So there's 'no inherit speed benefit' but it takes longer to do 64 bit math on a 32 bit processor? Sounds like that is a speed difference.
64-bit computing ( using 64 bit address registers ) does increase the addressing space by 2^32 times, I was not referring to the opteron which apparently limits that.
With regard to your comment regarding expanded pointer sizes, there's little reason I can see with why this can't be addressed using short addressing, obviously not every program will benefit or require the expanded address space, so it is feasible only to allow it 32 bit or less.
-Reid
The AMD migration path will allow you to mix 32 and 64 bit applications, and even portions of the same application can use a non-homogeneous number of bits. With Itanium, migration is painful, expensive, and difficult to do completely. With Operton, migration can happen transparently.
I think from the *user's* point of view, this will be an easy decision.
I am not a robot. I am a unicorn.
I am under the distinct impression that Microsoft had very little influence in any part of the x86-64 design.
Listen to the talk about x86-64 given by Kevin McGrath.. I certainly got the impression, from that talk, that this is a purely AMD design with maybe a suggestion here and there thrown in by the Microsoft and Linux community.
(This is very much *unlike* Itanium, where it is my impression that HP had a *large* amount of influence on the ISA design.)
As a programmer, I don't understand why anyone cares about preserving the legacy of 32-bit architecture by expanding it (yet again) for 64-bits.
0 3224711
As a programmer, how much would you charge to reprogram all those apps from 32bit to 64bit? As a CFO, how much would you be willing to spend in one lump sum on converting all your 32bit apps to 64bit ones b/c your new hardware wouldn't support all your old apps? In aggregate, that's a lot of money spent on an essentially non-productive endeavor. AMD's solution may not be technologically ideal, but it is finacially appealing, especially in an economic downturn. It allows a gradual upgrade to cheap low and mid-range 64bit technology. And Opteron and Athlon 64 are not without significant performance improvements either. It's quite competitive on a price/performance basis with the other major 64bit players.
http://www.realworldtech.com/page.cfm?AID=RWT0126
Flying is easy, just throw yourself at the ground and miss. -Douglas Adams
PCI already has plug pin-compatible 32 and 64 bit data width standards (if a 64 bit PCI card is plugged into a 32 bit connector it will reconfigure itself for 32 bit transfers)
NT is short for New Technology not N-ten. That's why I always cringe when I see the win2k bootup screen "based on NT Technology" indeed. Also the NT kernal has had a 64 bit derivitive for several years, including an internal test version for Athlon64 and the commerically shipping Itanium release of XP. I believe Server 2003 will ship with an Itanium version available.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
2K Pro will only use the first physical cpu and the hyperthreading unit of that physical processor if hyperthreading is enabled, which is why you disable hyperthreading in the BIOS if you are running 2k. XP Pro will in fact use two physical cpus and their accompanying hyperthreading units. Server 2003 will likewise recognize hyperthreading units, but the limits are such that if you want to have 8 physical cpus and hyperthreading you will need 2003 Datacenter where you could have run 2k Advanced server on that same unit without hyperthreading.
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
From what I've seen HP is actually improving the Proliants (which is hard to do sense they were so good to begin with). The new hotswap raid memory in the latest rounds of Proliants is simply amazing, with this and hotswap cards there will be almost no reason to take down a server (a failed cpu is about it, and I think that will be their next target)
There are 4 boxes to use in the defense of liberty: soap, ballot, jury, ammo. Use in that order. Starting now.
Yeah, I like humor too.
A firewall can not protect you from yourself. Turn off what you do not need. Do not use the firewall to do your work.
Yep...and it worked so well that they included an x86 emulator and 32 bit libraries -- even on 64 bit Alpha processors that had no native 32 bit support!
The more things change...well, I guess they don't change do they.
A firewall can not protect you from yourself. Turn off what you do not need. Do not use the firewall to do your work.
Which brings me to another point. What happens to buses and whatnot with the x86-64? Has AMD been quietly working away on a 64-bit replacement to AGP? Will we get rid of special graphics buses and go to a next-generation bus standard all round? Or will there be a collection of compatibility hacks to make it all work with existing graphics cards. Anybody care to speculate? :)
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)
For now, at least. Let's never forget the dark ages when many seemed to think a megabyte would never be seen put to use.
I don't have any links handy (and it's 3:45am and I'm trying to code here, so I don't really feel like finding them at the moment) but linux is already up and running on the Opteron and has been for a while. Do a search on just slashdot even and you should find some good info. (IE: the article about UT2003 running under linux on an opteron that ran a while back)
unless you make tricky optimizations, like packing 32-bit variables into a single 64-bit register and doing operations on them simultaneously (which, in general, isn't that useful, BTW).
Don't tell that to Intel and AMD! Or Apple, Motorola and IBM for that matter! What you just described is called SIMD, and it's exactly what MMX, 3DNow!, SSE, Altivec, etc. are designed to do!
IMO it is quite an interesting and rather flexible design, hardware and API-independent - the flexibility gives a performance penalty. But it is poorly implemented, not because of talentless programmers, but due to politics and marketing - thus extra performance penalties.
:o)
Parts of it has been rewritten and most of it with W2K, and I do not blieve that there's much of this flexibility left, the HAL has been reduced to minor hardware vendor optimizations/bugfixes.
Redmond has a long history of taking the quick road of performanceoptimization, all the way back to their OS/2 adventure - everything but the kitchen -sink has been moved into kernel mode - the next step will be moving the applications from user into kernel mode (Well what is left
To sum it up : Interesting design, but the implementation has been messed up due to politics.
Live long and prosper...
16 exabytes ought to be enough for everybody.
16 exabytes ought to be enough for anybody.
16 exabytes ought to be enough for somebody.
16 exabytes ought to be enough for nobody.
I look forward to 4/8-way smp opteron rigs with quad-channel DDR400 support, featuring 4-16 DIMM slots and multiple 64bit/66mhz PCI, multiple gig-e on hypertransport.
Imagine a Beowulf cluster of those...
You are in a maze of twisty passages, all alike.
Galileo: "The Earth revolves around the Sun!"
Score: -1 100% Flamebait
Sorry!
Galileo: "The Earth revolves around the Sun!"
Score: -1 100% Flamebait
I understand a 64bit cpu has .. more "bits" at it's disposal to calculate and therefore some calculations will be much faster due to a single calculation or not having to "trick' it's way round a mathematical computation.....
This is true, but it's quite rare. 32-bit integers give you a range of 4 billion, which is usually more than enough for calculations, however there are some situations where it's not (cryptography is one that jumps to mind). This was a much bigger issue in the change from 16-bits to 32-bit chips. Floating point calculations are almost always done in 64-bits, even on "32-bit" processors.
I understand it can address more memory also.
This is the real reason for going to a 64-bit chip. A 32-bit processor can only address 4GB of memory without resorting to some REALLY UGLY hacks. A 64-bit chip can access some rather ridiculously large amount of memory (1.8 x 10^19 if memroy serves).
but what about the ways in which we speak to the cpu / entire system?
This is more or less independant of the processor core. Every x86 chip since the Pentium has used a 64-bit bus for talking to the rest of the system. The Opteron actually changes this around quite dramatically, using two separate buses. The Opteron has a built-in memory controller to talk directly to memory (the vast majority of I/O traffic for a CPU) and than one or more other buses using AMD's new Hypertransport technology to talk to other CPUs and/or I/O devices.
do we need a "64bit" pci slot? will that help - will it only help due to increased bandwidth on the PCI slot or will a 32bit pci slot
64-bit PCI slots as well as the newer PCI-X slots do offer more bandwidth for I/O, particularly useful on servers for things like RAID cards and gigabit ethernet. However, you definitely don't need a 64-bit chip to use these, you can get 64-bit PCI slots or PCI-X slots on existing AthlonMP or Xeon motherboards.
does a 64bit exe file contain more lines of code,....... does it contain more physical datA?
No and maybe. A 64-bit exe file should contain the same or fewer lines of code, however the address pointers will be 64-bits rather than 32-bits, so it will contain more data. Integers may also be 64-bits instead of just 32, but as mentioned above, this is rarely needed and therefore the default is to keep them at just 32-bits.
do we need a different file system
Not at all. Filesystems already use 64-bit integers when they are needed (they are needed occasionally). For a 32-bit chip, these are hacked as 2 32-bit integers, with a 64-bit chip, they can be done as a single 64-bit one, but otherwise everything remains unchanged.
I guess I should remember that even 32bit boards had 8bit isa slots......... but they slowed the machine down in general when being accessed right?)
This was simply a matter of using ISA. ISA was the most bass-ackwards, poorly designed piece of crap bus ever to see the light of day. Using ISA always slowed the system down, but when it first came out everythign else was so slow that this wasn't a big issue. Once we got 32-bit chips, this had become a major problem, which is why IBM tried unsuccessfully to kill ISA when the 386 came out. Of course, the market didn't bite until 5+ years later when PCI came out, and even then ISA still kept rearing it's ugly head every now and then. Fortunately no current bus is anywhere near as broken as ISA.
Therefore, I wouldn't be surprised to see the RAM limits increased when more ram becomes necessary. Right now, MS publically says due to lack of PAE support in Exchange 2000, there is no reason to put > 3 GB of ram in a E2k box. Once apps start pushing things on a more regular basis, MS will probably have to relent. Right now I think the people who really need 64 bits have it, and are using *nix. People who would like it, find ways to deal. AMD hopefully will commoditize 64 bit computing, and make it something we don't need to worry about.
64 bit processors aren't just about supporting more physical memory. There's a lot of benifit to having a larger virtual address pool as well. People seem to overlook that when they talk about 64 bit processors being useless because people don't need more than 4GB of memory...
Sorry, try again... those are vector processing units, which perform much different operations. They provide specific support for packing multiple datatypes into a single array for performing simultaneous operations on them. 64-bit registers in the Opteron are NOT designed for this!
For example, in MMX you could pack a pair of 8 byte arrays into two MMX registers and add them together. If you tried to do this using 64-bit Opteron registers, you'd get overflow into adjacent bytes, which is NOT a good thing! As a result, you're highly limited on the kind of packed vector-style operations you can do in a 64-bit register on the Opteron.
But what if you are trying to describe the interactions between all of the particles?
In other words, I won't be upgrading to a new processor from AMD until either A> they remedy this problem or B> the sledgehammer processors come down in price considerably from what I now expect them to be.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
The previous poster was not talking about whether the x86 64 bit was an improvement over the 32 bit, they were talking about the x86 architecture as a whole being a 'hack' vs. chips like the PowerPC, Alpha, etc. That of course is another argument.
Linux is a lot easier to port to other architectures for several reasons:
1) Most Unix applications are written to be able to run on a variety of different Unix systems. That means most of userspace was written with some degree of portability in mind.
2) Linux isn't concerned much about binary compatibility. Each major version of RedHat is binary incompatible with the last. New GCC and libc releases regularly break binary compatibility. WinXP will still run software written for the DOS, Win3.x, Win9x, and WinNT product lines. The selling point of the x86-64 over other 64bit chips is that it runs your old software. It would be pointless for Windows to drop the back compatibility.
3) Microsoft has to make sure 3rd parties have time to develop and test drivers on the new system. Remember, not only do the companies have to spend time on it, but Microsoft has to take time to certify the drivers afterwards. I'm not going to get into whether or not certification really means anything, but if nothing else, it takes time for all the paperwork to be processed at both ends.
So there's 'no inherit speed benefit' but it takes longer to do 64 bit math on a 32 bit processor? Sounds like that is a speed difference.
What I was trying to get at is there are 32 bit CPUs with 64 bit registers. In fact, your floating point registers are all 64 bit wide on x86 CPUs. PowerPC's AltiVec unit has a 128 bit wide register for vector operations.
The world is neither black nor white nor good nor evil, only many shades of CowboyNeal.