AMD Opteron Due In April

updog writes "Here's an article from Infoworld claiming that the new 64-bit AMD Opteron is ready to launch on April 22. Some of the notable features of the new chip are an address space capable of addressing up to 1 Terabyte of memory, the ability to link up to 8 processors without any external chips, and backwards compatibility with existing 32-bit applications ..." PapaFSmurf, meanwhile, links to a disclaimer-heavy article posted at amdboard.com which says that 64-bit Athlons may arrive in June rather than September as previously expected.

Comments

What about intel

[Anonymous Coward]

Eh? What about news on their rival procs

Re:What about intel

[TheRaven64]

What about news on their rival procs

The article did include all details of all 64-bit Intel CPUs which are backwards compatible with x86 CPUs without emulation.

Re:What about intel

[cbreaker]

What about Intel? This is an article about one of AMD's products.

When Intel releases something new, it will be posted. There's been countless posts about Itanium, and any number of other Intel stuff.

Get a grip.

wow

[Zanek]

wow, One terabyte for address space . Think of all pron that can hold !!
But seriously, thats great to have the ability to use that much,
but who will is the really important question .

Re:wow

[No. 24601]

Actually, more people than you'd think. Most modern processors are limited to a 4 GB address space and one has to "hack" around this limitation to achieve larger spaces. The reason why I say more people than you'd think is that the 1 TB space would allow one to map any location on the largest hard drives available today to a virtual memory location. There would be a *tremendous* performance boost in doing this as opposed to other contemporary file management schemes.

Re:wow

[MonoSynth]

640 kb blah blah

Re:wow

[JebusIsLord]

Yes, as I recall Linus Torvalds made such a statement recently, saying that the hacks involved seriously hamper x86 servers performance when addressing more than 4gb memory. The x86-64 architechture fixes this issue, which should at the very least allow for some inexpensive and powerful high-end servers.

Re:wow

[Katana2]

Given the address space for a 32-bit (current x86 technology) is 4Gb I think alot of people will benifit from 64-bit. It's not that we will be using one terabyte but instead the ability to pass the current 4Gb barrier.

Re:wow

[edxwelch]

You're talking about accessing memory from a *hard drive*, which takes a massive amount of CPU cycles. In comparision the performance hit of using 32-bit paging instead of 64-bit direct addressing, is only a couple of CPU cycles which is negable.

Whoa

[Bluesman]

Imagine how fast THE INTERNET is gonna be now!

TWICE as fast, at least!

Re:Whoa

[Anonymous Coward]

I think you're confused. Only Intel chips make the internet faster.

Re:Whoa

[Bluesman]

I stand corrected. Thank you sir!

Do these chips have fun in them? Or is that just Intel too?

Somebody had better give me a reason to buy one.

Re:Whoa

[visgoth]

No fun in amd chips. They don't have bunnymen dancing around as they're made.

Re:Whoa

[yakko nef]

Yeah but thanks to this new development by AMD we will have the internet for computers now as once prophesied by the great H.J.S.

Re:Whoa

[Anonymous Coward]

Maybe that's why there faster and cheaper :)

Re:Whoa

[Anonymous Coward]

If you are using Mozilla, the intarnet might actually be twice as fast now.. ;-)

Not surprising

[rabtech]

Barton isn't enough to keep AMD going against Intel until Sept. Simply not gonna happen, and I think they have seen that coming and are trying to head it off by launching the Athlon64 closer to its originally planned release.

Once the Athlon64 is available and people are building systems using it, AMD just stole back the "King of PC processors" title and in a BIG way.

Re:Not surprising

[Anonymous Coward]

#1- They have always been planning on releasing Opteron (server chip) in April. The desktop version is still delayed until September (probably).

#2- Why don't you wait to see how Opteron actually performs before you start throwing around the "King of PC processors" title?

Re:Not surprising

[Anonymous Coward]

Imagine...

Re:Not surprising

[Loki_1929]

The nice thing about Opteron is that it's main competitor will be the P4 Xeon processor, which will soon start showing signs of age. The P4 was designed for high clock speeds, but it can only go so far. When Clawhammer is released this Winter, AMD will truly have a more rounded product line. For the time being, all they have is the Athlon, which seems like a big mistake to me.

AMD's future existence depends very much on the performance of the Opteron and Clawhammer with 32-bit applications, in a 32-bit environment. The best case scenario for AMD would be if M$'s next major OS release were 64-bit only. Unfortunately, this seems highly unlikely, but AMD can dream, can't they?

"Yes, sir, that new system with the latest Windows will be $12,000 if you go with Intel and the Itanium, or $699 if you go with AMD and the Athlon64. Yes, sir, it is pretty funny; I agree."

Re:Not surprising

[Anonymous Coward]

the P4 Xeon processor, which will soon start showing signs of age

And will also be replaced soon (Prescott, Yamhill, etc?)

Re:Not surprising

[IAR80]

I rather see it in the high performance (of course UNIX) servers than PC desktops for starting. It will deliver huge amounts of power but not to desktop aplications. So if you want to use it to play quake you are going to spend lots of money for 10-20% performance increase. If you want high performance servers this is another story. I also see future for it in scientific aplications. 8 64bit Opterons is a lot of FPU. Yummy. I can crunch the numbers much faster.

Re:Not surprising

[einhverfr]

I rather see it in the high performance (of course UNIX) servers than PC desktops for starting. It will deliver huge amounts of power but not to desktop aplications. So if you want to use it to play quake you are going to spend lots of money for 10-20% performance increase. If you want high performance servers this is another story. I also see future for it in scientific aplications. 8 64bit Opterons is a lot of FPU. Yummy. I can crunch the numbers much faster.

Maybe--- maybe not. Yes, there will be a performance increase. However, I don't think the cost will be that much higher in that I think that in a short time, the opteron and clawhammer will become the desktop processors that AMD pushes. This will spread out a huge R&D investment very widely and will enable them to compete in desktop space with server chips.

Um, yeah, we know

[Anonymous Coward]

Slashdot already reported April 22 as the release date, way back on January 31.

This is huge

[Toasty16]

Affordable x86 64-bit servers for the masses, this is going to revitalize AMD and really put it on the map as a serious challenger to Intel. I hope Chipzilla wakes up and sees that its incredibly expensive and backwards-incompatible Itanic 2 chips are the result of engineers developing for themselves instead of developing for the needs of their customers. Finally, AMD will be able to court the high profit business market, though I fear that they might alienate their hardcore enthusiast consumers in the process...

Re:This is huge

[sql*kitten]

Affordable x86 64-bit servers for the masses, this is going to revitalize AMD and really put it on the map as a serious challenger to Intel.

But there already are affordable 64-bit servers for the masses, cheap SPARCs, PPCs and so on. What does anyone need a 64-bit computer for? If you need to address large amounts of memory, or require compatibility with apps that are written to do so. The speed advantages of 64-bit per se simply aren't that great. People who need 64-bit can already get it, the apps are already written, etc. The one and only edge that x86 has is backwards compatibility and the use of very cheap commodity components.

hope Chipzilla wakes up and sees that its incredibly expensive and backwards-incompatible Itanic 2 chips are the result of engineers developing for themselves instead of developing for the needs of their customers.

But Engineers are the only people in the near future who need 64-bit processors on their desktops.

Of course, Intel has proven time and time again that you can use large-sounding numbers to sell things to people who should know better, so perhaps it will work out quite well for AMD.

Re:This is huge

[visgoth]

But Engineers are the only people in the near future who need 64-bit processors on their desktops.

Engineers are not the only ones. Visual effects people will snap up itanium and athlon64 based machines quite readily (I know I will!). I have a feeling the release of these new processors will be the final nail in sgi's coffin in the cg industry (yes they will still be around for other stuff, but for visual effects they are irrelevant now.)

Re:This is huge

[MonoSynth]

incredibly expensive and backwards-incompatible Itanic 2 chips are the result of engineers developing for themselves

Not for themselves, but for a market without serious competitors, like when they started with IA64.

Re:This is huge

[Anonymous Coward]

affordable? call any of the places taking opteron pre/sample-orders now, and you'll quickly find out that the two-way boxes cost about the same as a two-way itanium 2 box.

Re:This is huge

[paitre]

And to add to this, folks doing bioinformatic and genetic research will -=really=- put the Opteron to use, and for a -hell- of a lot less than Itaniu (and I2).

And no, Engineers -aren't- the only ones that could use them on their desktops. You also forgot graphics designers and testbed machines for applications programmers, and testbed machines for a larger 64bit environment. :P

Re:This is huge

[Natalie's Hot Grits]

"But Engineers are the only people in the near future who need 64-bit processors on their desktops."

To add to the list of this siblings' posts, most newer video games from the top manufacturers are also going to need 64 bit. Unreal will soon require 64 bit for their mod development tools, I'm sure id software is also going to have no problem improving performance on 64 bit platforms.

Did anybody really need 32 bit CPU's when intel went from 286 -> 386? Maybe not, maybe so. But that is what happened, and now everyone uses 32 bit, and needs AT LEAST that. It's only evolution that we move to 64bit cpu's, more general purpose registers, and shed the excess bloat that AMD is doing.

Sure, you might not find a use for it yourself, and if you haven't, you should keep buying 32 bit cpu's. But for me, and many more gamers and game developoers/modders (who are the ones that have pushing the desktop performance barrier higher and higher over the years) are going to use them, and NEED them. I plan on running the latest video games on my 64bit athlons when they arrive.

"But there already are affordable 64-bit servers for the masses, cheap SPARCs, PPCs and so on."

If you consider cheap SPARCs, PPC,s and so on to be cheap, high performance, and usefull on the desktop, you have another thing coming. For one, there are no, and never will be, drivers for commodity hardware for these platforms. Secondly the chips that offer reasonable performance aren't "cheap" as you call it. They are very very expensive, and their platform is also very very expensive (relative to x86-64, and x86-32). And there is only a small handfull of software packages(that the mainstream doesn't use, gamers can't use, etc..) that run on them compared to what AMD is offering.

"The one and only edge that x86 has is backwards compatibility and the use of very cheap commodity components."

though this is not the only edge x86 has (ie, look at its high competition market, which drives performance up way faster than the platforms you mentioned previously) you have hit this nail right on the head. One of the biggest advantages x86 has is its compatibility with commodity components. Untill those SPARC, 64bit PPC, and Alpha machines run my webcam, my USB Scanner, my USB printer, my IDE hard drives for a reasonable price without having to buy proprietary and expensive compatibility components, your argument stands mute.

Re:This is huge

[Ryan Amos]

Right on. The main speed advantage of a 64 bit processor is that it processes 64-bit integers faster. There really isn't THAT much difference in speed for applications that use 32 bit integers. Most desktop PC applications use 32 bit ints and larger floats (which are already handled by SSE2/x87) The only real gain in speed will be in applications that use 64 bit ints, namely scientific computing.

In other words, 64 bit processors allow you to process LARGER numbers, but not more of them. Yes, 64 bit has a use, but right now I really don't think the home market is ready for a 64-bit platform. The workstation market.. maybe. But the vast majority of PCs bought are home/office PCs, which definately don't need 64-bit yet. MS word runs just fine on 32-bit systems. It probably won't be much faster on a 64-bit system. Neither will most games. Now, there could be some benefit if the PLATFORM is upgraded, like faster bus speeds (hypertransport, etc) but the CPU itself probably won't speed too much up. So basically, 64-bit isn't twice as fast as 32-bit. Any performance differences will probably be marginal.

Re:This is huge

[Sgs-Cruz]

ARGHH. I _love_ Slashdot and _hate_ it too.

Yes, I realize it's different people posting. Yes, I realize there is no group consensus on Slashdot. But it's so frustrating when every single freaking article that involves something with backwards compatibility has 40 people saying "Companies should just drop the cruft. I mean, these processors have legacy code hanging off them like sludge or something. Go for performance, not backwards compatibility"

And then another 40 people say 'The Itanic (ooh... I bet that name comes from the same school of thought as calling Microsoft 'M$') dropped backwards compatibility, therefore it sucks. Go AMD for doing x86-64.

From everything I've heard, the IA-64 architecture is a bitch to code for, but is really, really fast, and good. Expensive, though. So maybe it's good for a different application than x86-64? Basically, is backwards compatibility good, or not? Performance, or legacy support? Okay, rant over.

Re:This is huge

[be-fan]

You miss a very simple point here: The benifets of 64 bit outweigh the costs. Making a 64-bit machine is not expensive at all. Most parts of the system are already 64-bit or wider. The only thing is the increase the widths of the ALUs, AGUs, and address lines. It's almost trivial, to the point where you see even cheap game machines with 4MB of RAM (the N64) and 16MB PDAs (Casio EM-500) having 64-bit CPUs. In fact, if it weren't for x86 architectural restrictions, we'd already be using 64-bit CPUs. Now the benifets to 64-bit are significant. You get slightly better performance for certain applications. You get to address more than 2GB (the practical limit for a 4GB virtual address space) of physical memory, which will be a limit within the next couple of years. You get to eliminate lots of stupid code inside the OS that has to manage with not being able to map all of physical memory at the same time. You get to do nice stuff like prelink all binaries (important for stuff like KDE) without worrying about running out of address space. So the question isn't: why 64-bit. It's: why not?

Re:This is huge

[sql*kitten]

Did anybody really need 32 bit CPU's when intel went from 286 -> 386? Maybe not, maybe so. But that is what happened, and now everyone uses 32 bit, and needs AT LEAST that. It's only evolution that we move to 64bit cpu's, more general purpose registers, and shed the excess bloat that AMD is doing.

The addition of hardware MMU (preemptive multitasking and protected memory) and on-chip FPU meant that a 386 added considerable value over a 286. The one advantage of 64-bit processors like Opteron is that it can address >4G of memory. It's more like the jump from P2 to P3 in terms of actual impact on the market.

Sure, you might not find a use for it yourself, and if you haven't, you should keep buying 32 bit cpu's. But for me, and many more gamers and game developoers/modders (who are the ones that have pushing the desktop performance barrier higher and higher over the years) are going to use them, and NEED them. I plan on running the latest video games on my 64bit athlons when they arrive.

You're focussed on performance, but performance isn't an inherent property of 64-bit-ness - memory is. Sure, more memory is good, and 256M or 512M on the desktop is common now, but no game developer can afford to require >4G of memory.

If you consider cheap SPARCs, PPC,s and so on to be cheap, high performance, and usefull on the desktop, you have another thing coming. For one, there are no, and never will be, drivers for commodity hardware for these platforms. Secondly the chips that offer reasonable performance aren't "cheap" as you call it. They are very very expensive, and their platform is also very very expensive (relative to x86-64, and x86-32). And there is only a small handfull of software packages(that the mainstream doesn't use, gamers can't use, etc..) that run on them compared to what AMD is offering.

Not really, you can buy a Sun Blade for roughly the same price as a PC. Gamers tend to see themselves as the most important part of the market, but the most important market is actually the corporate desktop. People who, every now and again, buy PCs by the thousand or the ten thousand.

hough this is not the only edge x86 has (ie, look at its high competition market, which drives performance up way faster than the platforms you mentioned previously) you have hit this nail right on the head. One of the biggest advantages x86 has is its compatibility with commodity components.

If Intel and AMD are smart, they'll do it like Apple - essentially throw away their old hardware and rely on emulation for compatibility. It is inevitable that AMD will have had to make compromises to ensure backwards compatibility, and those will hurt it in the long run.

Re:This is huge

[Natalie's Hot Grits]

"The addition of hardware MMU (preemptive multitasking and protected memory) and on-chip FPU meant that a 386 added considerable value over a 286. The one advantage of 64-bit processors like Opteron is that it can address >4G of memory. It's more like the jump from P2 to P3 in terms of actual impact on the market."

Actually, since you didn't do your research first, I'll enlighten you with what "going to 64 bit" means.. In AMD's case, with x86-64, not only are the general purpose registers doubling in size, but they are doubling in number. And since they are double in size, and the circuitry is there, you can do math on 64 bit integers in the same time as you can do 32 bit integers on a 32 bit machine. When adding more general purpose registers, this means the CPU doesn't have to access main memory nearly as much, and the same holds true for L1 and L2 cache. Top this off with the fact that AMD is doubling its L2 on chip cache, and you have yourself a much much much faster machine. Even if you run it in legacy 32 bit mode, you still have a faster machine than an Athlon. Run in native 64 bit mode, and you have 16 General purpose registers, plus all the other advantages mentioned above. To top this all off, AMD has improved on their SIMD units, and it looks as if they will be able to do floating point math much faster than traditional x87 code if developers choose to use it. (note, this also holds true for intel's latest offerings)

"you can buy a Sun Blade for roughly the same price as a PC. Gamers tend to see themselves as the most important part of the market, but the most important market is actually the corporate desktop. People who, every now and again, buy PCs by the thousand or the ten thousand."

So what if I can buy a Sun Blade for roughly the same price as a PC? Are you going to use that on the home desktop? (which BTW, is the second biggest market for PC's our there, second only to coporate desktops). You think nVidia made all their money on coporate desktops? I don't see my geforce 4 running on any Sun Blades around here...

Secondly, I never said that gamers were the most important part of the market, I simply stated the fact that the video game market is traditionally what pushes the home PC to a faster and faster standard.

"'One of the biggest advantages x86 has is its compatibility with commodity components.'

If Intel and AMD are smart, they'll do it like Apple - essentially throw away their old hardware and rely on emulation for compatibility. It is inevitable that AMD will have had to make compromises to ensure backwards compatibility, and those will hurt it in the long run."

If you will re-read my post, I stated exactly what you think AMD should do. With the x86-64 platform, AMD has discontinued real modes, and other outdated parts to the x86 ISA to shed the bloat. But you are still mistaken on what exactly compatibility means. I never said anything about old hardware I needed to be compatible with. I mentioned USB scanners, printers, and webcams. Since when are these old hardware, and since when does apple not support the same things? Where EXACTLY is the 64 bit system from Sun and everyone else that will run this COMMODITY hardware that is in every computer store on the planet? Short answer: there isn't.

In closing, you have not refuted a single statement I have made, and insted went off on a tangent argument in every single rebuttal you made. You haven't shown anything except repeat your origional premise that nobody needs 64 bit, and have given no new evidence, or argument. Next time maybe you should try to think about it before you get labeled as a troll.

Re:This is huge

[Ex-MislTech]

One thing you are missing .

Athlon was designed using technologies from
DEC Alpha 21064 and 2162 RISC processors.

Their "farther" Dirk Meyer came to AMD and
brought in an engineering team who succesfully
developed the Athlon .

The lead design engineer for the Athlon
was the engineer that built the DEC Alpha EV6,
and thus a higher base architecture
was designed in and was already forward thought
for 32bit and a 64bit world .

This was not well known by most ppl outside
of design circles, and hardcore hardware ppl .

The former DEC designer has insight into
64 bit design when Athlon was released,
and the fact that the old irongate boards
were based on the DEC EV6 lend credence .

A big suprise may be coming if they pull this
off, and I cannot wait to see it if it works .

Benchmarks for 1st pass BIOS's are "always"
lower until they are tweaked and optimised .

For that matter the OS's often have to be
tweaked to take full advantage of them .

In time the Clawhammer, and Sledgehammer will
show their true abilities .

Peace...
Ex-MislTech

Re:This is huge

[Ex-MislTech]

Another point I forgot to mention, do
not mod this up, I will look like the
almighty karma whore .

Microsoft has mentioned its preference
of the AMD 64 bit architecture, why do you ask ??

Perhaps it is the fact that they have written
a 64 bit version of their OS before for the
DEC Alpha, and that is what AMD used for the
foundation of their Athlon .

This will be a factor I am betting .

Peace...
Ex-MislTech

Re:This is huge

[You're All Wrong]

"
64-bit servers for the masses, cheap SPARCs, PPCs
"

PPCs aren't 64-bit. _Power_ are 64-bit, but PPCs are 32-bit.
The next generation of PPCs will be 64-bit.

However, you're right, you only need to look at e-bay to see cheap old (ten years old, some of them) 64-bit machines: Alpha, HPPA, Sparc, MIPS. Most are Sparcs, admittedly.

YAW

Just in case

[Anonymous Coward]

AMD's enchanted April
AMD's 64-bit alternative

By Tom Yager March 14, 2003

After years of hype, the AMD Opteron 64-bit processor will debut in April. The company and its shareholders might curse the rotten timing, but the current contracted market is actually the perfect setting for AMD's new technology. While other chipmakers scramble to adapt, AMD seems to have designed current business challenges and priorities into its architecture. Considering how long Opteron has been in engineering, AMD is either very smart or very lucky. Opteron may be an opportune solution for customers looking to consolidate their servers and reduce operating costs.

The advantages of AMD's new design are many. The most talked-about feature is the CPU's support for 64-bit applications. Unlike previous 64-bit processors, Opteron implements the full x86-32 instruction set. Software that runs on a Pentium III or AMD Athlon now will run unmodified on Opteron. Opteron-based servers will likely spend the majority of their time running the 32-bit Windows and Linux programs that businesses use today.

Software written to exploit Opteron's 64-bit capabilities will break through the barriers that prevent the x86 from running extremely demanding server and technical applications. A vastly expanded address space (up to 1TB of physical memory), a larger set of high-speed registers, and new instructions will take affordable servers to a higher level of performance. Running in 64-bit mode, an Opteron application can crunch through mountains of in-memory data and perform blazingly fast data transfers to network and storage devices.

Unlike other x86 processors, the Opteron CPU has the inherent ability to link up to eight processors without specialized chips. Every processor has three HyperTransport bus controllers for fast connections to other CPUs and devices. Instead of using an external memory controller, which complicates system design and adds latency, AMD links memory directly to each CPU. The design has plenty of headroom to accommodate faster memory and I/O devices. The only speed limit is the 19.2GB per second capacity of each chip's combined HyperTransport channels, which exceeds the top speed of the most capable PC server bus.

In systems that require more than eight processors, Opteron will rely on external chipsets to provide communications between CPUs. The fact that HyperTransport is already on the chip simplifies the engineering. Systems running two- and four-CPU configurations -- which account for most x86 server sales -- will ship in 2003. How soon larger systems appear depends entirely on market demand.

Answering critics

The chief criticisms leveled against the platform by Intel and critical analysts -- mainly that Opteron is immature technology and that Microsoft is dragging its feet porting Windows to it -- will prove groundless. The well-respected and thoroughly debugged Athlon x86 processor is the foundation of the Opteron chip. The remarkable HyperTransport bus that AMD uses to tie Opteron chips to each other and to I/O devices is already in widespread use. The DDR (double data rate) memory that AMD has chosen for its first implementation is inexpensive and readily available. AMD's chipset implements standard PCI-X and AGP ( Accelerated Graphics Port ) peripheral buses. System manufacturers and customers will have relatively few adjustments to make.

The Windows question is slightly trickier, but it isn't an issue Intel should press too hard. Yes, the sole 64-bit version of Windows runs on Itanium and Itanium 2. However, Microsoft has repeatedly stated that it strongly prefers AMD's 64-bit architecture to Intel's. Opteron is not stuck in the same spot as Intel at the launch of Itanium. Intel had to wait for Microsoft to announce its Itanium-specific port of Windows . Opteron already runs 32-bit Windows at full speed, while other 64-bit CPUs must use emulation to run most Windows software. Microsoft's engineering task, which it needn't hurry to accomplish, is to

Re:Just in case

[Anonymous Coward]

He (or you) didn't have to be so stupid.

No June Launch

[ruiner5000]

There will not be a June 3rd launch of the Athlon 64. It isn't going to happen. June 3rd doesn't even line up with Cebit America dates. Major AMD product launches almost always correspond with some trade show, and an release of this size definitely would.

AMD's naming scheme...

[DJPenguin]

Duron? Opteron? Who comes up with these? They sound like types of rubber, or possibly fuel additives...

Re:AMD's naming scheme...

[JudgeFurious]

I think they sound like Decepticons personally but maybe that's just too much time spent watching Transformers as a kid.

Re:AMD's naming scheme...

[IAR80]

Let me remind you of Celeron, Deshutches, Klamath, Mendocino .... It might be the same guy!

Re:AMD's naming scheme...

[misof]

Duron, Opteron.. I suppose the next one will be called Moron ;)
(after considering other names like Dodecahedron, Rhododendron, CowboyNeal-drone, etc.)

Re:AMD's naming scheme...

[Selanit]

Makes sense to me. The -on suffix has a nice ring to it, sort of techy in a Greek kind of way. Plus it's an English word in its own right with generally positive connotations, turning you on to the product.

The prefixes are just common sense, conveying to the potential buyer what the company would like you to believe about their product. Athlons are Athletic. Durons are Durable. Opterons are Optimized for Optimum performance, suggesting that it's the best of the best. The Latin optimus, from which the prefix is ultimately derived, simply means "best".

Course, they had to throw an "er" into that last one, 'cause otherwise it would have sounded like some sort of legion of boxy doom robots in a low budget SF invasion flick: Oh my God, the Optons are coming! Run . . . RUN!

But even then that extra joining syllable was carefully chosen. They could have picked an i, making the product Option. That would pick up on the Opti- prefix of optimum and optimize, but it would also make the word an English word that merely means "one possible choice", clearly not the best message to send to a potential buyer looking for the best.

-er- works much better. Not only does it call to mind the English comparative adjectival ending ("It's not only opt, it's opter!), it also rhymes with the middle syllable of their highly successful Duron line, suggesting by extension that these new chips will carry on that tradition of excellent value for the money.

Over-analysis, you say? That which we call a CPU would perform its function by any other name? Certainly the silicon would work the same; but the perception of the chip would be different, perhaps worse for AMD, and for that reason the name matters. Names are words, and words are how we define reality. You might ask me "What's a figgin?" And I would say "A figgin is a type of pasty, with chicken inside and raisins on top." Prior to that you probably had no notion what a figgin was, and I have now created in your mind the idea that it's a chicken pasty with raisins on top, and I may also have conveyed the idea to you that it is tasty and filling. And I have done it using words. This is exactly what the AMD marketing people are trying to do: they are using words to create a reality in which people believe that AMD's products are superior (Ha! there's the next one: the Superion), and buy from AMD rather than its competitors as a result.

Typical attention spans are short, moreso now than before we got so saturated with advertising, so they have to pack as much meaning into as few words as possible. Figgin is a rotten product name. It has no linguistic history; it could just as easily be a bodily organ, often removed in unpleasant ways by sadistic tyrants. Athlon, Duron, Opteron are much better: they are made from phonemes that have positive connotations to English speakers, and in several other languages as well, notably the Romance tongues.

Sensible buyers will not make purchasing decisions based solely on this, of course, but the initial impression remains and is reinforced every time you see, hear, or think the name. That counts . . . it may not be logical, but humans are not always logical. (Seldom, in fact.)

Basically, I think AMD's marketing team has done an outstanding job picking these names. Even you, oh parent poster, must have picked up on the positive connotations, even if you then realized how silly the whole thing is and mocked them for it. AMD's going to need every advantage they can get to win serious market share from Intel, but if their naming team is anything to judge by, then they've made a good start.

Re:AMD's naming scheme...

[Anonymous Coward]

Better than the Pentium 7, eih?

Re:AMD's naming scheme...

[Anonymous Coward]

Lots of Intel code names are geographical features in Oregon and the Pacific Northwest. Deschutes River, Klamath (Falls/County), Willamette River, Yamhill County. Etc.

Note that Intel doesn't use these names for release processors. They use the same sorts of words as AMD. Also, they seem to have a fetish for names that sound vaguely like elements - Pentium, Xeon.

Re:AMD's naming scheme...

[UniverseIsADoughnut]

>> I think they sound like Decepticons personally but maybe that's just too much time spent watching Transformers as a kid.

Glad to hear i'm not the only one who thought that.

Re:AMD's naming scheme...

[Anonymous Coward]

Course, they had to throw an "er" into that last one, 'cause otherwise it would have sounded like some sort of legion of boxy doom robots in a low budget SF invasion flick: Oh my God, the Optons are coming! Run... ...They could have picked an i, making the product Option.

When I first saw "Optons" I'd already parsed it as "Options", as in "My AMD options aren't worth shit...".

You might ask me "What's a figgin?" And I would say "A figgin is a type of pasty, with chicken inside and raisins on top."

And when I saw "figgin" my mind read it as "friggin". Like when my CPU fan died, and my friggin AMD fried itself.

I've got to say that I'm looking forward to the Opteron though, because getting an Itanium isn't a friggin option.

Re:AMD's naming scheme...

[TeknoHog]

IMHO it should have been Octeron, since it's the 8th generation CPU, aka K8. Though I admit that Opteron sounds cooler in an anime/mech sense.

Also, Athlon is a nice name for multiprocessor systems, even if you'd have to settle for a Biathlon instead of Triathlon.

As for Athlon64 (the consumer/desktop version of Opteron), it'll be fun it people will call them simply 64s because that's how the C=64 was generally known (at least here in Finland).

Re:AMD's naming scheme...

[moertle]

The first time I was in a real chemistry class was 8th grade not too long after the Pentium was released. The first day I kept scanning the periodic table to see what the symbol for Pentium was. :)

Re:AMD's naming scheme...

[ianezz]

Opterons are Optimized for Optimum performance, suggesting that it's the best of the best. The Latin optimus, from which the prefix is ultimately derived, simply means "best".

To me (and my newphew), "opteron" mainly evokes the image of a pterodactyl of the jurassic age cross-breeded with an eigth-legged monster of some sort (which is relevant, since the opteron can do 8-way SMP).

My mother tongue is italian, should I worry if I didn't catch the "optimum" pun at all until I read your message? :-)

Re:AMD's naming scheme...

[0x1337]

Decepticons forever..... Hey, you never know when your mobo will transform on ya.

Only 1 TB?

[MagPulse]

64 bits should give you 18 exabytes right? What's the limiting factor?

2^64 = 18,446,744,073,709,551,616
^EB ^PB ^TB ^GB ^MB ^KB

Re:Only 1 TB?

[cheezedawg]

The "limiting factor" was AMD's decision to only use 40 bit addresses at first.

Re:Only 1 TB?

[vlad_petric]

I'm pretty sure this is the physical address space and not the virtual one (which should be 2^64). So the external bus of the processor can "only" address 1TB (i.e. 40 address bits)

Re:Only 1 TB?

[sirsex]

Appears that they will use a 40 bit address bus. Since, for now, the 14 MSB bits could never be used, (jeeze, its aready 1000 sticks of 1G RAM) they can save the routing space and pin count. Every 18 months, just add one more bit line. Stir as needed.

Re:Only 1 TB?

[Tailhook]

Can they provide more bus lines in the future without ISA modifications? If so, then this is a smart design decision. Using only 40 bit addresses reduces the cost of the complete system while providing enough address space.

Re:Only 1 TB?

[Anonymous Coward]

oh shit i smell another keyboard controller a20 line hack.

Re:Only 1 TB?

[Anonymous Coward]

Re:Only 1 TB?

Don't worry, 1 Terabyte is the theoretical maximum of the chip in its first spin, basically the Opteron 1 X86-64.

The Opterion 2 X86-64, and Opteron 3 X86-64 should greatly expand that.

Besides, if you're reading slashdot you should know that 640GB should be enough for anybody.

Mmmmmmm, 64bit DOS nuts.

Re:Only 1 TB?

[Anonymous Coward]

I'm pretty sure this is the physical address space and not the virtual one (which should be 2^64).

Nope. 40 bit physical addresses, and 48 bit virtual addresses.

Re:Only 1 TB?

[Anonymous Coward]

The 14 MSB? For a total of 54 bits? Dumbass.

Re:Only 1 TB?

[JamieF]

So, if you like storage systems that can hold >1024 TB...

does that make you a "peta-phile"?

Yuk yuk yuk.

Re:Only 1 TB?

[Anonymous Coward]

Don't worry, 1 Terabyte is the theoretical maximum of the chip in its first spin, basically the Opteron 1 X86-64

I thought the 40 bit physical address was defined in the x86-64 spec, so increasing it will require a change to the architecture.

Re:Only 1 TB?

[BrianB]

Yes, much like the original 68k (68000, 68010) chips could only address 16 meg, but the instruction set was 32 bit and therefore able to go to 4Gig with subsequent chips without any problems (well, except the Microsoft-written AmigaBasic used the 8 upper bits as flags and therefore broke when the 68020 wanted to actually treat those bits as part of the address....horrible, horrible kludge).

Anyway, future opterons will be able to address a larger amount of memory without modifying the instruction set, and let's face it, by the time 1 TB of memory is affordable/useful, that original opteron is going to be long surpased.

Re:Only 1 TB?

[FPCat]

By using 40 bits, they save on the memory space every pointer uses... Using 64 bits, they would double the space used vs the current 32 bit system.

Re:Only 1 TB?

[dreamchaser]

No, they would not double it. A 64 bit address space is 4 billion times the address space of 32 bit land. Even 40 bits is many (2^8 times more to be exact) than 32 bitness.

Re:Only 1 TB?

[dreamchaser]

DOH! I just re-read your post, and you are correct. I misunderstood. Yes, using only 40 bits instead of 64 reduces pointer size slightly, thus overall code size. It also makes it easier from a motherboard design standpoint.

Re:Only 1 TB?

[BusterB]

It still uses a 64 bit address pointer register, and all addresses are 64 bit with the Opteron, code-wise. This makes it simple to, down the line, add more physical bits and have code automatically be able to address more memory. Integer sizes default to 32bit with it, for code size (most programs only need 32 bit), but a long will be a lot faster to deal with if code needs it.

Re:Only 1 TB?

[Anonymous Coward]

Ahh.....
Good times........

Re:Only 1 TB?

[eht]

If you had said 1000 > TB you'd be more correct, as a peta is 1000 according to the IEC
a PebiByte is 1024 TebiBytes

Re:Only 1 TB?

[Weirsbaski]

The linear (virtual) address is 64 bits, but only 48 bits are implemented. This means that pointers will only have the bottom 48 bits "arbitrarily" chosen. (The upper 16 bits are a sign-extension of bit 47). Future x86-64 revs can implement up to 64 bits if desired. Advantage: only 48 wires are needed to pass linear addresses around within the cpu.

The physical addresses are 52 bits, but only 40 bits are implemented. This means that the page tables can only assign pages to 40-bit physical addresses. Future x86-64 revs can implement up to 52 if desired. Why 52? The upper bits in the page tables that would be used for larger addresses are instead marked "available for software use". Advantage: only 40 wires are needed to pass around physical addresses, the caches only have to store 40-bit physical tags.

So in theory, one task could use 2^48 bytes of memory, but only 2^40 bytes would be in memory at any one time, the rest would be swapped out. The virtual-memory-manager (not the task iteself) would be responsible for keeping track of which pages are currently in memory.

Re:Only 1 TB?

[Anonymous Coward]

640 GigaBytes should be enough for anyone.

Re:Only 1 TB?

[Anonymous Coward]

afaik, you can probably even point up to the high areas. there will probably be a 40bit (physical) ->64bit (virtual) address translation table like how 32bit stuff was done. (that's why most windows calls are made when the MSB=1. ex: 7FF0C000)

Re:Only 1 TB?

[0x0d0a]

I seem to remember software breaking when the extra eight bits were used, and Apple having a "32 bit clean" flag in the Memory control panel to kick the system into compatibility mode.

I just keep thinking of hard drives and all their backwards compatibility issues...

Re:Only 1 TB?

[morcheeba]

ha ha. Sony already dodged this problem with their petasite, which has been around for years.

Re:Only 1 TB?

[be-fan]

It's even more funny than that. It's a 48-bit virtual address mapping to a 40-bit physical address. The remaining 16-bits of the virtual address are reserved for future use, so all pointers have to be 64 bit, but you can only map stuff to the first 48-bits. This is done to keep the page table to a sane number of levels. Even with the 48-bit virtual address, it already uses a 4 level page table, because they stick with the stupid 4KB page size.

Re:Only 1 TB?

[You're All Wrong]

It;s been a long time since I did the maths, but the user-level virtual address space is about 8192 tasks, 8192 descriptors per task, 2^64 bytes per selector. i.e. 2^90, or 2^77 per task.

The linear address space is still of course 2^64, but user-land doesn't get to see linear addresses, only virtual (descriptor:offset) ones.

Of course, many OSes don't implement memory protection via descriptors, and so each task will have just one 'segment', and rely on linear address protection facilities.

YAW

Re:Only 1 TB?

[Hoser McMoose]

No, the pointers are still 64-bits, it's just that the top ~20 of those bits will be rather useless for the time being (the Opteron supports 40-bit physical addresses, but something like 45 virtual address lines). The previous poster was correct, the idea behind this was that the number of pins for the address bus.

To answer the previous poster's question, yes, the chip can be expanded to right up to the full 64-bits of physical address space without any changes to the ISA.

Where are they getting this information?

[handsomepete]

The only thing the article references is that Newisys is leading the charge, but I don't see anything I would consider a source. Racksaver is claiming that evaluation units are available now, but mostly there's just a lot of Opteron Server Evaluation signups. Does anyone have any real information? Cost? Non-evaluation release? Anything?

Re:Where are they getting this information?

[fobbman]

Why are you doubting them? It's on /. now, and that's all the proof that I need.

Re:Where are they getting this information?

[DeadMeat (TM)]

Why are you doubting them? It's on /. now, and that's all the proof that I need.

Plus, if you need more evidence, you can just wait for the story to be reposted.

Re:Where are they getting this information?

[Anonymous Coward]

You can buy Newisys evaluation systems now, but they go for about 10k for a dual Opteron 1U rackmount (we actually got a quote for it about a month ago with a three weeks lead time). The same vendor also gave us a prelimary quote for their own dual Opteron 1U unit which was about 2.5k. That prelimary quote stated shipment on/after April, 22.

Oh, here it is.

[handsomepete]

AMD.com

Oh wonderful...I just upgraded

[Guysdrinkingbeer]

No, I ma not making a joke. Yesterday, Friday March 15 2003 I bought my new Soyo KT400 Dragon motherboard and a AMD 2600 Athlon with 333 bus. I have been cutting edge for 24 hours.
I never will win.

On a more positive note, any body know of motherboards for these monsters yet?

Re:Oh wonderful...I just upgraded

[Anonymous Coward]

The best thing is not always the Right Thing.

Re:Oh wonderful...I just upgraded

[Pharmboy]

No, I ma not making a joke. Yesterday, Friday March 15 2003 I bought my new Soyo KT400 Dragon motherboard and a AMD 2600 Athlon with 333 bus. I have been cutting edge for 24 hours.

Um, I hate to break this to you, but you should have read the . You weren't cutting edge for even a day. Close, but no cigar.

I never will win.

Yea, that pretty well sums it up if you want state of the art.

Re:Oh wonderful...I just upgraded

[Anonymous Coward]

But were you seriously ready to plunk down the money for a Opteron? Since you didn't even buy a Barton 3000 it doesn't sound like your even close to being able to afford an Opteron. So why the regrets? You obviously wounldn't have bought one.

I think a lot of AMD users are misguided if they think 64bit chips are going to be what most users would call "afforardable" in the next year. Corporations might get ahold of these but only the most hardcore of hardcore will spend what's necessary to go 64bit x86 for a while.

Re:Oh wonderful...I just upgraded

[Anonymous Coward]

i love it. there is an inversely proportional relationship between the likelyhood to use PREVIEW and the odds of writing clean code in posts.

Re:Oh wonderful...I just upgraded

[Dylan2000]

I think a lot of AMD users are misguided if they think 64bit chips are going to be what most users would call "afforardable" in the next year.

I'm an AMD user
I don't believe I'm misguided
I know many computer users
I don't know one person who has referred to the chip professionally, casually, verbally, or in any other form, as "afforardable".

Where are you getting your informatinion?

sorry, couldn't resist ;)

Re:Oh wonderful...I just upgraded

[Tokerat]

By the way, besides your HTML skills being a bit under par, you forgot to mention to him that yesterday was the 14th.

;-)

Re:Oh wonderful...I just upgraded

[Anonymous Coward]

"I'm an AMD user
I don't believe I'm misguided
I know many computer users"

Good for you.

You should have resisted. ;)

The AMD fanboy sites as well as /. has been buzzing over 64 x86 for a long time now. Many AMD users have been acting as if they'll actually be buying them. Stop by any Amd forum and read posts saying how people "can't wait" till these chips come out. Most is it is fanboyism because they're hoping they'll be faster than Prescot, thus the bragging. A lot of it is talk whether to buy now, or wait to buy 64bit chips.

My information is from AMD forums and tech review sites all over the net. Initial AMD 64 bit chips will very expensive and certainly initially won't be bringing 64bits to the masses.

Re:Oh wonderful...I just upgraded

[Dunkalis]

It won't be as cheap as other x86 chips, but remember, it is NOT an x86. Opterons will be expensive, but their desktop/workstation counterparts, the Athlon64, probably will be competitive with the fastest Intel chips in price and performance. If AMD isn't fudging its numbers, and a 2000 MHz Opteron is competitive with a 3600 MHz Pentium 4, I think that the future is bright for AMD and their chips. I do want one, and I will buy an Athlon64 machine to run Linux and Windows. If AMD can keep the prices on the Athlon64s comparable to the Pentium 4 or Barton, people will buy them. Hammer is not just for high end workstations and servers, AMD is aiming it at desktops too. This will replace their current line of Athlons, so I should hope its priced well.

Re:Oh wonderful...I just upgraded

[Anonymous Coward]

I never will win.

Nobody ever does. Somebody else's penis is always bigger.

Re:Oh wonderful...I just upgraded

[Dylan2000]

too bad your information doesn't come based on understanding the joke...

(hint: "informatinion" is not a word)

btw, your information, from various tech sites all over the net is as "no shit" as it comes. been around long? it's been the same for every chip generation since the 8088 and yes, of course it is fanboyism, and no, I don't give a shit.

Re:Oh wonderful...I just upgraded

[mcrbids]

I've always upgraded by the rule of "order of four". This gives me what I need at minimal expense.

I generally hold off upgrading until I see at least 4x performance improvement, barring special needs. My recent upgrade from a 450 Mhz PII >> 1800 Athlon is no exception. I probably won't upgrade again until at least 4 Ghz unless there is a real need for it.

Also, I never buy the "top end". If you look at prices, you'll find that prices start kinda cheap, rise slowly for a while, and then suddenly climb, as you go from low to high end.

The last item before that spike is the one I buy.

Notice: hard drives. (pricewatch.com =)

$49 = 20 GB
$58 = 30 GB
$59 = 40 GB (who'd get a 30?)
$67 = 60 GB
$77 = 80 GB
$100 = 100 GB (small spike = 80'd be ok)
$106 = 120 GB (don't bother with the 100)
$151 = 160 (Big spike. Go with an 80 or a 120)

Or, perhaps, AMD CPUs?
$50 1500 Mhz
$49 1600 Mhz // here, availability is the issue
$50 1700 Mhz
$58 1800 Mhz
$63 1900 Mhz
$71 2000 Mhz
$83 2100 Mhz
$95 2200 Mhz
$122 2400 Mhz
$170 2500 Mhz (Big spike, get a 2200 or 2400)

This is true in almost every part of the computer industry. At this point, a 2400 Mhz AMD is considered "commodity" while the 2500 is not. Same with 160 GB HDs vs. 120s.

With this, I'm almost happy with my equipment, and still have money to spend on my 5 (yes, FIVE) children.

-Ben

Re:Oh wonderful...I just upgraded

[WasterDave]

still have money to spend on my 5 (yes, FIVE) children

Screw the money, what about the time? I have 1 (yes, ONE) child - a toddler who's 17 months. At this stage my partner and I are _still_ constantly amazed that the species manages to perpetuate at all. Let alone five.

Jeez :)

Dave

Re:Oh wonderful...I just upgraded

[cebarro]

I was cutting edge for about 14 minutes once. But that was a very long time ago......

Re:Oh wonderful...I just upgraded

[You're All Wrong]

Don't worry - you did the right thing.

There are only 2 times to buy a PC - now, and never.

YAW.

Re:Oh wonderful...I just upgraded

[UnknownSoldier]

> Once upon a time mcrbids doth spoketh ...
> The last item before that spike is the one I buy.
> Notice: hard drives. (pricewatch.com =)
> $49 = 20 GB
> $58 = 30 GB
> $59 = 40 GB (who'd get a 30?)
> $67 = 60 GB
> $77 = 80 GB
> $100 = 100 GB (small spike = 80'd be ok)
> $106 = 120 GB (don't bother with the 100)
> $151 = 160 (Big spike. Go with an 80 or a 120)

Hey thx for the table! You're right about the 80 GB and the 120 GB as being the best buy -- $0.96/Gb, and $0.88/Gb respectively.

> Or, perhaps, AMD CPUs?
> $50 1500 Mhz
> $49 1600 Mhz // here, availability is the issue
> $50 1700 Mhz
> $58 1800 Mhz
> $63 1900 Mhz
> $71 2000 Mhz
> $83 2100 Mhz
> $95 2200 Mhz
> $122 2400 Mhz
> $170 2500 Mhz (Big spike, get a 2200 or 2400)

No, not a 2200 or 2400 !! They suck for overclocking, since the multiplier is too high. The 1700 and the 2100 are the best* CPUs for overclocking currently. Basically you can easily reach the same MHZ as the model number!

Check the scores at: http://www.cpudatabase.com/

*Best = Fantastic performance at a great price if o/c'd. Since a 1700+ can hit over 2 GHz (Air), and 2100+ around 2.2 GHz (Air) !

Cheers

--
The more corrupt the republic, the more numerous the laws.
- Tacitus, 55-120 AD

32 compatibility mode vs. true 64 bit apps...

[mekkab]

Maybe I'm trying to open a can of worms similar to VI vs EMACS, but so be it.

More addressable memory: cool.
8 way processors: cool*. (insert "do you have parallel apps to take advantage?" disclaimer. And some apps do.)

But if all my stuff is running in some 32 bit compatibility mode, then what real gain am I going to realize? The article did mention something about faster registers- does that mean small fixed point instructions run faster? (like increment register i, or add n to x, etc.)

And how many apps for 64 bit exist in the market?
(I'm sure there are some grad students and some code breakers with some custom software, but that's not my aim...)

I guess what I'm asking is- aside from custom code, what are the reason for me to early adopt a 64bit architecture? Or should I just save my money for when things get cheaper?

Re:32 compatibility mode vs. true 64 bit apps...

[vlad_petric]

The width of the machine is not the number of threads it's able to run at a given time (SMT like P4), but the number of instructions you can execute in a cycle (in a parallel fashion, of course). The difference between x86 and Itanium is that on Itanium the instructions are "explicitly parallel", i.e. the compiler bundles instructions together and the processor knows it can execute them in parallel, while on x86 the processor is smarter and determines itself which
micro-ops (actually it's rops for AMD) it can execute in parallel.

You're very likely going to see speedups on 32 bit code, simply because Opteron is twice as wide as P4 (and this has nothing to do with the size of the operands, which probably is not going to matter as much).

As far as the 64 apps are concerned - Linux already runs on it.

Re:32 compatibility mode vs. true 64 bit apps...

[TrekkieGod]

Well, this is how I see it.

Unlike the Itanium, AMD's 64-bit procs are not emulating 32bit compatibility...so there shouldn't be any decrease in performance there, it's supporting those natively. That doesn't mean decrease in performance with current 32bit chips, it means extremely large decrease in performance between 64bit apps and 32bit apps, such as what happens with the Itanium. Comparison with current chips depends on how good the new architecture is, and whatnot.

64-bit specific code is just a matter of recompiling existing code with a compiler that is aware of the newer architecture features, I've done that for my work at the university I study. Thus, I think Linux early adopters are going to see many more advantages than Windows early adopters because of all the open source software we can just recompile.

Then again, in a purely monetary view, it's always best to wait instead of buying things as soon as they come out. It's always way too expensive in the beginning.

Re:32 compatibility mode vs. true 64 bit apps...

[Anonymous Coward]

You're slightly missing the point ...

While true 64 bit apps are nice, and 8 way smp is cool, some of us are looking forward to 64bit because it increases the memory size per process.

Right now, very large applications (databases, etc) are limited to 4 GB - (kernel space) - (all other running processes). Assuming everything else is small, there's still a cap around 3GB for most rational operating systems. Or, if you have a system doing a LOT of networking, you need a LOT of kernel memory to keep track of the network clusters, and once again, the 3GB limit hits the kernel, and everything stops.

The 64bit chips raise the 4GB limit to something ridiculous, which gets rid of the 3GB limit for all of the others.

Re:32 compatibility mode vs. true 64 bit apps...

[PhoenixK7]

64-bit specific code is just a matter of recompiling existing code with a compiler that is aware of the newer architecture features

not exactly... i'd hazzard a guess that most code is not 64-bit clean and would required some cleanup/modification before compiling as a 64-bit app.

Re:32 compatibility mode vs. true 64 bit apps...

[Tailhook]

Unlike the Itanium, AMD's 64-bit procs are not emulating 32bit compatibility...

All AMD CPUs are "emulating" the x86 ISA. Athlons decompose x86 instructions and execute them on an essentially RISC core, no? AMD CPUs happen to do it very well, routinely matching higher frequency Intel parts.

Re:32 compatibility mode vs. true 64 bit apps...

[updog]

And how many apps for 64 bit exist in the market?

If you use a source-based distribution, like Gentoo, everything in your system will be compiled for your 64-bit architecture when it's installed. You'll be able to take advantage of your new 64-bit architecture right from the get-go.

Re:32 compatibility mode vs. true 64 bit apps...

[plugger]

I think that's also true of Intel 686 or later.

Re:32 compatibility mode vs. true 64 bit apps...

[Anonymous Coward]

Actually, so does Intel, at this point. They have a hardwired instruction decode for the few operations that are used enough, or are 'important' enough to do so; then the rest of the instructions (which aren't hardwired in) are taken care of in microcode. Microcode essentially does conversion of the 'CISC' instruction to a series of 'RISC' instructions; although when it comes down to it, it's really quite rare to find a pure RISC design anymore; People quickly point to the PowerPC chip as a RISC chip, but I think it's a misnomer in that case; RISC means Reduced Instruction Set Computer. The primary goal of RISC was to reduce transistor counts because a purely hardwired instruction decode (As was the case in the chips of the time) took up nearly all of the die area and transistors for the system.

So RISC was about putting more of the transistors to work more of the time. (By eliminating redundancies in the instruction decode). They did this in several ways; moving to a purely load/store architecture eliminated the complexity of many of the instructions. There's no need to have a register->register 'move' instruction at all; simply OR the contents of the register with zero, and store the result in the destination register. Assembly programmers find RISC even more tedious, because they have to jump through even more hoops (and using quite a few more instructions/lines of code) to 'get the job done.'

So if a RISC chip is about reducing the number of instructions (compared to a CISC chip), then the PowerPC hardly qualifies. It may have been intended to be one at the beginning, but as it is now, PowerPC has more instructions (without Altivec, which adds even more) than exist in a Pentium 4 or AthlonXP (Even if you include MMX/SSE/SSE2/3dNow!).

Re:32 compatibility mode vs. true 64 bit apps...

[sl3xd]

64-bit specific code is just a matter of recompiling existing code with a compiler that is aware of the newer architecture features,

Of course, this is the great equalizer: Sure, you can generate 64-bit code; but the new 64-bit architecture is new enough that for the time being, it's pretty unlikely that any of the current compilers will be able to optimize 64-bit AMD code as well as they can optimize 32-bit x86. So, sure, you can be assured you're running in full 64-bit mode... but unless you need to address that much memory space, the app would run faster compiled in 32-bit mode. This is espescially true in light of other comments about the unlikelihood that everything in a generic Linux distro is '64-bit clean'.

Re:32 compatibility mode vs. true 64 bit apps...

[Anonymous Coward]

Unfortunately, if you use a distribution like Gentoo, you are also a acne-ridden, pear-shaped, no-life loser nerd with no sexual future.

Re:32 compatibility mode vs. true 64 bit apps...

[MonoSynth]

And how many apps for 64 bit exist in the market?

And how many 32-bit OSses and apps were there back in 1985, when the 80386 was released? At the time it was released, it was treated by most users as an even faster 8086. It took ten years before a semi-32bit OS was accepted mainstream, and on top of that another 7 years before every sold PC had a full 32-bits OS. The success of the 386 was in its backward compatibility, and so will the success of x86-64 and the failure of the Itanium as mainstream-cpu be.

Re:32 compatibility mode vs. true 64 bit apps...

[sl3xd]

There are previous posts repeating this, but there's one problem with the theory: That taking advantage of the 64-bit architecture from the get-go will actually give a speed benefit. It also assumes that all of the source code is '64-bit clean', although that is likely not going to be much of a problem.

The problem is that, sure, everything is compiled into 64-bit mode. Fine. But can the compiler optimize the 64-bit code as well as it can optimize the older 32-bit code? Will the compiler make good use of the extra registers? I'm willing to bet that, for the first while, 64-bit AMD compilers will generate slower code than their 32-bit counterpart. (For 99% of all applications -- those which do NOT need more than 2 GB of memory).

And, of course, there is my major question: What kind of context switch and/or process latency can we expect from the Athlon64 & Opteron? I realize nearly all hard real-time apps don't need anything this powerful-- most good engineers will just put a microcontroller in to handle the hard real-time, and buffer things enough so that it doesn't matter that the workstation isn't hard real-time; but it does have a serious impact on other aspects of system responsiveness, as well as overall system performance for a microkernel architecture (such as HURD, Darwin, or QNX).

For that matter-- how will the Opteron's context switch time compare to other 'server' processors, like the UltraSparc, POWER4, Itanium, and, for good measure, PowerPC? Most of the arguments I've seen about modern x86 having a horrible context switch time don't seem to hold up to benchmarks I've seen-- where identically-clocked PowerPC and Pentiums take nearly the same time (and hence nearly the same number of clock cycles) to context switch...

Re:32 compatibility mode vs. true 64 bit apps...

[Nothinman]

Depends.

Linux has been running on 64-bit architectures for some time now, so 64-bit'ness isn't a new thing. And with distros like Debian that support everything from m68k up to Alpha a large portion of the issues have been taken care of already. Debian 'sarge' currently has 10058 packages, all of which compile and run (I'm not going to say all are 64-bit bug free, that would be stupid).

And also most Open Source apps are used on Solaris, Tru64, HP-UX, etc which are 64-bit. Windows may have a big transition ahead of it, but for the rest of us it'll be just like any other motherboard upgrade =)

Re:32 compatibility mode vs. true 64 bit apps...

[Garion911]

I wonder what kind of library compatiblity problems something like that would cause... Image I'm running a Gentoo x64 64-bit compiled system.. SO what happens when I decide to install some product that I don't have source to?? Would I have to have a 32bit glibc hanging around? Or would it be able to work?

Re:32 compatibility mode vs. true 64 bit apps...

[Nothinman]

And how many apps for 64 bit exist in the market?

A lot more than you think. Most Open Source apps compile and run on 64-bit OSes like Solaris, HP-UX, Tru64, Linux (when run on Alpha, PPC64, etc), etc.

There will be work that needs to be done to clean up some apps, there's no denying that. But a large chunk of the work is already done on the unix side. It's the Windows side that has a lot of catching up to do.

I guess what I'm asking is- aside from custom code, what are the reason for me to early adopt a 64bit architecture? Or should I just save my money for when things get cheaper?

If you can't think of a reason to get one, you probably won't need it any time soon. Most people won't get any benefit from a 64-bit (well, 40-bit it would seem at first release) address space.

Re:32 compatibility mode vs. true 64 bit apps...

[Vasilis Vasaitis]

Blockquoth the poster:

The problem is that, sure, everything is compiled into 64-bit mode. Fine. But can the compiler optimize the 64-bit code as well as it can optimize the older 32-bit code? Will the compiler make good use of the extra registers? I'm willing to bet that, for the first while, 64-bit AMD compilers will generate slower code than their 32-bit counterpart. (For 99% of all applications -- those which do NOT need more than 2 GB of memory).

Your caution in this is reasonable, but don't forget that the GNU community is using GCC, which supports a gazillion of architectures, so:

• It already supports 64-bit architectures.
• It already supports architectures with a lot of registers available. x86-64 has still got few registers, compared to some.
• It has already been targeted to x86-64.

So I would expect that the code GCC produces for x86-64 will be comparable to that produced for i386, and probably faster from the beginning, because of the extra registers. The extra size might affect some cache critical programs, but with the default integer size still being 32 bit, I wouldn't worry too much about that.

Re:32 compatibility mode vs. true 64 bit apps...

[calica]

You will probably need a 32bit glibc. 32bit code runs in a different processor mode (think the V86 mode used in dosemu)

Re:32 compatibility mode vs. true 64 bit apps...

[Neon Spiral Injector]

I compiled everything from source for my Alpha, not one problem with 64-bit cleanliness. That was a problem in the early 90s. There are enough 64-bit chips around now, that every serious application has been compiled on one at some time.

GCC is already ready for the Hammer chips. If compiling for x86-64 you get code generated that can make use of mmx, mmxext, 3dnow, 3dnowext, sse, and sse2. It actually prefers those instructions over the old x87 functions by default.

I've heard that the Hammer is 2.5 times faster per clock than the lastest Athlons in 32-bit mode, and faster still in 64.

I WILL be building a machine around 2 of these chips as soon as the first Tyan board ships with PCI-X slots.

Re:32 compatibility mode vs. true 64 bit apps...

[captaineo]

...and take a speed hit as all the mundane programs like 'cat' and 'ls' now have double the memory footprint...

64-bit code should be applied selectively, only to the prorgams that actually need the extra address space.

Re:32 compatibility mode vs. true 64 bit apps...

[updog]

Compiling applications 32-bit vs 64-bit is not really an issue of address space - the OS deals with that for you. What will make a difference, however, is how the compiler can take advantage of of the wider registers and internal data paths. I would bet that gcc does good job of optimizing code compiled for 64-bit architectures.

Oh, and a 64-bit architecture doesn't necessarily imply larger instructions - there are some architecures (like ARM) that are 32-bit, but you can choose between either 16-bit instructions or 32-bit instructions.

Re:32 compatibility mode vs. true 64 bit apps...

[Luyseyal]

Oh man, I'm so excited about:

gimp2.0_x86-64 provides gimp
gimp2.0_x86-32 provides gimp

yes, the little things in life...
-l

Re:32 compatibility mode vs. true 64 bit apps...

[Anonymous Coward]

gcc is very portable and it's nice to stick up for it since it's our favorite son, but it has never been a very good optimizing compiler. The x86 code generation has certainly improved a lot over the last few years but it still is soundly beaten by, i.e., the Intel compiler. Other architectures which have had less attention are even worse, for example on PPC and Alpha gcc is blown away by the Metrowerks or Compaq compilers.

Re:32 compatibility mode vs. true 64 bit apps...

[be-fan]

It also assumes that all of the source code is '64-bit clean', although that is likely not going to be much of a problem.
>>>>>>>>>
Most Linux code is 64-bit clean. The initial Alpha port of Gentoo was done by one guy in *two days* It took more than half a decade for Microsoft to go from 16-bit to 32-bit.

Re:32 compatibility mode vs. true 64 bit apps...

[be-fan]

They really shouldn't have double the memory foot-print. Code size increases only slightly (from an average of 3.4 bytes per instruction to 3.8 bytes per instruction). Integer size remains the same. Only long integers and pointers double in size. You'd probably get more of a noticible growth by compiling with GCC with inlining enabled.

Re:32 compatibility mode vs. true 64 bit apps...

[sl3xd]

I seem to recall it taking far more than half a decade for Microsoft to make the switch from 16 to 32; it seemed more like a decade. (espescially true considering some fairly major parts of Windows 95, 98, & ME were still 16-bit. (Not just real-mode things either... remember that protected-mode != 32 bit; OS/2 1.x was 16-bit protected mode.)

I recall that back in the day of the Pentium Pro, it was a much-ballyhooed 'flaw' about the PPro: It couldn't run 16-bit code very well at all. 32-bit code, great! 16-bit, slow. And, because significant portions (of the running code) of Windows 9x was in 16-bit mode, then to really get any kind of performance out of a PPro, you had to use Windows NT, which was... much more 32-bit than Windows 9x.

And we're only now getting rid of Windows 9x, and its 16-bit leagacy...

My grandkids are probably going to have to suffer something with 'old' 32-bit code from Microsoft in 30-40 years, rather than a 'modern' who-knows how many bits... The 64-bit address space is so obscenely big, it should be plenty even then... but, then again, so was 4 GB when the '386 came out...

Re:32 compatibility mode vs. true 64 bit apps...

[nusuth]

The 64-bit address space is so obscenely big, it should be plenty even then... but, then again, so was 4 GB when the '386 came out...

When 386 came out we already had exhausted 20bit adress space we had and trying funny hacks with accessing the memory thorugh the keyboard controller chip (not really but A20 thing was close enough.) And 32 bits merely 4096ed it. The transition from 32 bits to 64 bits will increase the addressable space 4 billion times, which is a million times higher than four thousand of 386. That should make this transition quite a bit more durable.

Re:32 compatibility mode vs. true 64 bit apps...

[innovate64]

The stuff is coming. Actually, its already here. Valve Software has announced that a 64-bit version of its Counter-Strike server software and it runs 30% faster than the 32-bit version on AMD's upcoming Opteron processor. The server is already on the market (prior to the April 22 Opteron launch). http://www.tech-report.com/onearticle.x/4851

please just say 40h bit--no decimal .

[Anonymous Coward]

Isn't 40h more beautiful than 64? Decimal is ugly in this case.

Re:please just say 40h bit--no decimal .

[PhoenixK7]

or 01000000

binary is soooo much cooler than hexadecimal!

Re:please just say 40h bit--no decimal .

[rsheridan6]

Screw that, let's be extra obscure and do it in balanced trinary

1(-1)101

Re:please just say 40h bit--no decimal .

[Anonymous Coward]

There's a big mistake in the article.

A binary digit is a bit. A trinary digit is not a trit, it is a tit.

Slashdot a Little Slow?

[Grip3n]

The release date of April 22 was released a long time ago

(ie: January 21, 2003, just incase you didn't get the picture)

Re:Slashdot a Little Slow?

[Anonymous Coward]

The release date of April 22 was released a long time ago.

Wow, you just realized slashdot isn't about the fastest cutting-edge news! No one has ever realized that before! You get a cookie!

Re:Slashdot a Little Slow?

[brejc8]

The new story about this should be AMD's numbering stratergy

Palladium?

[rmohr02]

The big question is whether or not Palladium will be built into this chip. Anyone know?

Re:Palladium?

[Anonymous Coward]

I'm betting it is in there for testing but it won't be official for a few more releases. I can't wait until it is released, though. Quality digital media available cheap is going to give the industry the shot in the arm that it needs.

Re:Palladium?

[Anonymous Coward]

WORD TO YO MOTHER DAWG!!! Functional, trustworthy DRM combined with micropayments will create a revolution on the scale of Steve Gutenberg's printing press! Dow Jones will skyrocket to 20,000 - BUY MICROSOFT NOW! Bling bling!

Opteron and Athlon 64 are not the same CPU

[ruiner5000]

The Opteron is being launched on the 22nd of April. It was code named Sledgehammer, and is what Newisys and others are using. It is the server version of Hammer.

Athlon 64 is coming out in September. It is the desktop and mobile version of Hammer that was codenamed Clawhammer.

Re:Opteron and Athlon 64 are not the same CPU

[Anonymous Coward]

thanks! I was trying to figure the difference between Opteron and Athlon 64, but neither Google nor AMD's web site were any help..

Re:Opteron and Athlon 64 are not the same CPU

[calica]

But what are the actual technical differences? The Duron and Athlon had different size caches. Is that how they're gonna differentiat them? How about the CPU interconnects?

Re:Opteron and Athlon 64 are not the same CPU

[aka1nas]

The Opteron is more scalable(4+way), while the Athlon64 sounds like it will be 2way SMP at best. The Opterons also have more cache.

Re:Opteron and Athlon 64 are not the same CPU

[Luminous Coward]

Straight from the FAQ.

Q: What are the differences between the AMD Athlon 64 and AMD Opteron processors?

A: The upcoming AMD Opteron and AMD Athlon 64 processors are designed for different markets. For the server/workstation market, the AMD Opteron processor will undergo more stringent validation and reliability testing. Another difference will be in the number of HyperTransport links embedded on the chip. The AMD Athlon 64 processor will contain one HyperTransport link offering 6.4 GB/s data transfer while the AMD Opteron processor will offer three links. The processors will also contain different amounts of cache.

ClawHammer = Athlon 64
SledgeHammer = Opteron

nekkid Opteron boards!

[gearheadsmp]

here
and Athlon64 boards here

Re:nekkid Opteron boards!

[florin]

Unfortunately, the Opteron boards in the pictures on amdboard all have one thing in common; the lack of an AGP interface. If this this story is correct, that's not an inherent limitation of the 8200 chipset. The Gigabyte board for instance is supposed to come with an AGP8x slot.

While I realise that AGP is not a requirement for AMD's envisioned Opteron target market, I can't imagine I'm the only one who is toying with the idea of getting an Opteron setup as a personal workstation, what with the earlier availability as well as the larger cache (and the resulting higher performance compared to the consumer CPU, presumably).

Anybody have any clues if there will be Opteron boards with AGP at launch?

Re:nekkid Opteron boards!

[Neon Spiral Injector]

Some of those boards have to have PCI-X slots. They can be faster than AGP. Just have to find a PCI-X video card.

(If anyone knows of a maker, point me to them, I think PCI-X is the future, AGP was just a gap bridging hack, who's time has passed.)

Re:nekkid Opteron boards!

[Anonymous Coward]

I don't think those are PCI-X slots, just PCI 64-bit. Those have been on server boards for years..

No videocards for that available methinks

Re:nekkid Opteron boards!

[florin]

If the Inquirer link is correct then there will indeed be boards that have PCI-X, and I agree it will certainly be a big thing in the future.

But AGP is likely to be around for a while still (all current and up and coming chipsets support it), and I already own a great dual DVI card which I'd like to use with the new board.

Re:nekkid Opteron boards!

[Neon Spiral Injector]

There are 66 MHz 64-bit PCI video cards, but the cost for them is so high, as there is no demand for them. I'm hoping that changes with PCI-X, I'm aready seeing more boards with that interface than 64-bit PCI.

But yes, my Alpha which is quite old has 64-bit PCI. And I have two Pentium 2 boards with ServerWorks chipsets that have 64-bit PCI. But that is what I'm saying, for as long as the spec has been around I don't see that many boards, just about every server class board ready to be released has PCI-X now.

Re:nekkid Opteron boards!

[hovik]

Those are regular PCI 32 bit. Thats only used for desktop computers, so it seems a bit like the Athlon64 is meant for desktop use from day one.

PCI 64/66 MHz is a must in server systems to benefit from Gbit NIC.

Re:Check out the linux bechmarks with optron!

[Sandman1971]

I don't see why Sun should worry too much. Sun offers something that AMD doesn't: Service contracts. At my shop, all our servers, monitors, etc... are under contract. We have a problem with a CPU, or memory, or a mobo or anything else? A Sun tech is at our door within 4 hours to replace it. That's something you won't get moving to a Linux/AMD solution.

Re:Check out the linux bechmarks with optron!

[gearheadsmp]

Or Sun could just use Opteron chips in their servers, like they do Athlon XP mobiles for their blade servers.

Why get a VIA 400?

[Anonymous Coward]

Why not get nForce2?

Re:Why get a VIA 400?

[listen]

nVidia are being pretty cagey about some hardware specs, and are releasing binary drivers for linux... Its a shame, seems like a good chipset.
But I would go for a VIA atm unless you are entirely MS bound.

Re:Can you imagine...

[WindBourne]

I suspect that AMD is hoping that many will, in fact, imagine just that. They may be a bit pricey, but beowulf will benefit from them.

1 TB of RAM??

[Alpha_Nerd]

Looks like a nice way to prevent sites from being /.ed

Re:1 TB of RAM??

[You're All Wrong]

Yes, it instantly increases your network bandwidth, obviously.

Remember - even sites with a single text page which would be in cache (memory) get stashdotted, so it's not the hard disks that are the problem.

Or were you trying to whore for a +1 funny?

YAW.

Re:1 TB of RAM??

[Alpha_Nerd]

Anytime a connection is made to a webserver it uses RAM , and that is one of the major causes of /.ing.

If pages are only text, most of the time bandwidth would be perfectly adequete.

Clawhammer earlier than Sept?

[Tumbleweed]

Other rumours (see the Inquirer) suggest that it may be _later_ than Sept - _maybe_ arriving in time for the xmas season, though maybe not.

When betting on Clawhammer release dates, you'd be best advised to always count on the later date, rather than the earlier.

I'm actually more excited about PPC 970, thus finally allowing Apple to have a decent amount of hardware horsepower.

Re:Clawhammer earlier than Sept?

[imbezol]

These guys (gg webdesigner) seem to think June 3rd.

Re:Clawhammer earlier than Sept?

[Tumbleweed]

A fan site can dream on about whatever they want, but I stand by my advice - don't count on any release date of a Hammer processor being _early_. Hammer CPU release dates seem to follow the MS OS release date paradigm. :(

Re:Clawhammer earlier than Sept?

[Seq]

I'm actually more excited about PPC 970, thus finally allowing Apple to have a decent amount of hardware horsepower.

I too, am waiting to see how this chip compares. I doubt I would buy apple hardware still (although I would love a powerbook and OSX), but the idea of possibly building your own low power (comaratively) 64-bit PPC machine is tempting, especially when I try to imagine the power consumption from an Athlon64..

Opteron vs. Athlon64?

[sconeu]

What's the difference between an Opteron and an Athlon64?
</QUESTION>

Re:Opteron vs. Athlon64?

[Anonymous Coward]

Where's the DTD so I can parse this correctly?

Re:Opteron vs. Athlon64?

[Anonymous Coward]

Opteron is for servers and it supports 8-way SMP. Opteron also has a dual-channel memory controller.

Athlon64 is a desktop version and has a single-channel memory controller.

Re:Opteron vs. Athlon64?

[gmack]

Simple... The Opteron is designed for server use and has the massive cache etc while the athlon64 will be priced for home use.

Re:Opteron vs. Athlon64?

[gordyf]

Opteron is the Stronghammer, and Athlon64 is the Clawhammer.

I don't think the Athlon 64 will support SMP at all (similar to Athlon XP), and will probably have smaller caches than the Opteron. Other than that, I believe they're functionally equivilant.

Re:Opteron vs. Athlon64?

[TheRaven64]

Opteron will cost more :)

Re:Opteron vs. Athlon64?

[You're All Wrong]

One will have its name chaged by marketeers within the next 3 months, the other won't.

Never try to get inside the minds of marketeers.

Given The Register's ( http://theregister.co.uk/ ) habit of mutating processor names (Celery, Itanic etc.), which annoys the maufacturer(s) greatly (as it dilutes their trademark) I think that AMD we unwise to call one 'Opteron' - I think that it might be called the 'flopteron' if it doesn't catch on quickly.

You heard it here first.

YAW.

Re:Opteron vs. Athlon64?

[Hoser McMoose]

That's actually a very good question, since the rumors as to the difference seem to change about every second day!

Originally there were two chips, the Sledgehammer and the Clawhammer. The Sledge had 1MB of L2 cache, 3 Hypertransport links and a 128-bit wide memory controller. The Clawhammer had 256KB of L2 cache, 1 Hypertransport link and a 64-bit wide memory controller.

At first it seemed like the Sledgehammer was just going to get the official name of "Opteron" while the Clawhammer would get the official name of "Athlon64". However, it no longer seems that simple, since there are THREE different versions of the Opteron (the 100, 200 and 800 series models). What's more, it's looking like AMD might shy away from using only 256K of cache on their desktop chip since Intel will just be releasing new desktop chips with 1MB of cache at the same time.

So, where does that leave us? Well, I dunno for sure, but this is my guess:

Opteron 100 series: 1MB of cache, 1 Hypertransport link, 128-bit wide memory bus, only for single processor systems
Opteron 200 series: 1MB of cache, 2 Hypertransport links, 128-bit wide memory bus, for dual-processor systems
Opteron 800 series: 1MB of cache, 3 Hypertransport links, 128-bit wide memory bus, for up to 8-way SMP systems

Athlon64: 1MB of cache, 1 Hypertransport link, 64-bit wide memory bus, only for single processor systems
"Athlon64-Lite" (Duron64?): 256KB of cache, 1 HT link, 64-bit memory bus, single-processor only.

Of course, it could be a bit trickier than that. Also, I've seen documents from AMD where their "single" 16-bit Hypertransport link could be split into two 8-bit HT links and allow for dual-processor operation. So, perhaps the Opteron 100 and Opteron 200 series will be identical except that the latter will be certified for use in dual-processor systems? If this is the case, than the Athlon64 would also (at least theoretically) work in dual-processor systems.

It's also quite possible that AMD will have one or more processor with 512K of cache, rather than the 256K or 1MB originally planned.

Long story short, until AMD announced the official specs, no one really knows for sure.

chicken before the egg?

[Rcknight]

And how many apps for 64 bit exist in the market?

umm, do you really expect alot of 64 bit apps, when the processors arent avaliable yet?

Re:chicken before the egg?

[Neon Spiral Injector]

Well I was messing with Bochs emulating (90% of) the x86-64. I was cross compiling and then copying the executables to the drive image. But a 2.6 MHz Opteron just isn't too much fun.

So anyway, I have a 2.4.20 Linux kernel, glibc, gcc, binutils, make, bison, flex, ncurses, perl, texinfo, man, bash, fileutils, textutils, findutils, inetutils, grep, groff, grub, file, less, vim, util-linux, e2fsprogs, procps, sysvinti, sysklogd, shadow suite, sh-utils, sed tar, gzip, bzip2, openssl, openssh, and who a few other things ready to go for when the chips ship.

Not everything runs and I had to watch my level of optimization (had to use the x87 math instead of the mmx/3dnow/sse stuff that gcc wants to use by default). I will actually be compiling on the machine for the machine when I get one.

Re:Check out the linux bechmarks with optron!

[Anonymous Coward]

Yeah, unless you go with a Linux/AMD vendor.

Duh.

Congrats to AMD

[Anonymous Coward]

Nice to see them going out and competing in a market that Intel has seriously flubbed. I only buy AMD boxes exclusively now as they are as fast as I need them to be and are cheaper.

Re:Check out the linux bechmarks with optron!

[iggymanz]

I've worked with x86 Linux servers in datacenters from Compaq and Dell with 24x7 4 hour contracts. I'm sure major vendors of AMD based servers will have them too. Sun is VERY worried. Sun is doomed. I give them 4 years tops.

Sun SHOULD worry

[Anonymous Coward]

Since firms running huge database applications will leave their overpriced/underpowered Sparc chip in droves. Please tell me how Sun justifies their 10X price differential over x86? I just don't see it. Neither do most corporate computer buyers.

They built it, now who will come?

[glrotate]

Vendorwise I mean. Dell won't touch AMD flakiness with a 10 foot pole, I don't recall seeing many IBM AMD servers, and HP has too much tied up in IA64. So you're going to build your datacenter around Gateway? MOO.

Re:They built it, now who will come?

[Pharmboy]

Vendorwise I mean. Dell won't touch AMD flakiness with a 10 foot pole

I don't know, they have used AMD off and on, and any look at TV shows that Dell is going for the server market with a vengence. I could possibly see only using AMD for servers and not the regular athlons. I own several dell servers, they are aggressive as hell on price and marketing, and frankly, make a good entry to entry/midlevel server, IMHO.

So you're going to build your datacenter around Gateway? MOO.

Yea, thats not gonna happen here. I have had the worst luck with their products, regardless of cpu. Then again, there are lots of lesser known makers that don't have the mindshare of the big guys, but could snag some of the marketshare.

Re:They built it, now who will come?

[leviramsey]

However, Michael Dell has historically been Andy Grove's bitch...

I dount Dell will ever use an AMD chip, even if it means losing sales.

No Matter How Much Ram You Have

[The Analog Kid]

Windows is still going to use the swap file.

Re:No Matter How Much Ram You Have

[TheRaven64]

No Matter How Much Ram You Have
Windows is still going to use the swap file.

Why was this moderated as funny? It's certainly true for the NT series (including 2K and XP). Their VM strategy allocates all the ram (except the part reserved for the kernel) as disk cache, and all allocated memory is swap, cached by the main memory. This allows the kernel to dynamically tune the amount of disk cache used according to how much is required. It sounds insane, but is actually quite an elegant solution (in theory at least, I'm not convinced it works in practice. A lot of disk I/O throughput will kill system performance as all your apps get swapped out in favour of disk cache).

Re:No Matter How Much Ram You Have

[/dev/trash]

Tell Windows not use a swap file.

Re:No Matter How Much Ram You Have

[OneArmedMan]

conservativeswapfile=1 it wont always use the swap file ...

Re:No Matter How Much Ram You Have

[Anonymous Coward]

A lot of disk I/O throughput will kill system performance as all your apps get swapped out in favour of disk cache

Whis was an annoying problem for me in Win2K. In XP you can disable disk swapping by setting the swapfile size to 0MB. The cache subsystem sees swapfile is full and instead nukes some non-dirty page from the disk cache. Process memory is still converted to swap and sent to cache, but it never gets flushed to disk. Just need to have enough memory (512MB min) to do this .

Re:Check out the linux bechmarks with optron!

[Tailhook]

"That's something you won't get moving to a Linux/AMD solution."

Major vendors (IBM, HP, etc) offer this level of support (and better) for there Linux products right now. Sun does too. When these vendors sell x86-64 CPUs, they will offer the same level of support on day 1. This is understood.

Dual 64 boards

[BrookHarty]

I want to know if there will be dual boards out that are affordable. Currently the only options for a dual cpu environment is the P4 with hyper threading, Dual Xeon, or older AMD MP boards. All which are very affordable.

I've been using dual boards for awhile, then switched my workstations to a fast single CPU setup. What a mistake. With I/O taking so much CPU time, a dual setup is the best solution. I have not tried P4's hyper threading, but I wonder if thats the cheapest way to get the performance on a single CPU board.

AMD shouldn't abandon its MP CPU line. Wheres the MP chips with the 333mhz bus?

Re:Dual 64 boards

[gmack]

The nice thing about hypertransport is the fact that the boards are simple design wise and cheaper to manufacture. The current Athlon boards are expensive to make and difficult to get right and that lead to the lack of vendor support and the general lack of quality. It's good to see they have learned from past mistakes.

Also I'm not saying this is the case but if your having a general IO slowdown on your workstations that goes away on dual CPU systems then you may want to check and make sure DMA is enabled.

Re:Dual 64 boards

[BrookHarty]

Actually its more the way IDE spikes when your under heavy load. Having a dual cpu machine helps, scsi would be the ideal if the prices were no so high. Storagereview.com use to list how much CPU a HD would use under load. Sad to see that go.

Re:Dual 64 boards

[Wesley Felter]

I'm guessing that dual ClawHammer boards will be pretty affordable since the chipset and sockets are the same and HyperTransport is easier to route on the board than old-style wide shared buses.

Re:Dual 64 boards

[DancingSword]

Take a real good look at AMD's roadmaps:

Dual Cla-whammers are GONE.

AMD evidently decided to force the enthusiast/mini-server market to choose to buy-up ( dual sledge-hammers, and at the prices involved .. NBL ), or

buy-down ( Dual Athlons

Mean ( well, not really mean, but .. wahh!! . . ), but effective ( for their bottom-line ).

Mind you, there are 2 other significant concerns in replacing my system ( I'm in the segment they .. decided to ignore ):

1. Silent System: .. those Cla-whammer HSFs look huge/possibly-noisy, or, if the chips really are low-wattage, then they'll be really silent under a copper HS with a Verax.de fan on it, and

2. as someone else mentioned, HD CPU usage, but the solution for that ( waitaminit, we dissolve stuff to fix our PC's? ) is to use Serial-ATA ( non-blocking, and non-redundancy-of-commands ).
And with Linux, the Silicon Image chip based S-ATA controllers are supported in 2.6, so grab these, then, rather than the non-open-source HighPoint, or the outright opposed-to-open-source Promise.

Lost Circuits Benchmarks ( stunning ), and CyberCPU.net ( it's the low-CPU, 8% vs 44%, that puts S-ATA into the phenomenosphere ), and .. I'd heard that Seagate is implementing out-of-order-execution for its upcoming S-ATA drives, which oughta make 'em punchier..

( for the TLA-challenged, the CLA in Cla-whammer, the new AMD desktop chip, stands for the Canadian Luge Association, and if these chips are able to flatten luge , they're damn capable, and..
the above usage of NBL stands for Not Bloody Likely, as rememberers of the film-version of Pygmalion may remember.. My Bloody Fair Lady, I think it were callethed.. hmmm.. )

Re:Dual 64 boards

[BrookHarty]

I know Opterons will be upto 8x cpus, but i hope they keep the chipset cost down for home users wanting dual/quad boards. The MP chipset line is end of life via its roadmap.

Guess its just, hurry up and wait.

Years of optimizations are reaching their end

[godIsaDJ]

Since the advent of 32 bits processors there's been a huge amount of research trying to shrink the space taken by data-structures and similars. I've read of techniques devoted to save a couple of bits in Java classes for instance. Now we get addresses 64 bits long... I guess we can forget all of that... Any thoughts guys? Is it really worth trying to squeeze bits in object code if every 5/10 years processor technogy makes a leap?

Re:Years of optimizations are reaching their end

[John Courtland]

I would really say yes, that wasting cycles/space is a horrible practice. If you can market your program to a very wide audience and not have to redevelop it to meet space criteria on different architectures (especially embedded) then all the better and cheaper.

Another way to look at it is the less processor time and resources you dedicate to a process, the more processes can be run at a time, making the OS more responsive and the computer more productive.

Re:Years of optimizations are reaching their end

[Katana2]

These optimization techniques that result in smaller code will almost always result in faster code. Bottom line is the CPU takes time to process data, more data = more time.

Re:Years of optimizations are reaching their end

[rtaylor]

Heck yeah. Even if you have unlimited memory, you still have finite bandwidth between the memory and the CPU. Any *real* reduction will have a speed improvement on all systems.

Anyway, the real trick is to apply those techniques without actually doing any (much) work. Sometimes as simple as a define and some preprocessor work, othertimes you toss a perlscript into the toolchain that builds & compiles the code.

That, or if possible have the technique applied globally by the compilter itself. Send a patch to Sun ;)

Re:Years of optimizations are reaching their end

[Herkules]

"Is it really worth trying to squeeze bits in object code if every 5/10 years processor technogy makes a leap?"

Well yes small code/data makes more of the program can fit in the cach witch makes cach hit rate better!

Amd claims the code size will grow an avrage of 10% when moveing from 32 to 64 bit.

8 way interleave

[rf0]

For me its the 8 way SMP that excites me but I do have to wonder if they have any performance enhancements in? Do they have some sort of shared cache so they can talk cache->cache without have to do a lookup (TLB?) in main memory. Also what is the performace like, say compared to an Itanium 2 or Xeon? Anyone got any recent benchmarks?

Rus

Re:8 way interleave

[10Ghz]

The benefits should be substantial. Besides the normal benefit of havin 8 CPU's you also get massive memory-bandwidth. Each prosessor has a memory-bank of it's own with 128bit mem-bus to it. But they can also use the mem-banks of the other CPU's (altrough there is a latency-penalty when doing that. But the latency is pretty good when compared to other architectures). So you could say that 1 Opteron has 1 x 128bit mem-bus, 2 Opterons have 2 x 128bit mem-bus, 4 have 4 x 128bit bus and 8 have 8 x 128bit mem-bus.

Re:8 way interleave

[Neon Spiral Injector]

That is one place the OS can help too. I know a lot of work has gone into Linux's SMP to try balance favoring one CPU to keep in its cache and the processes running on it in sync, vs. keeping all the processors active.

But yes, part of the Hyper Transport is to allow better cache sharing between chips.

Re:8 way interleave

[Wesley Felter]

Do they have some sort of shared cache so they can talk cache->cache without have to do a lookup in main memory.

Yes, I believe the coherence protocol allows direct cache-to-cache transfers.

Performance should be very good, although we'll have to wait until April for the official benchmark result.

Re:8 way interleave

[Anonymous Coward]

Each CPU has its own cache with its own cache coherencey, but the new magic is with the 3 hyper-transport controllers. This means that with a 2-CPU machine, the two CPUs can access each other's cache with one hop, 4 CPUs can do it with two hops max, and 8 CPUs will have a 3 hop max.

The opteron has a mesh networks to connect all the CPUs so that near CPUs can be accessed in only one hop, while the farther can be reached with 3 hops. Each additional CPU adds 3 more bandwidths, while also requiring more for inter-CPU memory access. SGI can scale their design up to 512 CPUs, but AMD limited to 32 (each CPU must have its own address for the cache protocol). However AMD recommends motherboards to only implement 8 CPUs and to use clustering to get more CPUs. There must be some scaling problems with the current implementation.

Sun is also using this mesh in its dual one chip CPUs to get the memory access times down with multiple chips.

AMD could also make multi-core chips using both internal and external hyper-transports to get to 16 or 32 CPU machines. I doubt there is enough demand for this now, but hopefully cheap opterons may boot-strap demand by showing how much more productive Linux can be with 32-CPUs.

1TB

[vandel405]

This is not a troll, but don't any of you find the 1TB memory ceiling stupid? I mean, doesn't this remind you of 640kb limit? Why only a 40bit address space? Why not 64bit? I mean if you're gonna go 64bit, go 64bit! this isn't even half assed it 3/8's assed.

Re:1TB

[Tailhook]

No, the ceiling is not stupid.

Limiting the physical memory to 40 bits reduces the cost of building other system components, such as chipsets and motherboards, dramatically. Further, 1TB of RAM is sufficient for the current market. That is 1000 1GB parts, to give it some perspective.

As customers begin to approach 1TB requirements, AMD only has to implement more lines. No need for any segmentation hacks. The ISA needs no modification. This is a pragmatic and wise design decision.

64 bits of address space is still very useful without having actual RAM to back it up. It means you can map large quantities of storage into RAM directly. For example; if you have 10TB of disk, you can map all of it into a single virtual address space and address it with simple offsets. Obviously this is useful for large databases.

Re:1TB

[phfpht]

Cost.

These chips (for the first few years *at least*) will be low end or mid range small servers. Small in that they won't be competing with Sun Fire 15K servers which themselves only support physical memory of up to 576GB. No one will need those extra 24 memory address lines, so why build chips or motherboards with them?

Another way of looking at it is real-estate.

Currently 1GB dimms are the normal upper end, with a few 2GB dimms around.

At that rate, can you imagine the motherboard that could actually reach the 1TB RAM limit? 512 or 1024 dimm slots?....

Would that fit into the ATX form factor?

Re:1TB

[TheRaven64]

Actually, it makes a lot of sense. Digital did the same thing with the Alpha. The architecture supports 64-bit addresses, but the current implementation does not. Who actually needs 64-bit addresses now? I suspect that anyone who does is not going to be using x86 hardware. Software can be written now which can take advantage of 64-bit addresses, and it will not need to be recompiled for newer versions of the series.

AMD Should Rename Chip to "VAPTERON"

[Anonymous Coward]

An AMD release date is as strong as silly puddy

Why Release a Server and Desktop Version?

[Anonymous Coward]

Could someone enlighten me as to why they would release to versions of essentially the same chip? In fact wouldn't it be more cost effective to just have the 1 line of chips instead of splitting manuf and development costs?

Re:Why Release a Server and Desktop Version?

[TheRaven64]

In fact wouldn't it be more cost effective to just have the 1 line of chips instead of splitting manuf and development costs?

Manufacturing ICs is not an exact science. Very small impurities can render a die, or a part of a die unusable. The manufacturer can either throw these away, or disable the broken part and sell it as a crippled version. Intel's Celerons were just the P2/3s that had failed part of the cache tests. A lot of the difference between the Athlon64 and Opteron is cache size. A second is SMP support. The Opteron has 3 HyperTransport controllers on die that allow it to communicate with other CPUs with no 'glue' architecture. The Athlon64 will only have one, so any Opterons which have one defective HT controller can be sold as 2-way SMP parts, and those with 2 defective HT controllers will be Athlon64s (those with 23 defective HT controllers will be marketed as paperweights).

AMD has only developed a single CPU, and it will only manufacture a single CPU. It will market this as the most expensive Opteron. All the other versions are simply failed versions of this, with the broken parts disabled.

As an aside, you actually could enable the other half of the cache on some Celerons by mutilating the CPU a little. It was a particularly bad idea, since errors in cache tend to make your system rather unstable, but it could be done...

Re:Why Release a Server and Desktop Version?

[LiquidEric]

Even if the Athlon64 and Opterons were identical it would still be more profitable to sell one full-featured chip and one chip that has been "crippled". By offering two versions of the chip you can charge a lot for the fully-featured chip to business and hobbyist. At the same time budget conscience people will by the lesser version at a lower price. There is a term for this practice it is called "price discrimination". For example the cable company charges more for the deluxe package.

Re:Why Release a Server and Desktop Version?

[Wesley Felter]

AMD has only developed a single CPU, and it will only manufacture a single CPU. It will market this as the most expensive Opteron. All the other versions are simply failed versions of this, with the broken parts disabled.

I don't think this is true. I guess in September we can pop off the heat spreaders and compare die sizes.

*ahem*

[otterpop378]

uh oh uh oh here comes the hammer!

Re:*ahem*

[gilesjuk]

Wonder if they'll miss the obvious advert?

"Can't touch this....." :)

Re:*ahem*

[Anonymous Coward]

I know you're making a joke, but when was the last time you saw an AMD ad on TV or heard one on the radio?

AMD needs good products but they also need a marketing campaign. The phrase "so good it'll sell itself" is untrue no matter how good a product is.

Re:*ahem*

[gilesjuk]

Thing is, do Intel's adverts make any difference? the average user just buys a PC or laptop by looking at the price and the looks. Ask them what the difference between AMD and Intel is and they won't know. This is why Intel are so determined to stop AMD and Transmeta eating away at their laptop market. Comparing CPUs isn't like comparing cola.

Re:*ahem*

[otterpop378]

durring the first season of "who wants to be a millionaire". Although they may need to make a slightly larger profile marketing move, the fact is, they have one of the most successful word-of-mouth campaigns of any company. ever.

Re:*ahem*

[VAXman]

Intel is the fifth most valuable brand in the world. Intel has something like 85% of the desktop CPU market, even while it charges a substantial premium against the competition. That's the value of the brand, right there. If its brand was less valuable it wouldn't be able to charge a premium while maintaining market share.

Re:*ahem*

[gilesjuk]

I understand branding, however this isn't like McDonalds and Burger King. Many people walk into a PC store and buy what they think is a good price, they probably ask the sales guy if it can do XYZ. If there was an Intel PC store, or an AMD PC store I could understand the need for lots of advertising.

The fact that Lindows PCs have been selling well sort of proves that price makes a difference to some. The fact that the PC is running an almost alien OS doesn't put some people off.

Main Competitor = Itanium; Not Xeon

[Nazmun]

Xeon's are below $1000 while the opterons will be several thousand. They are going into super high end servers. Athlon 64's will be for us normal people.

Re:Main Competitor = Itanium; Not Xeon

[paitre]

And you're getting this pricing for the Opteron from where? Everything I've been hearing has the low end stuff -=undercutting=- the current Xeon pricing, while still running more than standard Athlons/P4's.

Geez.

Re:Main Competitor = Itanium; Not Xeon

[cbreaker]

I don't think this is true. The Opteron will be replacing AMD's current Multi-processor offering (Which is the Athlon MP) and the Athlon will replace the uni-processor offerings (Athlon XP.)

There will probably be some "High End" Opteron chips with large cache, but overall the chips will most likely be priced for market share.

I do agree with the part about the main competitor being Itanium, but not price-wise. Ultimately I believe that Itanium will offer better performance, but at the price of compatibility and cost.

I can't wait to have a 64-bit workstation!

Re:Main Competitor = Itanium; Not Xeon

[Loki_1929]

Opteron is AMD's first real entry into the server market. AthlonMP was never truly meant for anything other than workstations or very low level entry servers. Itanium is Intel's offering for very strong, high performance servers. Their Xeon chip is the high-end workstation to mid-range server chip. All indications are that AMD will be targetting low to mid range servers with the Opteron.

Intel's point is that they don't believe anything other than high-end servers will use 64-bit chips effectively. AMD's point is that anyone can use whatever they like. A dual Opteron with 2MB L2 will most likely be targetted against dual Xeon machines. AMD will try to offer a better price point and the ability to run 64-bit applications to potential customers in their attempt to win partners. The launch of the Opteron had best go off MUCH better than the launch of the AthlonMP if AMD hopes to make it in the server business. Already weary of new products, big businesses will be looking for any excuse not to go with an Opteron. Even Intel has trouble convincing long-time customers to invest in new products, as evidenced by their dismal Itanium 1 launch. If the chipset and board problems we saw in the Tyan AthlonMPs creep up in any of the Opteron boards, AMD's cash cow will be seen in the business community as little more than the 'roo meat at McD's. That, potentially, could end up being the nail in AMD's coffin.

They have a lot riding on this launch, so let's all hope it goes off without a hitch. If it does, I think Xeon processors will be collecting dust within a year's time while Opterons slowly replace what's currently in the workplace. In very, very few circumstances will a company look at Opterons as an alternative for Itaniums. In terms of performance? Who knows; we haven't see benchmarks on production Opterons yet. If it's everything we're told it is, it may very well outperform the Itanium 2s vis-a-vis.

Re:Main Competitor = Itanium; Not Xeon

[Nazmun]

Thats what i originally thought but the opteron is the multiprocessor that does more then dual.

The clawhammer (athlon 64) will do dual processors. Thats right, dual processing for the mainstream processor!

If athlon 64 processors go like xp's have been going (around $50-$60 bucks for the 1700) near the lower end then we might have very cheap duallies :)

Re:Main Competitor = Itanium; Not Xeon

[cbreaker]

Ohh, very cool. I was still under the impression that Athlon64's would be single-processor systems. In that case, I can't wait for a Dual Processor, 64-bit workstation!

As for price, they will have to be cheap to stay competitive with Intel's 32-bit offerings. The new 64-bit core of the Opteron/Athlon64 will replace the entire line of AMD's processors, so there will be cheap Athlon64's in time.

I just hope that the AMD chips aren't used solely as "Fast 32-bit" processors; I would hope that Microsoft and other software developers keep to the promise to have 64-bit versions of windows and other packages. Once again Microsoft will be developing Windows for multiple platforms; it used to be x84, MIPS, PPC, and Alpha. Now it will be x86, x86-64, Itanium.

Of course, Linux already has x86-64 support, but for mainstream acceptance we need everything to support it, even Windows. (ie- let's not turn this into a Linux vs Windows fight.)

Re:Main Competitor = Itanium; Not Xeon

[cbreaker]

The so-called "very low end server" market makes up 90% of all x86 servers sold. Most x86 based data centers, albiet Linux or Windows, use single or dual-processor servers in load-balanced clusters. Most of these servers run Pentium 3 processors.

These servers handle most of the workload, while a small few "high end x86" servers can handle database things. Sometimes, the database sits on non-x86 platforms.

The AthlonMP is a higher-performance alternative to the dual-Pentium 3 server/workstation, and can even perform well against a dual-xeon box.

It's hard to know where AMD is trying to position this processor, but it seems to me that the new processor core is positioned to compete with Intel on all fronts. Perhaps the processor will perform so well and be so much cheaper then the Itanium that it will bring very-high performance computing to places that could normally not afford it. We don't know; we'll have to wait and see.

Re:Main Competitor = Itanium; Not Xeon

[dabootsie]

The enormous price difference between 2 Athlon MPs and 2 Athlon XPs (with closed L5 bridges making them essentially MPs) was also a black mark against AMD that didn't help the MP's sales at all.

Re:Main Competitor = Itanium; Not Xeon

[fitten]

Nothing that I've seen on the 'net even hints that the Opterons will be that much of an increase in speed over the Athlon that exists today running at near the same clockspeed. The best figure that I've seen is only 20% improvement in speed in 64-bit mode because of the new registers.

The best database server comparison I've seen in a while is http://www.anandtech.com/IT/showdoc.html?i=1747 which shows that while the AthlonMP can keep up with the non-HT Xeons that are clocked near its PR rating, the Xeons are walking off from that point just by virtue of clockspeed, even before you add in HT.

Remeber, 64 != 32x2 in this case. Clock for clock, the Opterons are probably not going to be *that* much faster than an Athlon on many things. Personally, I'll be impressed if we see a 33% speed improvement at the same clock speed and I would bet that we could attribute a lot of that to FSB speed increases over the AthlonMP.

The Opteron's ace probably isn't going to be a huge performance increase. It's going to be that it can address a lot of memory.

Re:Main Competitor = Itanium; Not Xeon

[Anonymous Coward]

You mean like how Intel screwed themselves with the celeron-A?
You'd think AMD would have seen it coming.

Re:Main Competitor = Itanium; Not Xeon

[cbreaker]

Nothing hints to this? What about this very same slashdot article? They are giving the Athlon64's the "PR" ratings of Barton-core AthlonXP's at a much lower clock speed. This seems to be an indication of a pretty good speed increase in just running 32-bit software.

When running 64-bit software, it will depend on how heavily the software utilizes the extra data width. What type of application. And so forth.

It all depends on your perspective. Perhaps in your case, addressing 20GB of RAM is important to you. For me, 4GB is more then enough for any application I've used besides large database applications.

For me, a 20% speed increase out of the box before 64-bit software comes into play is quite acceptable. IMO, as long as the Athlon64 remains competitive with the Pentium 4 in terms of performance, then I'll love buying and using AMD chips now and in the future.

Besides, can you see of any other way the mass-market of computer users is going to be able to get their hands on 64-bit computers? I'm more excited about higher performance desktops then the high end server market.

Re:Main Competitor = Itanium; Not Xeon

[fitten]

Yeah... like the Barton that has a higher PR rating than the 2800+/333 T-Bred but the T-Bred beats it in most of the benchmarks you find...

I've seen the 3dmark2001 scores (high 7000s - far from spectacular) and some of the other more "real" benchmarks. All have been fairly unexciting and been "explained away" by using a variety of statments from #include

I'm just not going to go crow about performance figures that are still fantasy *especially* when they are based simply and completely on some numbering scheme used for marketing that has already been proven to be inaccurate, and especially when I've seen plenty of evidence already that seems to say that there will be only modest speedups on existing IA32 code. To be honest, I'm as excited about Banias in a desktop as much as the Opteron or Athlon64 because it means that I may be able to have practically silent machines again :)

Re:Main Competitor = Itanium; Not Xeon

[cbreaker]

I don't care if my computer is silent or not. I want more performance, more speed.

Can I see this evidence you speak of that says that there won't be much speedups on x86-64 chips?

Re:Main Competitor = Itanium; Not Xeon

[fitten]

Some of these I found on the web by doing a search on Google using
athlon64 benchmark

http://www.theinquirer.net/?article=8313
http:/ /www.blanos.com/benchmark/bprint.cgi

Here is the link. Unfortunately, many of them are in languages that I can't speak or read.
http://www.google.com/search?hl=en&ie=UTF-8 &oe=UTF -8&q=athlon64+benchmark

Re:Main Competitor = Itanium; Not Xeon

[cbreaker]

The first link is a very short non descriptive article with no information about the machine, video board, etc. Except "sample board sample chipset beta drivers."

The second link contains no benchmarks for Athlon64 and Operon chips.

From what I've seen, on benchmark sites such as:

http://www.heise.de/ct/english/02/26/018/

The new CPU core has a lot of potential. It's not going to be twice as fast. It's not going to change the world. But, I think it's going to be a great CPU for a great many purposes. I can't wait to own one. Unfortunately I'll have to wait until I can afford one. (And find a new profession..)

I purchased an AthlonXP 2200+ last year, so it probably won't be until next year when I start getting the urge to upgrade.

Re:Main Competitor = Itanium; Not Xeon

[fitten]

I read VIA chipset with Geforce4Ti4800 in the first one along with the "not finalized drivers, etc" stuff.

Anyway... I agree, I think it would be cool to own one, but I honestly have no real use for 64-bits at home right now other than the cool factor. A faster clocked Athlon or P4 would be just as useful to me. The Banias at higher clockspeeds (I already have machines faster than the P4@2.2GHz at home) would be interesting as well because of the low heat generation.

Re:Main Competitor = Itanium; Not Xeon

[Hoser McMoose]

AMD's Opteron processor will be offered in three varieties:

Opteron 1xx series - Single processor workstations/servers
Opteron 2xx series - Dual processor workstations/server
Opteron 8xx series - Up to 8-way SMP servers

The first two are pretty directly targeting the market that the Xeons currently sell into, but the last one starts to touch on the Itanium's market. The Xeon tends not to scale well beyond 2 processors, and in fact, most Xeons won't work at all in anything more than 2-way systems (only the Xeon MP is certified for use in 4+ processor systems).

As for price, AMD hasn't announce any prices, but I think it's quite reasonable to assume that the 1xx series of Opterons will be quite reasonably priced (probably not significantly more than an AthlonXP or P4), the 2xx series will be priced similar to the Xeon (about 10-20% more than a P4 of the same clock rate), while the 8xx chips will cost about the same as a Xeon MP ($1000-$2000).

Of course, when it comes to servers, the processor is only one relatively small part of the equation, and so far all the Opteron servers I've seen have been pretty high-end systems, so I don't expect them to be cheap.

Re:Main Competitor = Itanium; Not Xeon

[TheCrimsonUnbeliever]

Windows XP supports 64bit (You need a special version I belive)

And Longhorn will no doubt have 64bit support

What I want to see is applications supporting 64bit dual operations

Well.....games but - You know what I mean

Re:Main Competitor = Itanium; Not Xeon

[rsborg]

As for price, they will have to be cheap to stay competitive with Intel's 32-bit offerings.

What, you mean the Xeons they will be competing against? Even if they sell for 2/3 the cost of the Xeon, they'd be rakin' in the bucks, if they get decent yields (the Itanium, for example, has horrible yields due to it's large L3 cache and die size).

The Xeon is Intel's low-end luxury model; If AMD can provide a full-luxury model at lower rates, then they'll pull a Lexus on Detroit.

Re:Main Competitor = Itanium; Not Xeon

[cbreaker]

> What, you mean the Xeons they will be competing against?

Well, no. Opterons will be the only ones that will support over two processors. I would think that the price of an Opteron will be competitive with a high end Xeon, while offering much better performance, but who knows.

I mean the Athlon64. If you look at their roadmap, they are replacing all of their 32-bit CPU's with the new 64-bit CPU core. This means that the Athlon64 will have to compete with the Pentium 4, too. This is good, since it will force the prices down.

Re:Main Competitor = Itanium; Not Xeon

[cbreaker]

Yea me too. Obviously you need applications support or it's no use.

Hopefully games adopt the new 64-bit stuff quickly, as these will show better the type of performance gains that can be had. Video Editing type software will most likely also gain a significant advantage while rendering.

For general purpose software, like Office, we're not likely to see much in terms of performance gains. I mean, OfficeXP runs great on a 400Mhz Pentium II.

super duper

[Anonymous Coward]

imagine a beowulf cluster of these

possible practical uses of opterons:

* super duper fast game server
and then of course, um.......

tebibytes

[aminorex]

Just a word to root for the ISO standard terminology,
which disambiguates the base-10 and base-2 scale
prefixes: 2^40 bytes is a tebibyte.

Also, although the Hammer implementation may be
limited by physical address lines, I believe the
AMD architecture uses a 48-bit address space,
for 256 tebibytes of addressable memory.

addresing?

[simpl3x]

with longhorn-ish systems on the horizon, where the os contains database functions, will this substantially increase performance? being able to address everything into a specific memory allocation would seem to imply some application performance gains in say image processing...

Cost of Multiprocessing

[LordMyren]

Its going to be really sad how quad motherboards are still going to cost at least $800, even though they've got to be downright trivial to make compared to modern quad system. No central switching logic, just interconnect buses between processors! PCB and sockets aint that expensive, there's really no excuse. But its going to happen anyways.

Speak nothing of the many-thousand dollar eight processor boards.

Damn cushy profit margins.

The Abit BP6 was my introduction to low cost SMP. Now I've got a craving for more, but I dont think its going to happen. Even thought it could.

Maybe someone will get smart and make a enthusiast board. I seriously doubt it though. Not when there's bigger fish to fry. How long is it going to take for someone to realize that although less profitable, there will be untapped demand for non-server class quad systems.

Myren

Re:Cost of Multiprocessing

[Papineau]

If it's really that easy to do, somebody ought to design a motherboard for it and sell it cheaper. Or, it could be that people actually needing those boards are ready to put some more cash on the table than you...

That first part might sound crazy, but would it be possible to do? Doing the layout should go quite smoothly, if you can sufficiently simulate the signals (for noise, cross-talk, etc.). Then, it's gerber files, and to the manufacturing. Getting the specialty parts (south bridge, BIOS, etc.) might be trickier, but it should be possible to do (although probably not for an individual). Somebody wants to start his own company?

Re:Cost of Multiprocessing

[zenst]

I find it supprising that the chipset manufactures dont release some OEM board dirt cheap doing 4 or 8 cpu's. That way they would make there money on the CPU's. Kind of a loss leader in elling chips in a way though akin to McDonalds giving you the burger letting you chose your prefered brand of drink and charging you an arm and a leg for every frenchfry errr sorry freedom fries, O bloody hell there chips, small chips; Tiny lickle bits of potatoes cuts realy small and mean because nobody can be assed to cook chips properly as it takes time (tip is to cook them partualy then alow to cool and then carry on cooking ;). So in summary cheaper multi CPU boards from the manufacturers sells more CPU's and more memory == profit/rather swanky large collection of underpants.

Re:Cost of Multiprocessing

[Anonymous Coward]

They do - sort of. They're called reference designs, and they're fscking expensive.

But yeah, you're right. Fear not however, Taiwan or China will come to out rescue.

Re the "more memory" thing, maybe Crucial should be making 4 and 8 way mobos...

Re:Cost of Multiprocessing

[LordMyren]

The processor interconnect nature of the opeteron's are trivial compared to previous SMP architectures, in that the processors directly communicate with each other rather than using a switched architecture (one which requires very high speed fancy flying hard to use logic to function properly. aka: the hard part).

But you still have to design a motherboard. That alone is well beyond the scope of most mortal men. Electricity does some crazy things at 2.0 GHz.

Continually more and more, we're left at the hands of companies to deliver this future. True entrepreneurs aside, individuals have less and less power in issues like this.

Re:Cost of Multiprocessing

[nusuth]

Quad capable boards are useless if you don't pair them with four of quad SMP capable cpus. That will definetly dwarf 800$ price tag of the board, at least until atlon64 is the mainstream cpu and opterons are being mass produced too.

Re:Cost of Multiprocessing

[Herkules]

Well one posetiv thing is that the AMD CEO has been quoted of wanting to keep avrage sales price at 100 - 150 USD.

Could mean cheep Optrions! (in the 100 and not 1000 of usd)

Re:Check out the linux bechmarks with optron!

[Anonymous Coward]

You take RTFA to a new level- you didn't even read the post you responded to!

On sale

[whereiswaldo]

Some of the notable features of the new chip are an address space capable of addressing up to 1 Terabyte of memory

That's awesome because, Dude, Dell is having a sale on half terabyte memory modules this week!

But seriously, how could one ever pack that much memory onto a board? It sounds physically impossible unless you have some crazy expansion pack but that would slow the system down so what's the point?

Re:On sale

[sprouty76]

Addressable memory != physical memory Pages stored on disk are also addresable, they get swapped in when accessed.

Re:On sale

[BenjyD]

The current 4Gb limit is pretty easy to hit: 4 1 gb dimms would only be around $1200 or so. Maybe ten years ago memory was, what, $100 for 4mb or something? Continuing that trend means that 1Tb will only cost $3600 in ten years.

If you're going to make an architecture change you might as well make sure you don't have to do it again for a while. Nobody wants to be caught saying "640kb should be enough" any more!

Re:On sale

[whereiswaldo]

Addressable memory != physical memory Pages stored on disk are also addresable, they get swapped in when accessed.

Yes, of course, but I don't know anybody who'd want to run their app relying on swap space. Can you spell SLOW? You'd be better off having many computers with 4GB of RAM per.

Re:On sale

[Anonymous Coward]

Don't recall where, but I do remember reading something about Opteron boards being produced with 8 DDR DIMM slots. can't help ya find it, and too damn lazy to find it myself, so I'm not being too terribly helpful, but....

Re:On sale

[QuadGoatBoy]

... shouldn't have read that with enthusiasm... I just went browsing through Dell's website looking at what kind of prices to expect on a half terabyte memory sale... LOL! Mondays...

Quadgoatboy

High Availability?

[VortexVertigo]

Will it be able to compete with the big boys in High Availability 64-bit computing? It will need hot-swappable CPUs, PCI-X slots, and maybe even memory (which means being able to define a particular memory slot as inactive and then swapping it out).

Even without these features, I see the Opteron competing aggressively with Intel's Xeon line of processors. If it has HA features Intel will have its hands full trying to quench the Opteron's flames with FUD.

NUMA Implications?

[Lank]

In the article, it said "...AMD links memory directly to each CPU", which right off the bat makes me think NUMA (Non Uniform Memory Access). NUMA machines are made up of nodes, which have a processor or two, and a bank of local memory. But one CPU can access any memory in the system, it just takes longer if it isn't local - hence the name NUMA. Did anyone else get that impression?

Re:NUMA Implications?

[Wesley Felter]

Yes, Hammer is NUMA. AMD's marketing department is trying hard to convince people that it's neither NUMA nor UMA but something in between. If the technical details that have been released are correct, then a NUMA OS won't be needed, but I guess we'll have to wait for some actual experiments to know for sure.

Re:NUMA Implications?

[Hoser McMoose]

Yup, the systems will be NUMA. However, AMD is trying to coin a new phrase for their chips, SUMO (Sufficiently Uniform Memory Organisation, I think). They even have a little drawing of a sumo wrestler on some of their documents :>

The idea is that they will use their Hyptertransport links as a high-speed, cache-coherant, memory bus. Their claim is that the interconnect links will have enough bandwidth and sufficiently low latency that you can completley ignore the fact that it's a NUMA system and treat it exactly the same as a UMA system.

In reality, the chips will probably benefit from some NUMA optimizations, however the setup does look pretty good. Typically accessing remote memory will still be just as fast as accessing memory through an external memory controller, as is done in all current x86 systems.

There's more to it than that.

[kangasloth]

What makes the Opteron a server chip is the presense of three hypertransport links, the bus used for communication between multiple CPUs and other components such as the motherboard chipset. The Athlon64 will have only one. This is important since hypertransport, unlike say PCI, uses point-to-point links. The AGP and PCI bridges could be on separate hypertransport links and in theory we could see things like gigE controllers directly attached to the hypertransport bus.

Also, last I heard, the Opteron will use Dual DDR memory, while Athlon64 will have to make do with single-channel DDR. Recall that both Hammer chips (SledgeHammer, aka Opteron, and ClawHammer, aka Athlon64), have the memmory controller integrated onto the CPU.

For both of these reasons, the Opteron and Athlon64 sockets are incompatible (Socket 754 vs Socket 940). There's an old review with plenty of information here

Re:Check out the linux bechmarks with optron!

[Anonymous Coward]

Sun has Linux. Sun will introduce Opteron systems just as soon as the preformance whips SPARC's ass. Why do you think they kept Solaris x86 alive, eh?

And they'll keep the service. Solaris is more-or-less free, the hardware isn't that expensive considering the quality, and the service is top notch. If they move with the market, I'd give 'em a few more years of life before an Oracle merger.

Caveat

[bstadil]

So in summary cheaper multi CPU boards from the manufacturers sells more CPU's

Yes, but it does not generate more revenue or profit. The BP6 (I am still using this board as my main machine even though I have 2 faster single cpu machines) caused Intel to disable SMP in the celeron . It was such a succes that it ate into the "premium" brand P3.

AMD need the support of the "serious" vendors so the last thing they will do is themselves offer somethng that will undermine that support. Expect SMP boards to be high prices for a while.

Dell is a distributor

[bstadil]

Dell is not a computer company. They are a distributor for Intel.

In the sc industry any national distributor will normally only carry one major line. This has been the Rule since late 60's.

Re:Dell is a distributor

[Pharmboy]

Dell used to sell AMD chips in their systems, at two different periods, so it certainly is possible.

Here's where 64-bit becomes useful:

[MtViewGuy]

Multimedia.

The problem is that with increasingly capable multimedia capabilities of desktop machines we're starting to approach the limits of 32-bit computing. We could use the capabilities of 64-bit computing for serious image-editing (especially now with 5 megapixel and higher digital cameras dropping in price rapidly!) and serious video editing (e.g., editing videos downloaded from your MiniDV format camcorder and then burning a DVD-R disc from the edited video).

Also, games could benefit from 64-bit processing. Imagine being able to process larger amounts of data to do 3-D graphics that will be of much higher quality than currently possible--games will start to look something akin to participating in a photorealistic movie.

Re:Here's where 64-bit becomes useful:

[sql*kitten]

Also, games could benefit from 64-bit processing. Imagine being able to process larger amounts of data to do 3-D graphics that will be of much higher quality than currently possible--games will start to look something akin to participating in a photorealistic movie.

You are assuming that a 64-bit processor can actually process all the data in its larger address space in the time it takes a 32-bit processor to process data in its smaller address space. That would mean a 64-bit processor would have to be thousands of times faster than a 32, but really you'll be lucky if it's twice as fast.

Yes, you can process larger amounts of data in one go, but not that much faster than breaking it into 4G chunks and doing the processing sequentially on a 32-bit.

Re:Here's where 64-bit becomes useful:

[Natalie's Hot Grits]

"You are assuming that a 64-bit processor can actually process all the data in its larger address space in the time it takes a 32-bit processor to process data in its smaller address space. That would mean a 64-bit processor would have to be thousands of times faster than a 32, but really you'll be lucky if it's twice as fast."

Actually, YOU are assuming that is what he means. What he really means is that currently, video games need to do 64 bit math. Currently, on a 32 bit CPU, you have to do multiple instructions to do 64 bit integer math. With 64 bit computing on the new AMD x86-64 platform, you will be able to process 64 bit integers in LESS time than you can on currently available Athlon CPU's. Your parent post said nothing about >4G of memory. He is talking about data processing speed. The facts are that on AMD's new 64 bit platform, nearly all processing (both 64 and 32 bit chuncks of data) is inherantly faster. To top it off, processing of 64 bit numbers is more than two times faster than on traditional 32 bit machines.

(Faster, because there is more cache, and more general purpose registers, and memory latency is reduced by having on chip memory controler, and on 32 bit machines, it takes at minimum of 2 instructions to process 64 bit integers [and in actuality, more than that])

But require that the high bits are zero

[Animats]

In early Motorola 68xxx machines, only 24 bits of the 32-bit address space were implemented. The higher 8 bits were "don't care" bits. When the address space expanded, much effort went into making Mac software "32-bit clean", and various compatibility hacks persisted for years.

Let's not make that mistake again.

You hit the nail on the head

[carlmenezes]

Yeh...and guess what? It's going to be double your RAM size :)

Condon

[Anonymous Coward]

Condon, Boron,

athlon = no halt
duron = round
barton = born at
opteron = one port

pentium = put me in
celeron = cern ole (tribute to the web?)
deshutes = seed shut
klamath = ham talk
mendocino = coin demon

CowboyNeal = only be waco
computer = cpu metro
processor = sore corps
monitor = it moron
motherboard = her mad robot (red bathroom? hatred broom? hard bet room?)
microprocessor = crisp score room (?)
slashdot = dash lost (stash old?)
microsoft = is comfort (???) -> scoot firm!
electronics = score client, circle stone?

(thanks wordplay).

Hey, jerk.

[Anonymous Coward]

Re: Your sig -- I moderate anything with "M$" or "Windoze" as troll.

I metamoderate posts marked as troll or flamebait simply because the poster decided to make a small, name-related insult -- independant of a far larger intelligent point -- as unfair moderation.

Re:Hey, jerk.

[Anonymous Coward]

Good for you. M$ Sux0rs a big fatty!

No nice uptime for Opterons

[G3ckoG33k]

Sorry, the uptime for Opterons will be worse than any 64-bit competitor. Why? you may ask. Basically, there are two reasons: ut2k3 and CS, both of which run natively on x86-64. So, what will all vigilant admins do? Just a guess.

Re:you forgo palomino....

[Anonymous Coward]

you forgot palomino:

(in Spanish) palomino = pigeon's shit.

Probably to late to be read...

[wray]

But AMD has shown that many times code is REDUCED in size due to fewer Load/Store instructions because of the increased number of registers available to the compiler. Additionally, for a while, they have chosen the pointer to be only 40 bits (one byte extra) even though the architecture handles 64 bits.

8 way SMP.

[Kibo]

I knew a guy with eight. I was very tired just watching a few of this very patient man's children bounce off everything they saw. I would imagine the worst thirty seconds of his week would be a very effective birth control advertisement.

256TB?

[p3d0]

AFAIK, Opterons have a 48-bit addressing--via 4-level page tables of 4KB each--allowing them to access 256TB, not 1TB. Maybe the chip only has 40 external address lines?

Last Post!

[alpg]

BOFH excuse #247:

Due to Federal Budget problems we have been forced to cut back on the number of users able to access the system at one time. (namely none allowed....)

- this post brought to you by the Automated Last Post Generator...