AMD Takes 25 Percent of Server Market 164
An anonymous reader writes "AMD has taken 25 percent of the server market for itself, according to a News.com article. This gives them some 21 percent of the entire x86 market, and is an increase from only 16 percent in the second quarter of 2005." From the article: "AMD has been picking away at Intel's server market share for several years based on the superior performance and power consumption of its Opteron processor. But Intel fired back last month with a new Xeon processor based on its Core microarchitecture that appears to be outperforming current Opteron processors on several tasks. Intel is pinning its hopes of resurrecting its market share--and its stock price--on the new Core generation of processors."
The Intel monopoly? (Score:2, Insightful)
Re:The Intel monopoly? (Score:3, Insightful)
Re:The Intel monopoly? (Score:5, Insightful)
This is why I see Windows as a monopoly -- in order to be certain of being able to run all of the Windows applications out there, you need to have Windows, not Wine or MacOS etc.
Competition is a good thing. I've traditionally run AMD chips in my machines, since I've had good results and gotten good value, but I wish Intel well, too -- if only to keep AMD honest.
Re:The Intel monopoly? (Score:3, Interesting)
Re:The Intel monopoly? (Score:3, Insightful)
In the same way, writing code portable between Windows and Linux is hard, so Linux isn't in the same market as Windows (unless it is a java app, or web-based).
Chips? (Score:2)
Re:Chips? (Score:2)
Re:The Intel monopoly? (Score:2)
That's easy... because it's false
freebsd will run linux binaries, newer versions of solaris have limited capacity to do the same.
Further, since most "linux apps" are free software & POSIX, you can compile them with no modification on solaris, any of the BSD's, hp-ux, solaris, irix, etc...
Plus (Score:2)
Re:The Intel monopoly? (Score:2)
That "Dell is playing Intel" comment was toungue-in-cheek, and I was modded accordingly. But I am concerned about the lack of availabiity. One of our colocation sites is Dell-only, and I really need to get a new 4-socket database server, and don't want to buy another NetBurst machine. So where are these Opteron-based Dells?
Mega hurts! (Score:4, Interesting)
Re:Mega hurts! (Score:3, Informative)
Re:Mega hurts! (Score:4, Interesting)
Because their chips were terrible. Compare a P4 with an Athlon 64 and there is no content. The Athlon runs cooler, could usually beat the P4 in many contests despite the much lower clock, etc. The fact is that MHz was all Intel had going for it, technology wise. Once that started to slip up that was AMD's big chance. For the past few years most things I've seen have put AMD's Opterons much better than Intel's Xeons.
But, in the great spirit of competition, that will change. Intel's Core 2 Duo architecture looks to be a real winner. If the performance is anywhere near where the early numbers look, then AMD could be in real trouble. If AMD can't pull something out with the Opterons... They won't have a new architecture (K9 for the sake of argument) ready until late '08 early '09 at the earliest.
There is major competition again, this is good for consumers, and should be fun to watch too.
Re:Mega hurts! (Score:2)
Re:Mega hurts! (Score:3, Interesting)
Right now AMD is changing nothing more than sizes and process. In the past year it was Intel who worked on a major stepping in CPU design. On top of that Intel is releasing kentsfield (quadcore) at the end of *this* year, not *next*
I'm not an intel fanboy but I'm certainly not an AMD-zealot. Things change in the CPU industry and AMD is not interested in becoming the fastest cpu maker anymore. The purchase of ATI proves that ATI is planning to become a
Re:Mega hurts! (Score:2)
Re:Mega hurts! (Score:3, Insightful)
Anyway, on the AMD side they should gain some of it back when they shrink from 90nm down to 65nm in the 4th quarter of 2006. I think that gets them some automatic power savings due to the process shrink and possibly a performance boost (higher frequencies?).
But the Intel Core 2 Duo chips are looking like very good chips which definitely catch up with AMDs offerings and even surpass it in some (all?) areas. Their pricing is also rather aggressive for bei
Re:Mega hurts! (Score:2)
Hmm, well I'm glad that's true but for someone like me who's looking to replace his motherboard/CPU and isn't an expert in the CPU market, it's pretty difficult. I'm not sure whether to go with AMD, Intel, or wait. Isn't 'buy now' and 'wait' always the dilemma, given that there are always constant improvements in CPUs and 5 minutes after you've purchased some smartass will say "haha, your chip is slow as
Re:Mega hurts! (Score:2)
Re:Mega hurts! (Score:2)
Hardest part is then figuring out what motherboard to use.
(And I don't care what the smartass's have to say. I buy wherever the "knee" of the price/performance curve is. Usually whatever was "hot" 12 months ago now sells for a good price.)
AMD's advantage is being first-to-market (Score:4, Informative)
Re:AMD's advantage is being first-to-market (Score:2, Insightful)
Re:AMD's advantage is being first-to-market (Score:3, Interesting)
Which was unlike other attempts to move to 64bit which required compromises (running 32bit code in an emulation layer or taking a performance hit on 32bit code). Or that required that you recompile everything into 64bit mode in order to take advantage of the new architecture.
I've said it fo
Re:AMD's advantage is being first-to-market (Score:2)
It will be curious to see what the Core 2 Duo proc (and related intel procs) does. According to reviews I've read (and Maximum PC who's 2006 dream machine contains a Core 2 Duo), the processor is amazing. Under full tilt, it doesn't overheat even with passive cooling, which is a major departure from my Pentium 4 - watercooled to keep the noise down.
~X
But does it do multi-CPU? (Score:2, Interesting)
I think Conroe's advantage is really only apparent in 1P solutions, and thus, to get the biggest mindshare/perception shif
Re:AMD's advantage is being first-to-market (Score:3, Insightful)
Re:AMD's advantage is being first-to-market (Score:4, Insightful)
high performance? Only in floating point math. FP is a big important part of our world these days, it's used even in audio processing, all our games are all FP... But the Athlon's performance is superior in every other category, and it was about a tenth the price or less for the processor alone. The price of a complete solution...
iTanic has gotten precisely the treatment it deserves.
Re:AMD's advantage is being first-to-market (Score:3, Informative)
Curious that you don't mention the Itanium by name... It may have its niche, but expecting the PC market to drop everything and adopt the 'Itanic was pure folly. AMD's solution may have been conservative, but by maintaining backwards compatibility with no performance penalty (often the opposite) there wasn't a good reason _not_ to buy and AMD64 processor. Yes, the x86 ar
Re:AMD's advantage is being first-to-market (Score:3, Informative)
Intel Conroe (Score:2, Interesting)
technology (HyperTransport etc) make it that much more desirable? Or will/should Conroe gain more market share back to Intel?
Re:Intel Conroe (Score:2)
Re:Intel Conroe (Score:3, Informative)
Which I believe are the Woodcrests.
Woodcrests - Server.
Conroe - Desktop.
Merom - Portables.
Re:Intel Conroe (Score:2)
Re:Intel Conroe (Score:3, Informative)
Well Opterons badly lose against equivalently-priced Woodcrests (from the benches we've seen up until now), but the impressive HyperTransport links and the way they open it (custom 'drop in' chips for HT motherboard) will keep it extremely interresting. Plus the aforementioned HTT links make Opterons architectures much easier to scale up.
Re:Intel Conroe (Score:2)
Only for a time. Once you saturate the bandwith on the HTT you are screwed and need bridge glue just like with the Intel offerings. Problem there is the Intel chips expect glue and partitioned memory, the AMDs don't, so your glue will likely cost more.
-nB
Re:Intel Conroe (Score:2, Insightful)
Re:Intel Conroe (Score:5, Informative)
Re:Intel Conroe (Score:2)
Re:Intel Conroe (Score:2)
Re:Intel Conroe (Score:2)
Re:Intel Conroe (Score:3, Insightful)
Re:Intel Conroe (Score:2, Informative)
Re:Intel Conroe (Score:4, Interesting)
First, HyperTransport is an open standard, not an AMD technology, and from what I know, it kicks ass.
AMD has gained recently, especially in the server/HPC market because of a few things. Price, performance, power consumption, backwards compatibility, and 64bit offerings. These are the key variables for server/HPC computing. Power consumption and to some degree 64bit-ness are newer and these variables have increasing weights in today's markets.
Personally, I'm not a fan of the EM64T or other hacks for providing 64bit capabilities over a native 64bit architecture, but then again I've never dealt with these x86 extensions first hand, so I could change my opinion with new information. From what I know the original Intel Core was only 32bit, but the core duo included 64bit abilities.
Re:Intel Conroe (Score:3, Informative)
Core 2 = the foundation of Intel's next generation CPU which are as follows...
Merom - laptop chip - T55xx, T56xx and T7xxx
Conroe - desktop chip - E6xxx, X6800, X6900, etc. - (Core 2 Duo, Core 2 Extreme, etc.)
Woodcrest - server chip - Xeon 51xx
Re:Intel Conroe (Score:2)
I proudly claim ignorance of the solo, duo, 2, duo 2, solo 2 stuff.
The Intel marketing people should be drowned in Sun's java.
Re:Intel Conroe (Score:5, Informative)
(BTW, EM64T = shameless clone/re-branding of x86-64, which is an open standard created by AMD. A rare case of Intel not succumbing to Not Invented Here syndrome. From here on, I'll lump them both under the name "AMD64".)
FWIW, I have very little hands-on experience (not being a frequent programmer of x86 assembly), but there are two big features of AMD64 that stand out: more registers (which helps compilers especially), and addressing relative to %rip (the 64-bit Instruction Pointer). The former lets you compute more things on-the-fly without reserving stack space for temporary variables, which can cut down on round trips to L2 or main memory -- thus making AMD64 a bit more like a RISC system, while leaving behind the ivory tower "orthogonal" (read: code-bloating) instruction sets that RISC forces on you. The latter lets your code reference constant things like strings (which are generally compiled into the .text section, right alongside the code that uses them) without [PIC] reserving a register for it, or [non-PIC] hardcoding the address. This simplifies the build process for a LOT for programmers.
Quick tutorial on PIC:
Let's say I have a function, void hello() { printf("Hello, World!\n"); }. If I compile and link this code normally, I get something that looks like push $0x80484b8; call printf, where 0x80484b8 is a hard-coded address located in the .text section (or else a section for data constants that can be found relative to .text). If you're building an executable, that's fine, since the location of .text will be known at link-time.
However, if you want to bundle your code into a shared library, that won't do at all. Each program that loads your library will load it at a different address, so .text could be anywhere in memory. On a modern system, you can add a fixup so that the dynamic linker patches your code on the fly, but now your "shared" library has one copy in memory per instance, even if it's all instances of the same program. That's worse than a static library! The solution is called PIC, Position Independent Code, and is invoked with -fPIC when using GCC. On x86, it usually looks something like this: call .Lfixup; .Lfixup: pop %ebx. Since x86 provides relative jump/call instructions, you can call to .Lfixup without knowing the absolute address, which pushes %eip on the stack as the return address. After the pop, %ebx now contains the absolute address of the .Lfixup label at runtime, and you can safely access your constants relative to that. (All that fuss just because you can't use %eip directly.)
On the downside, you've now eaten a register (on the already register-starved x86 architecture) and you've blown away most branch predictors, forcing a pipeline stall. Not a biggie if you just do it once in main() or similar, but since this might be a library function, you have to do it each time the function is called, in each function that needs it. Ew. It works, but it's not elegant, and it eats performance very badly if you call a PIC function from within an inner loop, so a lot of programmers just tell their tools to compile the entire program twice: once with PIC, and again without. (That's what all those *.lo files are from GNU libtool.)
AMD64 allows compilers (and assembly writers) to unify PIC and non-PIC code into a single, efficient path. Instead of jumping through hoops to copy %rip to %rbx and locate your constants relative to %rbx, you can just address your constants relative to %rip directly. There's no longer any penalty for using PIC, so compilers can just turn it on by default, saving the world from millions of tiny hassles that add up to one big Ick. It's probably the single most real-world useful thing they could have possibly added to the x86 instruction set.
Competition (Score:3, Insightful)
AMD is about as old as Intel (Score:5, Informative)
For more information on AMD, see: wikipedia on AMD [wikipedia.org]
Re:Competition (Score:2)
"On May 1, 1969, Jerry Sanders and seven friends founded Advanced Micro Devices in the living room of one of the co-founders." (source [amd.com])
Re:Competition (Score:4, Informative)
Not so so Fast, Intel may be getting it all back (Score:2, Informative)
I just read a review on Inetl new C2 chips and from the specs, it apparently is faster by almost an order of magnitude than anything AMD has (im not a intel fan boy as everthing i have right now runs AMD)
Anyway, the most interesting thing about these C2 chisp is how much cooler they are at the same time. I've read on article that said they were able to run them fanless.
anyway, heres another articles http://www.extremetech.com/article2/0,1697,1989036
I think i might be upgrading to these w
Re:Not so so Fast, Intel may be getting it all bac (Score:2)
Of course with that said a lot of data centers as deploying blade based systems with the norm being 2 socket blades... so I believe Intel is targeting the large aspect of the market.
Re:Not so so Fast, Intel may be getting it all bac (Score:4, Insightful)
Re:Not so so Fast, Intel may be getting it all bac (Score:2)
Yeah, all those Cray's don't scale well at all (Score:4, Informative)
Seriously though, Newisys and IBM have chipsets to do 32 Opterons, but why? That market doen't need it for the trouble it would be. Right now, you can do four way glue-less and eight way with little trouble. The next revision, in Decemeber - March-ish timeframe, K8L adds more interconnects, the ability to split HT connections to 8 bits to double connections, and 4 cores per die. This all adds up to 32 way glue-less for a total of 128 cores. The real reason why you don't see large scale single bus style Opterons, is that the combination of the current HyperTransport (ver. 1) and NUMA make for a very chatty bus, which causes performance issues related to scale. The point of HT is that it is routable and switchable by HT chips on the bus-lines, a la Cray. It's just hardly anybody does it.
They scale fine.
Re:Not so so Fast, Intel may be getting it all bac (Score:5, Insightful)
I do not think that means what you seem to think it means.
Re:Not so so Fast, Intel may be getting it all bac (Score:3, Funny)
Re:Not so so Fast, Intel may be getting it all bac (Score:2, Funny)
Re:Not so so Fast, Intel may be getting it all bac (Score:5, Insightful)
One, they are not an order of magnitude faster. I have seen some benchmarks on the Core 2 Duo CPUs versus Athlon X2 CPUs, and in a clock for clock comparison they Core 2 Duo were up to 20% faster in some integer operations. Floating point performance was almost equal, as was memory access. 20% is not an order of magnitude.
Two, we are talking about server CPUs, not desktop CPUs. That means that we need to be comparing Xeon CPUs with Opteron CPUs, not Core 2 and Athlon.
Three, the new Core 2 and Xeon CPUs may be faster one on one, clock for clock, than an Athlon X2 or Opteron, but they still have the same old problem that has haunted Intel CPUs since the birth of the Athlon 64: the FSB. Putting 4+ MB of cache onto the Xeon and Core 2 CPUs helps alleviate some of the FSB bottlenecks (for memory access), but they still can't touch the Hypertransport interconnect for performance. And where this really comes into play is in scalability. If you put two or four Intel CPUs into the same server, they share the FSB. If you put two or four Opteron CPUs into the same server, they each have a dedicated connection to the memory, etc. Opteron-based servers scale much much better than Xeon-based servers. This is especially important now that people are pushing virtualization more and more. Instead of buying 10 small servers to handle 10 different tasks, they're buying a single 4-way server and running 10 virtual servers on it to save money and make better use of the CPU and memory resources that they have.
Re:Not so so Fast, Intel may be getting it all bac (Score:3, Interesting)
At first, when I saw the 4MB numbers, I was worried because Opterons are 2x1MB L2. But once I dug into the real specs and saw that the majority of the Intel line is only 2MB L2 shared cache, I was less worried.
(And worried might be the wrong term. I'd like to see the two companies compete for the next 10-20 years rather then one or the other running away with the performance crown.)
Re:Not so so Fast, Intel may be getting it all bac (Score:3, Insightful)
You're right about that. It's interesting that almost all of the early benchmarks were done with the 4MB cache models, whereas the benchmarks on the 2MB cache models didn't come until later (if at all). The same with retail availability. The only Conroe CPUs available now (outside of buying a ne
Intel has done nothing to address the memory issue (Score:3, Insightful)
Intel has done little to address this.
Re:Intel has done nothing to address the memory is (Score:2)
Intel is using faster memory (for now), however, they still have to deal with the MIO, which increases their latency.
Dynamic cache does not buy you a whole lot if your application cannot find the data that it needs in the cache. Adds to latency.
If you are doing random I/O in RAM, there is not a whole lot that your prediction logic can do to save you.
The benchmarks that I am seeing that show an advantage for Intel are all synthetic tests that rely heavily on cache.
Well... (Score:5, Insightful)
AMD & Apple (Score:2)
I have been a share holder in both since 2001 (not a huge share holder, but a few hundred shares). Anyway, AMD & Apple announce financial results about the same time (maybe the same day) and I noticed that they could announce almost identical results on a per share basis. Yet Apple stock would soar on the news while AMD stock would drop like a rock.
Apple does do PR much better than AMD, but I have been a long term beleiver in AMD.
Re:Well... (Score:2)
Via C3? (Score:3, Interesting)
People are running servers with Via C3 processors?
My desktop machine is powered by a C3/866 but it is a cheap low power (in all senses) processor. If the C3 even makes it onto the radar then it sounds like the statistics are by volume rather than by price. It is a pity that AMD have also stopped producing their Geode, that was aimed at the same markets.
Re:Via C3? (Score:2)
Re:Via C3? (Score:2)
Bet that's where at least 5% of that 5.x% number came from.
-nB
Re:Via C3? (Score:2)
Or maybe I'm just trying to rationalize a use for my old 600MHz C3. My current plans are to do the DHCP / DNS with it so that I can shut other boxes off when they aren't needed.
Re:Via C3? (Score:2)
just because a box's primary job is providing a service over a network (rather than supporting a local user) doesn't nessacerally mean it needs to be fast.
Re:Via C3? (Score:2)
Standard business cycles in CPU industry. (Score:4, Interesting)
Expect Intel to take share back in the 2P and below market (largest market) while AMD will hold onto their lead in the 4P market until at least early next year, and possibly a while longer due to the technical superiority of their HT-based interconnect. Conroe and Woodcrest are undeniably the better uarch's, but when you start scaling to more CPUs the interconnect becomes more and more important.
It's impressive, to say the least, than Intel has managed to make Conroe perform so well without an integrated memory controller. A lot of uninformed fanboys will claim they "cheated" by using so much cache, but there's no cheating in the microprocessor field and even the 2M Allendale units with less cache have stellar performance. I can't wait for them to come out with their next gen chips with CSI and an integrated memory controller, those will be stunning perforers in all sectors.
Re:Standard business cycles in CPU industry. (Score:2, Interesting)
Re:Standard business cycles in CPU industry. (Score:2)
ARS Technica [arstechnica.com]
Soft32 [soft32.com]
Intel's core has it's weaknesses (Score:5, Interesting)
There is no doubt that AMD's solution for connecting multiple cores and processors is superior to Intel's. And when we start to see coprocessors being popped into one CPU socket providing super-accelerated services such as encryption... the shift to AMD will accelerate. I imagine a secure webserver that is able to handle twice the number of concurrent connections is quadrupled because all of the encryption is handled in hardware by a $600 coprocessor. Sure Intel's system will be faster for general purpose activites, but when your talking paying $600 for a coprocessor, or several thousand for additional servers... well you get the idea.
I think that though Intel currently has a leg up, it's only a matter of time before AMD knocks their other leg out from under them.
Now I'm no fanboy, I'm anxiously waiting for the Core 2 Duo to become widely available before I build my next workstation. But I still believe that AMD is eventually going to become the king of server processors, if not the desktop.
Re:Intel's core has it's weaknesses (Score:3, Interesting)
2) If AMD does not match Intel, they will be forced to create coprocessors to supplement the CPU because Intel currently has the faster CPU
3) The time it takes AMD to match Intel gives Intel the same amount of time to stay ahead. The real question: If AMD were really ahead of Intel, why didn't AMD create an Opteron/Athlon killer in the time it took Intel to create their C2D?
4) Intel's shortcoming was sticking to Netburst two years too long. AMD's advan
Re:Intel's core has it's weaknesses (Score:2)
Re:Intel's core has it's weaknesses (Score:2)
In other words, as soon as AMD releases a faster CPU, so can Intel.
The question holds, what can AMD pull out of their hat to BEAT Intel?
Re:Intel's core has it's weaknesses (Score:2)
Umm, because AMD has the better architecture, and Intel isn't going to just throw theirs out overnight.
The only comparisons I've seen have been for 32-bit software. When it comes to 64-bit, Intel is lagging behind. And still there, they're only winning on the float. In a few months you can expect AMD to leapfrog Intel, and then Intel to leapfrog AMD again, etc. If AMD can just continue to maintain pa
Re:Intel's core has it's weaknesses (Score:2)
I can imagine it with a $100 PCI hardware crypto card as well...
Stock Price (Score:3, Informative)
Bang for the buck (Score:5, Insightful)
Re:Bang for the buck (Score:5, Interesting)
Have you seen Intel's pricing for Core 2 based CPUs? They compare if not out compete AMD on price for performance.
Intel's [intel.com] Xeon 5100 series starts at $209 (@1.6GHz) and tops out at $851 (@3GHz) while AMD's [amd.com] dual core Opteron series starts at $316 (Model 265) and tops out at $1051 (Model 285).
Re:Bang for the buck (Score:3, Insightful)
Re:Bang for the buck (Score:2)
Businesses like Google that get under the hood of their servers are the exception not the rule.
I think a huge factor here is the enthusiastic adoption of AMD chips by the big server manufacturers.
Re:Bang for the buck (Score:2)
On the desktops, the new Intel dual-core CPUs are still a bit expensive ($200-$250). But it forced AMD to drop their X2 price
Well... (Score:5, Funny)
Illustrates the inefficiency of the market (Score:5, Interesting)
AMD has taken 25 percent of the server market for itself,
During the time period that this data refers to, AMD's products had a clear lead in price/performance. But they only got a quarter of the market, instead of >90%, which they would have got if purchasers had been knowledgeable and rational.
Re:Illustrates the inefficiency of the market (Score:2)
Re:Illustrates the inefficiency of the market (Score:3)
Re:Illustrates the inefficiency of the market (Score:3, Informative)
Re:Illustrates the inefficiency of the market (Score:2)
They are continually building fabs, they have contracts with 3rd parties to supply cores if they can't meet demand, and are now taking on ATI as well.
Intel is the one that has had recent supply problems. Their serious chipset shortages at the end of 2005 seriously raised prices, and forced many companies to go elsewhere.
Re:Illustrates the inefficiency of the market (Score:2, Insightful)
NYLF (Score:4, Interesting)
Re:NYLF (Score:2)
amd edge endure w servers more than desktop (Score:3, Interesting)
Re:Single Processors are boring (Score:2)
I know you were trying to be funny, but any machine sold at office min or worst buy can be included in a cluster :P
Re:Moot Point Post Core Duo? (Score:2)
With the Core 2 server chips, Woodcrest, Intel is the clear winner in 2 and 4 core systems. However Woodcrest can only be used in 2 and 4 core system
Re:AMD over Intel ANY DAY (Score:2)
I've owned 4 AMD systems... two were K6-2 35