Sun Joins Apple in the Intel Camp for x86 Chips

An anonymous reader writes "Don't worry, SPARC isn't being replaced by Itanic. However, Sun will start using Intel Xeon CPU's in their X86 servers. Further evidence that Intel's Core microarchitecture is winning back a lot of the business that AMD won with Opteron." More coverage at CNN Money and the International Herald Tribune.

Oh Gosh! Sun 386 all over again?

[140Mandak262Jamuna]

Way back in my previous life, as one of the starving PIGS[*], I was the root (of all evil according to some of my students) in a CFD lab and I manage SunOS running on 386 chip. Deja Vu all over again!

Glossary: PIGS= Poor Indian Grad Students

Re:

[multipartmixed]

Let's just hope they don't bring back SunView. What a POS, even in its day!

Re:

[140Mandak262Jamuna]

But hey, they shipped pizzatool along with suntool in the OS. Oh! yeah!

Re:Oh Gosh! Sun 386 all over again?

[xenocide2]

The news here isn't that they're using x86 architecture. It's that they're using Intel chips. They already have AMD Opterons for sale, and they're adding Xeons. And Solaris 10 does run on x86. Some see it as a concession that their server CPU designs are little more than a niche market. That diagnosis is probably correct, and if Sun wants to ever dig themselves out of the "Sun is dying" meme, they'll have to start taking advantage of the fact that their competitors are engaged in a price war, one that's also cutting into profits. Sun can still pride themselves on quality server hardware, support contracts and integration, even if they aren't the designer behind most their chips.

Meanwhile, their Niagra still has some niche applications, and will grow as software is designed for dual and quad core chips. If Niagra does what they say it will, people will be forced to consider one Niagra unit versus 6 Xeon servers. Xeons may have fallen in price recently, but they're still not cheap, so that's a calculus that Sun might win some day.

Re:

[Daishiman]

I really, really don't get where this "Sun is dying" thing is coming from. Having a bunch of friends at IBM and several telcos and consulting businesses, it is simply amazing the number of Sun Fire 25K machines being bought everywhere. These are 72-processor monsters that will set you back a cool $2 million each, and they're in pretty hot demand.

In the market for very large servers, there's only three choices: HP SuperDomes, IBM p590s and p595, and SunFire 25ks. The Sun machines have by far the largest mar

Re:

[Anonymous Coward]

The "Sun is dying" meme stems mostly from the fact they haven't posted a profit in the better part of a decade

Re:

[MemoryDragon]

You should check the numbers, most of the losses stem from the fact, that their stocks have gone down from insance values into normal numbers, especially the billions of losses in 2001 and 2002, if you count out the stock devaluation, Sun has done very solidly, with many years of earnings and a few lof losses (mainly caused due tu buying other companies) The sun is dying rumor is caused mainly due to people seeing big reds but not knowing the exact numbers and those losses mainly are paper losses while sun

Re:

[suggsjc]

you and your technicalities...

Screw your pessimism, I want the FACTS!!!

Re:

[davecb]

xenocide2 wrote: Meanwhile, their Niagra still has some niche applications, and will grow as software is designed for dual and quad core chips

Actually the multi-threaded/multi-core chips run standard applications in parallel without change. This applies to everyone, not just Sun. Java (ie, threaded) and multi-instance apps get an extra benefit.

--dave

Re:

[h4ck7h3p14n37]

My condolences. I've managed Solaris systems for years, almost entirely SPARCs, but the new place I'm working for buys the cheap x86 junk. Now the x86 systems that Sun ships aren't bad despite some issues with the nVidia RAID chip, Intel Gigabit ethernet nics and the damn x86 BIOS, but running it on a generic 1U x86 server is a joke compared to SPARC. You've got no lights out management support (fortunately some of Sun's AMD systems come with a dedicated service processor for this task, but it requires

Thoughts:

[Ayanami Rei]

1) ALOMs on SPARC use a service processor also... so that shouldn't be a big surprise.
2) Set your LOM to defer to BIOS, and your BIOS for serial port emulation, with an OS-mitigated handover (if that config option is there). Then set up your OSs to boot with a serial console. Attach all serial ports to a Cyclades or other terminal server. Enjoy the deliciousness.
3) Buy SATA in pairs and use (software!) RAID 1. It's cheap, fast and easy.
4) Consider using something BESIDES Solaris on x86. Maybe linux or freeb

Re:

[h4ck7h3p14n37]

Regarding points:

1. For some reason I was thinking the ALOMs on the SPARCs were simpler than the service processors on the x86 systems
2. That's how things have been setup, but occasionally the BIOS is misconfigured to do something stupid like remain off after an interruption of power.
3. Doing that via SVM.
4. We're running various releases of FreeBSD, OpenBSD, Red Hat Linux, Solaris, Windows 2003, etc. We're forced to run Solaris for certain systems due to applications being built for a particular release or due

Re:

[_damnit_]

Nice options, but I really prefer to not use RAID 1 since it doesn't protect against filesystem corruption. Your boot device is probably pretty static if your data resides on a SAN array or NAS device. Cron a dd, LVM mirror resync or ufsdump to occur every night or once a week or even every patch cycle. It depends on how static your environment is. A place I visited only updated the mirror a few days after patching to make sure the patchset burned in nicely.

Re:

[drsmithy]

Nice options, but I really prefer to not use RAID 1 since it doesn't protect against filesystem corruption.

Huh ? This is like saying you don't wear a seatbelt because it doesn't protect against the engine catching fire...

Re:

[_damnit_]

Not really. Nice try with the bad analogy though. If you only have two drives, I prefer a second copy that is not part of a continuously mirrored set. There are generally a few bad things that can happen:

Dead disk. If this happens with Raid 1 you are usually OK. There are sometimes problems with performance as the drive dies but with a good RAID controller or software LVM, nothing dramatic should occur. With a dd copy or offline mirror you get a crash and boot from the alternate copy. If things are p

Re:

[drsmithy]

Not really. Nice try with the bad analogy though.

It's a pretty good analogy. RAID isn't supposed to protect you against software failures (be they in the operating system or the operator), it's supposed to protect you against hardware failure.

Saying you don't use RAID because it doesn't protect you against software error is, very much, like saying you don't wear a seatbelt because it won't stop the engine catching fire.

Dead disk. If this happens with Raid 1 you are usually OK. There are sometimes prob

Re:

[_damnit_]

I think it important to note that the point I was making was in regard to cheap servers with 2 to 3 disks. If you have mission critical servers in 1U form factor, that is a mistake in itself unless there is some clustering or such to protect.

Rebooting a server takes on the order of minutes (many minutes, if it has a lot of physical drives attached to it). How much money does your company lose in a few minutes while the relevant service(s) is/are down ? Would it be enough to pay for a RAID controller ?

If we

Re:

[drsmithy]

I think it important to note that the point I was making was in regard to cheap servers with 2 to 3 disks. If you have mission critical servers in 1U form factor, that is a mistake in itself unless there is some clustering or such to protect.

I would argue that the point I'm making applies to _any_ production server. Disk failures are such a common thing to happen, and RAID is such a trivially quick, simple and cheap means of mitigating most common types of disk-related system failures, that IMHO not usin

Re:

[E-Lad]

Not to nitpick, but to weigh in:

1) The only Sun x86 servers which are SATA-only are the low end X2100 and X2200. The rest (X4100 and up) are 2.5", 10k RPM SAS drives. There is one Sun x86 server which has SCSI internally, and that is the left-over V40z. I think Sun still keeps it in the product line up because it has 4 sockets.... but it's a product that will probably be replaced soon.

2) The iLOM included with the X2100/X2200 M2 and the X4x00 servers is actually really, really nice, and just as capable than

Re:

[h4ck7h3p14n37]

Thanks for the info on the Sun servers supporting SAS; we've only purchased X2100, X2200 and v20Z's. I believe we only have the model 1 X2100 and X2200's, but we fortunately did spring for the service processor.

I'm aware of a workaround for the boot archive issue, and yeah we do reboot when the patches require it :) I think what's happening is that we're doing some post-install work with cfengine and are not updating the boot archive. The systems aren't typically restarted when cfengine runs since that'

Re:

[drsmithy]

The only Sun x86 servers which are SATA-only are the low end X2100 and X2200. The rest (X4100 and up) are 2.5", 10k RPM SAS drives.

On the downside, the "RAID controllers" (I use the term loosely) in them are so basic they don't even support RAID10, let alone RAID5 or 6 (I was _not_ happy when we bought a few of them to discover this, and subsequently had to setup a RAID0 (hw) + RAID1 (sw) to compensate).

FFS. A bottom of the barrel 4-channel SATA "RAID" controller will typically support RAID10. To not f

Re:

[E-Lad]

The RAID controllers on the X4x00 servers (excluding the X4500) are built in to the SAS controller chip itself, which is a LSI Logic 1068.

I have a negative view regarding the utility of doing RAID 0 or 1 (or 10 for that matter) in hardware. There's really no performance benefit since those controller chips sit right on a PCIe buss, and RAID 0 and 1 don't require anything of the CPU in terms of parity calculation nor do they have the downside of partial stripe writes at higher RAID levels.

You're also removin

Re:

[drsmithy]

I have a negative view regarding the utility of doing RAID 0 or 1 (or 10 for that matter) in hardware. There's really no performance benefit since those controller chips sit right on a PCIe buss, and RAID 0 and 1 don't require anything of the CPU in terms of parity calculation nor do they have the downside of partial stripe writes at higher RAID levels.

The advantage is in transparency. Ie: you don't have to do any sort of additional funky configuration to make it work in a reliable fashion and your bootl

Competition benefit / AMD warchest

[cblack]

Core 2 Duo does seem to offer some benefits over the current opteron line and I think it is great that server vendors can so easily switch between them for new models. I believe Sun has a fairly sizable portion of the x86 server market and it was good to see a company have such success with AMD CPUs. Overall I think the competition is a good thing, but I do worry a bit that AMD will have trouble regaining sales even if they have the better next gen technology due to decreased profits as they lose server vendor sales. I look forward to a next gen battle based primarily on merit.

I'm sceptical

[btarval]

I agree that competition is good for the consumer, but I have to wonder what the effect will really be on their AMD servers. In the server biz, there's a LOT more to it than today's CPU-intensive benchmarks. The other big thing is IO bandwidth, and this is where AMD has been far more competitive than Intel is. AMD appears to be able to continue this lead, based on both companies claimed roadmaps, the last time I looked.

One can only shove so much data across a single bus, and AMD seems to have realized this, while I don't see this as easily done from Intel.

One of the cool things about AMD is the Hypertransport bus. This allows one to offload various peripherals easily onto separate busses, while still allowing them to be shared across CPU's. Offloading PCI peripherals (for example) onto different busses allows one to achieve higher IO bandwidth. In contrast, Intel's current approach seems to be to shove more and more CPU's onto the same bus.

It's as if Intel has completely forgotten about how to keep the CPU busy - that's the main name of the game, and has been for years (to say the least). Idle CPUs are useless, and the more idle CPUs there are, the sillier it is, IMHO.

And AMD appears to be capable of outdoing Intel in the bandwidth area, for both memory and bus bandwidth.

So it looks to me like AMD will continue to be ahead of Intel as far as top-end server solutions go.

In short, I find this particular move puzzling. Sun has traditionally backed the best performance design, and I see Intel still lagging here overall. This strikes me as more of a marketing move, not a real engineering one. It will be interesting to see how popular these Intel-based servers remain.

Re:

[RightSaidFred99]

Your premise is flawed. First, Intel has and will have dual FSB solutions coming down the pipe. Second, they will have a much better interconnect, in CSI, in the 2008 timeframe. Third - your assumption that "IO bandwidth" is the main factor in performance has been proved false. The Intel chips are winning the majority of benchmarks, and that includes dual-CPU, 8-core systems. It turns out that the "sky is falling!" mentality about the Intel FSB has been proved to be partly nonsense, and partly prematur

Re:I'm sceptical

[afidel]

Glad you bought into the desktop class benchmarks that all the little internet sites like to publish. Meanwhile I'm happily using Opteron servers for serious N-Tier architecture where their performance per watt and lack of I/O bottlenecks is great. I also use some Intel Core based Xeon's for less demanding workloads where they prove to a good match of price/performance and power use per workload. I guess I actually research stuff, test it in my specific situation, and select the best product, unlike the many hacks in IT =)

Oh no you di-in't!

[JeanPaulBob]

Suh-NAP!

Re:

[suggsjc]

Well...doesn't someone think they are Joe Cool...actually testing products before deploying.

I bet you also complain about people not reading the articles here on /.

Some people...

Re:

[RightSaidFred99]

Oooh! N-Tier! "lack of I/O bottlenecks". "less demanding". Good propaganda attempts, but you and I both know you're not saying anything of substance.

What specific "I/O bottlenecks" are you referring to? What on Earth does a "serious N-Tier architecture", which is a business application architecture, have to do with a CPU architecture? You're basically waving your hands around hoping nobody will notice your fanboyism, but in the end you're saying a whole lot of nothing.

What kinds of apps are you ru

Re:

[RightSaidFred99]

You call that a flame? I fell asleep like 3 times trying to get through that gibba jabba. Don't worry, you didn't irritate me. Seriously, I've been more inflamed and irritated by loose stoole, and I don't really get irritated by loose stoole.

Re:

[Joe The Dragon]

Intel's severs chip sets still suck and the amd ones are much better. NFORCE pro kicks Intel's a**
Intel need more pci-e lanes in the chip set and in a 4 way sever havening the ram controller in the cpus is a big boost over have all the cpus needs to go though the chip set add a lot of l2 can only go so far.

Re:I'm sceptical

[cbreaker]

I agree with you. The thing is, even if RightSaidFred99 over there thinks Intel is just as good at SMP configurations, it's only NOW just starting to become a reality. AMD has been using HyperTransport since the first Opteron, released several years ago. You've been able to use 4-way Opteron boxes and achieve MUCH better overall system performance then you ever could with a Xeon. Think VMware. When a dual-CPU Xeon outperforms a Quad-CPU Xeon, there's something wrong with the bus architecture.

The "core" CPU is finally, after over 7 years, perhaps better then the current generation of AMD CPU's, but again, it's still based on the same old North-bridge configuration. While Intel has managed to bump up the speed on this bus a bit, and they can more easily support new and faster RAM because the CPU doesn't have the memory controller, it's still the same old. If you're doing 4-way or more, with heavy applications like busy ESX servers, you're going to get a LOT more performance out of your Opteron system, including 4-way systems utilizing multi-core CPU's. Just because CPU's are going dual and multi-core, doesn't mean enterprise servers will ship with only one socket.

I say Good for Intel, the Core CPU is a good one. But, if you look at everything Intel has been doing with their CPU line lately, you'll see that they are generally copying AMD in a lot of places, starting with EM64T (aka AMD64.)

Re:

[TheRaven64]

I say Good for Intel, the Core CPU is a good one. But, if you look at everything Intel has been doing with their CPU line lately, you'll see that they are generally copying AMD in a lot of places, starting with EM64T (aka AMD64.)

Copying from AMD? Not really. Intel had dual-core before AMD (although you could say they copied this from IBM). They had SMT (HyperThreading) before AMD, although seem to have abandoned it in their recent chips (again, IBM had this first too). x86-64 extensions? Sure, but they are some minor tweaks to the ISA. SSE4 is a similar sized change, and they didn't copy this from AMD (although AMD are now implementing SSE in their chips...). What about dynamic shared caches? Nope, Intel got this one befor

Re:

[Jeff DeMaagd]

I haven't seen the benchmark where an Intel quad processor was beat out more than a little bit by an Intel dual of the same core design. The next four socket Intel platform (IIRC, Tigerton) will give each socket its own front side bus.

I don't think AMD was the first to integrate a memory bus into the processor, just that AMD was the first to make it in a mass market processor. Intel will come around on the on-die memory controller deal, that is a little puzzling why it's not until late this year or some t

Re:

[cbreaker]

Big company or not, I wouldn't assume anything.

But, yea - you're right - I mean, neither Intel nor AMD really create so much new technology on their own, they all base everything off each other's work. And IBM tends to be a leader in actually inventing things.

But, AMD made a lot of right moves in the X86 chip design and market, so Intel is now playing catch-up. You could nit-pick the little details like TheRaven64 did (obviously an Intel fan boy) but some of the major changes in X86 architecture have bee

Re:

[drsmithy]

I agree with you. The thing is, even if RightSaidFred99 over there thinks Intel is just as good at SMP configurations, it's only NOW just starting to become a reality.

And by "NOW", you mean "after AMD's brief stint at the top of the x86 [multiprocessor] heap, they've fallen behind again".

The "core" CPU is finally, after over 7 years, perhaps better then the current generation of AMD CPU's, [...]

"Current generation AMD CPUs" have only existed for about 3.5 years. Prior to that, intel held the crown in

Re:

[cbreaker]

Don't tell me what I mean. The NetBurst archetecture was dead on arrival, and you know that's what I was talking about. Pentium 4 always sucked. Intel had to drive them so hard, make them so hot, to compete. It's been that way for years.

Just an FYI, while you admit that AMD has had the "crown" for SMP for 3.5 years, there weren't many more years then that BEFORE that, with SMP intel boxes. Come on, how many SMP x86 boxes were around in 1996?

"Few server tasks are meaningfully bottlenecked by the CPU

Re:

[RAMMS+EIN]

To confirm your doubts: I have a friend who runs benchmarks on high-end systems, and he says Intel CPUs outperform AMDs as long as there are few threads, but grind to a halt when the number of threads goes up. He says it's because of the caching strategy (I know Intel and AMD use different strategies, but I can't remember who uses which).

Re:

[MojoStan]

So it looks to me like AMD will continue to be ahead of Intel as far as top-end server solutions go.

In short, I find this particular move puzzling. Sun has traditionally backed the best performance design, and I see Intel still lagging here overall.

I find your finding (that this particular move is puzzling) puzzling. Sun offers more than just "top-end server solutions." Some of their current Opteron servers are 1-2 CPU solutions, like the Sun Fire X2200 M2 [sun.com]. In this category, Kentsfield/Clovertown soluti

Re:

[Gr8Apes]

For 1 or 2 socket systems, at most, and for 2 socket systems, it depends upon the application(s) running.

AMD's still running on 3 year old tech/fabrication. Until their new lines come out starting this quarter, Intel's got the performance crown only. Multiple sockets is still ruled by AMD, something about NUMA....

This is a surprise

[Tim C]

I was at JavaUk06 [sun.com] last year, and in his keynote speech (one of) the Marketing VPs spent quite a lot of time extolling the virtues of their new line of SunFire servers, paying particular attention to their power:performance ratio compared to similar Xeon-based servers. Listening to him then, you'd have thought that Opterons were the best thing since sliced bread. Yes, I realise that his job is to push their current and up-coming products and solutions, but the main thrust of his talk was "Opteron-powered SunFire servers use far less power than those crappy, power-hungry Xeon servers".

Re:This is a surprise

[be-fan]

If this was last year, he was likely talking about the older (P4-based) Xeon, not the new (Core 2-based) Xeon.

Re:

[RAMMS+EIN]

``If this was last year, he was likely talking about the older (P4-based) Xeon, not the new (Core 2-based) Xeon.''

And the fact that people get confused about this just goes to show how important naming is.

Re:

[Jeff DeMaagd]

The site said that the convention was 15 March 2006, that definitely predates the Core2-based Xeons. It's quite amazing how much can happen in a year. That's not to say that AMD's offerings are a slouch with regards to power consumption. The FB-DIMMs require a pretty hot chip to operate them, and each FB-DIMM is effectively its own memory bus.

Re:

[Anonymous Coward]

"you'd have thought that Opterons were the best thing since sliced bread."

In fact it was at the time when it was competing against Xeon's based on the P4 core. Things have changed since the new Core2 based Xeon's have come on the market.

Re:

[DemoLiter3]

Wasn't Core architecture found eating about as much as AMD, considering that AMD CPUs include memory controller, and Intel has it on the Northbridge?

Re:

[DrSkwid]

9 months ago.

9

months

Does this mean...

[vought]

That AMD will change the rate at which that "Number of dollars saved using AMD servers" billboard counts at?

So, ahhhh...

[Anonymous Coward]

Sun's using Intel chips in their servers that use Intel chips?

Zounds!

No story here, move along.

Re:So, ahhhh...

[TheThiefMaster]

No, they're using Intel chips in their line of servers that previously used AMD chips. For the pro-AMD slashdotters, this is "a very bad thing"(tm).

Re:

[rivaldufus]

and for those who like a little competition in the computer hardware field, that's a bad thing. Guess what folks? Once Intel regains all the share it lost to AMD, those chip prices aren't going to go down.

Re:

[xjerky]

That's the way competition works. All AMD has to do to stay in the game is to release a new line of chips that can compete on price/preformance/power consumption.

Re:

[TheRaven64]

Yes, that's 'all' AMD has to do. If you think that's simple, here's a quick overview of how a CPU is developed:

1. Management asks the materials guys roughly how many transistors they think they are going to be able to put on a die in five years.
2. Management asks the market research people what factors the chip-buying public is going to find important in five years (e.g. raw throughput, multithreading performance, power usage).
3. Management tells a team of a few hundred engineers to develop a chip that is stron

Re:

[xjerky]

I never said it was easy. Just that it appears that Intel has won this round fair and square. But for the sake of continued competition, I hope AMD can top them again before long.

Re:

[hawaiian717]

Until AMD catches up and lowers prices to get some of that market share back, at which point Intel will have to lower prices again to keep from losing too much market share, and so on...

Re:

[Anonymous Coward]

No, they're using Intel chips in their line of servers that previously used AMD chips. For the pro-AMD slashdotters, this is "a very bad thing"(tm).

Not really, it makes AMD chips cheaper. Who cares if Intel has a 10% performance edge for an extra 30 watts? Me, I will stay with the more efficient and cost effective AMD. Especially since my P4 with a Intel PERL mobo died early.

Re:

[teg]

It's not "more performance for more power" - currently, it is "more performance, less power". At the system level, so effects of built in memory controller has been counted.

Intel currently has better performance per core, more cores per socket and less power per core than AMD - AMD has done pretty much one thing in many years: Dual core. Time to get off the laurels from that time and improve their chips again.

Re:

[saleenS281]

Actually, the *really bad thing* is that ANY slashdotter is *pro* *insert any corporation of your choice*. Corporations are out to make money, nothing more. You should be happy Sun is giving you a choice of both competitors because in the end, we, the consumers, win.

But the interesting bit is

[$RANDOMLUSER]

...and for Intel to endorse Sun's Solaris operating system...

What's that about??

Google is your... oh, never mind

[fm6]

A random story from Google News:

Although [Solaris] is currently certified for Intel processors, Intel will dedicate developers to help Sun fine-tune the software for advanced features on its processors such as hardware support for virtualization, storage and I/O. Intel makes its chip features available to developers as open standards, but help from Intel is expected to allow Sun to faster and better support those features its software.

Sun needs this

[Salvance]

Sun has been in trouble for years, and this is a smart first step to getting out of it. Their chips are no longer the powerhouses they once were, and we're truly moving to a commodity chip market anyway. I hope this marks the beginning of Sun moving entirely to Intel/x86 based chips, this way Sun can focus on their other ailing businesses. Sun (just like Mac) is not big enough to keep up with AMD and Intel on chip performance, so why spend Millions/Billions trying?

Re:Sun needs this

[nwhitehorn]

Sun is moving ahead with their SPARC servers, and just taped out a successor to the Niagara. If you'd read the article, you'd know they are replacing their (quite excellent) AMD servers with Intel ones, not SPARC with anything. Sun has quite happily been selling both architectures for some time now.

Re:

[Anonymous Coward]

I hope this marks the beginning of Sun moving entirely to Intel/x86 based chips, this way Sun can focus on their other ailing businesses. Sun (just like Mac) is not big enough to keep up with AMD and Intel on chip performance, so why spend Millions/Billions trying?

Don't count on it just yet. As a former Solaris jockey and proud owner of a Core 2 Duo E6700 on my desktop (it screams!), I think this is a smart move at the lower end of their server market. As far as the high end goes, I don't see Intel's stuff

Re:

[delirium of disorder]

What's with the FUD? Sun is not dying. Their stock [yahoo.com] has been going up over the past two years. Their latest CPU, the UltraSPARC T1 [sun.com] has 32 execution units (eight cores), massive cache and register files, and the highest throughput in the industry. It's not great for floating point tasks so it won't make a good toy system for playing games, but for a real work in a server, it's the most energy efficient, powerful architecture available.

Yeah, I think you're right

[Concern]

What I'm seeing now are people who went google-style with blades buying empty rackspace to cope with hosting providers' power per rack ratio.

Meanwhile Sun's sales guys are selling $14k 72 watt, 8-way, 32-thread T2000's that can replace multiple Opteron (or Core :) blades... IF you're within its application domain. Interesting gamble.

Most webapps probably are... not actually a lot of hot floating point, or math code in general, in that space. But you have to be very careful.

So, it's possible that Sun has turned their biggest disadvantage into their biggest advantage: they're in a niche! Yet they can design whole hardware architectures. So it frees them up to find ways to specialize, and it seems that there may be some payoffs there.

Re:

[InsaneGeek]

but for a real work in a server, it's the most energy efficient, powerful architecture available.

I tend to disagree with that statement, for traditional java, oracle, web serving, etc server loads the Intel/AMD processor has consistently had better performance and with our Opterons is much more power efficient as well. We have found that on those operations the Intel/AMD processors have traditionally outperformed the Sparc proc by 2 to 3 times. The benefit that sparc traditionally have given you is bus sp

Re:

[InsaneGeek]

Redhat AS 4.0 x86_64, using 10gR2 with ASM and attatched to an EMC DMX.

Re:Mac?

[Bastian]

SPARCstation has been in trouble for years, and this is a smart first step to getting out of it. Their chips are no longer the powerhouses they once were, and we're truly moving to a commodity chip market anyway. I hope this marks the beginning of SPARCstation moving entirely to Pentium/x86 based chips, this way SPARCstation can focus on their other ailing businesses. SPARCstation (just like iPod) is not big enough to keep up with Opteron and Pentium on chip performance, so why spend Millions/Billions tryi

Re:

[UnknowingFool]

Their chips are no longer the powerhouses they once were, and we're truly moving to a commodity chip market anyway.

The real issue is that the market is moving to commodity servers because Linux and Windows has made them good enough. Sun servers are good at certain enterprise functions but you don't neccesarily need Unix for file/print, web servers, and in some cases database servers. Before you had to pay $10K for a box and lots of $$$ for support for a Unix box. Today you can get the same functionality

Re:Sun needs this

[BrainInAJar]

"Today you can get the same functionality with several cheap Intel boxes with either Windows or Linux."

Or UNIX for that matter. Solaris is free to use, and a support contract is about half the price of RHEL...

Re:

[mihalis]

Sun has been in trouble for years, and this is a smart first step to getting out of it

With respect, this is a dated analysis. I think their start first step was, you know, launching an entire range of competitive x86 compatible 64-bit servers and workstations, getting serious about storage (storagetek, thumper), upgrading the CEO, upgrading Solaris with tons of awesome new features, cutting the fat in the workforce (at least) and revitalising SPARC with impressive throughput oriented hardware multi-thre

Re:

[Doctor Memory]

Sun (just like Mac [sic]) is not big enough to keep up with AMD and Intel on chip performance

I dunno, I haven't seen either AMD or Intel shipping 8-core chips yet. Sun is. And I'll bet that 8 1.8GHz cores will beat 2 @ 3.4GHz...

Re:Sun needs this

[TheRaven64]

Depends on the workload. For floating point, the T1 is pretty dire (the T2 is better). The thing it does really well is highly concurrent workloads. AMD have a faster interconnect than Intel at the moment, but they are both crippled by one thing: the speed of light (painfully slow). The amount of work that a modern processor could do while waiting for a cache miss is enormous. Intel currently have some very clever logic predicting and pre-fetching cache misses, but they still suffer from them. Sun, however, cheat. Each core on a T1 has 4 contexts (8 on a T2), so when you get a cache miss it just switches over to the next thread instantly and carries on executing. The T1 core will only sit doing nothing if all four cores are in the middle of a cache miss, while an Intel or AMD CPU will do this for a few hundred cycles on every single cache miss.

For desktop workloads, this isn't such a great thing; most current-generation desktop applications do all their work in one thread, so if that thread has a cache miss you still end up doing a load of waiting because the other threads are not using the CPU much. On a server with a few hundred (or thousand) concurrent users, however, there are always threads waiting to do something, so you can get a phenomenal amount of throughput from this. With the growth in web applications, I expect Sun to do very well.

Re:

[TheRaven64]

Intel wisely abandoned it and dedicated more chip area to cache, prefetch, and the like.

I was talking to an engineer at IBM a year or so ago, and he claimed that SMT gave about a 30-50% performance increase at a cost of about a 5% increase in die size. This is a much better return on investment than cache increases (which tend to give a linear performance increase for a doubling in size), so I wouldn't be at all surprised if Intel had another go with the Core 3 or 4.

In a couple of generations, I wouldn't be surprised if the line between contexts and threads is blurred even more. I would im

Re:

[MemoryDragon]

Sorry to say that but Sun not really has been in trouble, especially not nowadays, with a good product lineup, most of their losses basically were stock devaluations which do not resembly anything regarding Suns profitability, from 2000 on if you count out the stock devaluations Sun hat a rather solid earning with only a handful of quarters of losses (some of them caused by buying new companies), not a huge performer, but it is defintely not in trouble and its bank account has definitely more cash than in 2

Re:

[julesh]

"Their chips are no longer the powerhouses they once were,"

Huh? I mean, is the statement I've seen that "at the time of its release [slightly over a year ago] the UltraSPARC T1 [was arguably] the world's most powerful general-purpose commercial server processor, when considering multithreaded commercial workloads" just somebody's imagination? Or are you saying that because they haven't released a new model for a year now they're slipping behind? 'Cause they have a new core scheduled for release some time

Re:T2000 gentoo stage3 build, was Re:Sun needs thi

[Wdomburg]

I'd say the Niagra line of SPARC chips do quite well for server-oriented tasks.

"Server-oriented" is entirely too broad a category to recommend a single architecture. An OLTP server has different needs from a file server has different needs from a database server, etc, etc.

sun.com

[Daemonstar]

Of all the links posted in the summary, there's no link to the webcast on Sun's site [sun.com] about the story (01/22/07 @ 10:00 PST, Realplayer 10 required). :P

Sun is not abandoning AMD. Sun is adding Intel.

[gp310ad]

Which makes sense. When there are two competitive players whose product features and performance keep passing each other, why not give the customer a choice and at the same time exploit that competition to improve ones own position...

Re:

[foxtrot]

I dunno. It seems to me that building hardware around three different CPU architectures is an awful lot of duplication of effort-- while AMD and Intel are both "x86" and run the same software, it's not like they're pin-compatible. And Sun's not exactly buying Intel 965 or nForce chipsets...

I like the move, don't get me wrong-- anything that gives me a choice is probably a good thing from where I'm sitting. But I'm not sure it's a wise thing for Sun...

-F

Re:

[Wdomburg]

And Sun's not exactly buying Intel 965 or nForce chipsets...

Um, yes, they are. The nForce is exactly what's inside the current Sun Fire chipsets.

They're not likely to use the Intel 965, but the 5000 series chipsets are far from unlikely.

Questionable

[tomstdenis]

The Core 2 Duo has an awesome ALU, and it is definitely low power.

But they still suck for NUMA. Unless Sun is building desktops I don't see the point of the move until Intel starts rolling out CSI [which by that time AMD will be 65nm working on 45nm parts...].

For the desktop, hands down the Core 2 Duo is the winner. These things are just amazing. Even when overclocked the thing is so cold that the CPU fan turns off and the BIOS warns me (annoying... so I turned the warning off). In terms of IPC it matches the AMD offerings fairly well.

Tom

Re:Questionable

[Anonymous Coward]

AMD reached 65nm last quarter. Intel's already moved to 45nm, though, but AMD's trying to catch up. Still, one of Intel's big strengths is its silicon fabbing capability; some have said Intel isn't so much in the business of designing chips as building fabs.

Re:

[RzUpAnmsCwrds]

In terms of IPC it matches the AMD offerings fairly well.

Actually, the C2D has considerably better (~20%) IPC than the Athlon 64. K8L is supposed to close this gap, but it's not coming until (at least) the second half of 2007.

which by that time AMD will be 65nm working on 45nm parts

FYI, AMD is already shipping 65nm parts for revenue, and is already working on its 45nm process.

Re:

[IYagami]

Yes, NUMA is not avalaible in Intel Core processors.

But Intel has processors with four cores avalaible. 2P motherboards with 4Core processors are cheaper than 4P motherboards with 2Core processors.

You can find a review with more information at:

http://anandtech.com/IT/showdoc.aspx?i=2897 [anandtech.com]

and in:

http://realworldtech.com/page.cfm?ArticleID=RWT111 406114244 [realworldtech.com]

The conclusion is (more or less): yes, the scalability of Intel Core Processors is worse than AMD Opteron Processors. However, the price/performance ratio of

Re:

[Saffaya]

My overclocked Opteron fan also regularly stops.
Takes a while at boot-up before it feels the need to actually spin.

I am sure the Core duo praise is justified. Just wanted to say it isn't exclusive to it.
 

Re:

[tomstdenis]

Must be 65nm parts. I have two 885 Opterons [90nm] and while they're not super hot, they do run warm [~40-45C] and can peak ~50C at full load.

That said, the Sempron 2GHz I bought my parents runs wickedly stupidly cold. Basically like +2C over ambient at idle. It's crazy.

Yeah, technology!

Tom

Re:

[Junks Jerzey]

The Core 2 Duo has an awesome ALU, and it is definitely low power.

That's "low power" compared to other high power consumption CPUs, but hardly low power from an absolute standpoint (compare it to chips from ARM or just about any microcontroller, for example; it's out of their league by a factor of 10).

Re:

[tomstdenis]

Yeah, but clock for clock an Intel or AMD x86 processor can mop that ARM processor. ARM/AVR32/PPC/MIPS are good for when you don't need a lot of power, but flexibility [in the ISA] counts.

And when I say "low power" I mean compared to other desktop/server processors. As in, the thing runs so damn cold that even under load, overclocked by a GHz, it's still well within the temperature range.

Try overclocking an ARM by 41%, achieve the same MIPS as the Core2Duo and come back to me with a "it's low power!" rant

Re:

[RAMMS+EIN]

``The Core 2 Duo has an awesome ALU, and it is definitely low power.''

Depends. When running idle, even a 90nm Athlon (FX or 64 or whatever the normal one is these days) draws less than the one from Intel. I'd expect the 65nm AMDs to draw less; perhaps even when compared at full speed, but I haven't seen any numbers yet. At any rate, for the 90nm AMD vs 65nm Intel comparison, it depends on how much work your CPU does during the day. Mine, especially in the servers, are mostly idle, which means AMD wins.

Re:

[tomstdenis]

Try this: run two memory intense applications on an SMP box. Tell me how great that shared FSB works out for you.

Oh yeah, and DIAF.

Sun Benefits

[dupup]

FWIW, it seems to me that Sun agreed to launch a Xeon-based server line in exchange for Intel pushing Solaris x86, rather than the other way 'round. In other words, the emphasis, from Sun's point of view, is that Intel will advocate in favor of Solaris, for which Sun offers Intel some bidness.

IAASE (I Am A Sun Employee), BTW.

Re:

[DrSkwid]

No it was single cpu before anything other, but yes it was AMD then Intel.

Re:

[Agelmar]

Except that this article isn't about sun using Itanium, it's about sun using Intel's Xeon x86 line. Itanium (IA-64) is a totally different (and now mostly defunct) line.

Re:

[RAMMS+EIN]

``if they still want to make servers and somehow remain different from Dell.''

Ye gods no! There's a world of difference between Sun and Dell. With Dell, you think you're buying a PC, but you're not; you're buying a computer with the quality of a PC, but without the flexibility and compatibility (missing drive cradles, anyone?). With Sun, you think you're buying a Real Computer, and you are. Ever have a component break within warranty? Dell will happily send you a replacement part, but if it breaks every mon

Re:

[itlurksbeneath]

Apples rot in my refrigerator too. In fact, if left anywhere for any appreciable length of time, they will rot. What's your point?

Re:

[zootm]

Apples rot in my refrigerator too.

I'll be sure to remember that when I need a catchy gag the next time that Apple has something in common with Refridgerator Microsystems.

Re:

[Lars T.]

...but if done right, Sun dried Apples make for a tasty snack...

Thus Sun should use the Itanic for a better drying experience.

Re:

[multipartmixed]

Nope.