Submitted by “Customers demanded it and now it’s here. Sun is challenging Dell to a Benchmark Brawl. Up for contention are the titles for the server side Java benchmark. Floating point intensive calculation. Power consumption. And the jewel in the crown, price-performance. If Dell accepts, benchmarks will be conducted by a neutral, third-party lab running agreed upon industry-standard benchmarks. Dell has until January 31, 2006 to accept the challenge. Testing [is] to be done using Sun Fire x4100 and X4200 servers and the corresponding Dell Xeon servers that are commercially shipping in volume.”
Interesting. Has anything like this happened before?
P.S.: Oh … and may the best `Sun’ win btw.
P.S.#2: Not that they would ever challenge Dell if they knew there was even a slight chance of losing.
Sun’s stack would likely win such a ‘brawl’ hands down.
Regardless, Dell has no reason to submit to this challenge, and little to gain.
The decision a company makes on which product to buy is based on service contract, kickbacks from the vendor, and other meaningless criteria. This decision is seldom based on the merits of a vendor’s technology.
I love Sun’s products (soft & hard), and would love to see them dominate again. But, this sort of stunt will do little to help them.
Their marketing has been re-shaping Sun’s image with edgy advertising, but they are missing their audience. If their market was direct consumer sales (like Apple) this sort of thing would work. Corporate, what’s-in-it-for-me types won’t be impressed.
Edited 2006-01-25 21:50
i concur. there’s no incentive to be publicly smacked down for Dell. the Sun systems are most likely superior just in cpu performance alone using amd’s chips, whereas Dell is tied to Intel’s xeons. Its also clearly on Sun’s terms, w/ using Java etc.
Sun will probably end up with the bully-type bragging rights, such as “we’re clearly better, they wouldn’t even accept our invitation to a challenge!”.
Their marketing has been re-shaping Sun’s image with edgy advertising, but they are missing their audience. If their market was direct consumer sales (like Apple) this sort of thing would work. Corporate, what’s-in-it-for-me types won’t be impressed.
I doubt it – the fact is, corporate people want the best bang for the buck and most importantly, they want to integrate it into their business without any major issues occuring.
These corporates are currently purchase Dell hardware, SUN is simply saying that if you’re purchasing Dell on a price/performance perspective, you are getting ripped off and that SUN has a better product here.
What SUN need to do, which would be better is push the idea of ‘get everything under the one roof’ approach – get the lot from one vendor, and get a massive discount! that is what will attract corporates who are fed up with having to stuff around trying to marry different things together into something that semi-works – if SUN can offer a complete out of the box solution, you would find that more people would be attracted.
With this challenge, Sun has managed to point out one thing: that Dell is at a disadvantage by sticking with Intel. The situation is now put in clear light and is dificult to overlook.
So whether or not it will help Sun may be up for debate (my personal, gut feeling is, it could), but, boy oh boy does AMD love it!
Dell would probably be fine to just ride it out until Woodcrest is released. Unless Woodcrest is a lemon, they’ve already taken the bulk of the pain of being an Intel shop.
Agreed, Get a Life. Woodcrest is a different beast entirely, and it’ll be interesting to see if SUN still wants to challenge them when Woodcrest is out. 🙂
Agreed, Get a Life. Woodcrest is a different beast entirely, and it’ll be interesting to see if SUN still wants to challenge them when Woodcrest is out. 🙂
I may be one of “nay” sayers, but…
Same thing was said about any Intel CPU, not even a month ago about Yonah. And….? Hot-Water reinvention without any single benefit except being dualcore 2-3 months before Turion dualcore is out? Taking dualcore in equation Yonah is actualy dissapointent.
Yonah is not a server processor. It’s not competing with the Xeon nor the Opteron. Yonah is a dual-core Pentium M with VT, SSE3, and further improvements in FP performance and power conservation. What did you want or expect from it that makes it a disappointment? What does it have to do with Woodcrest?
1. This is why I mentioned Turion not Opteron.
2. Performance sucks for dualcore, 32-bit, overpriced, overhyped
3. Historicaly, Intel quality is going down, Intel BS is going up. Intel was probably just overtaken with PR who needed extra space and take over R&D building. Parent before me was overhyping Woodcrest, so it is only cool when you do that?
Neither the Turion nor Yonah have anything to do with the future performance of Woodcrest, which was the topic in question. Woodcrest is the server-oriented version of Intel’s next line of processors in the Merom and Conroe family. You know, the one that isn’t in production.
Here’s how your train of thought works: Yonah isn’t better enough in your opinion, therefore Woodcrest will be bad. If you think that sounds reasonable, then good for you.
Yonah was never slated to support EM64T, so that’s hardly disappointing. Its multithreaded throughput is pretty good. It would be better if it had a faster bus and an on-die memory controller. It could be better in many ways, but it’s really just the center of Intel’s Napa platform and the continuation of their aging mobile line. It improved the SSE implementation of Dothan, which was really the only weakness the Pentium M had with regard to the Turion other than cost. What did you expect from Yonah that made it disappointing? What was promised that wasn’t delivered?
Intel’s mobile line has been dominating the market. Their products in their mobile line have been Intel’s only strong suit since at least Prescott. Attacking it seems almost silly, but doing so in a discussion about servers is just out-right misplaced. You want to complain about Intel’s products and have it make sense? Well you can complain about their current Xeons, with their mediocre performance and high power requirements. And now that Intel is finally retiring the Pentium 4 line and moving on, you instead use the above reasoning to sandbag the platform.
With your reasoning, AMD’s K7 would have been a piece of crap, because its K6 line was only ever competitive in integer performance. The K8 would have been a regular furnace on its way to reaching the P4, because before moving to SOI AMD’s processors were becoming rather toasty. And let us not forget VIA’s early chipsets. These are technology companies, not abusive husbands. Woodcrest might not be miraculous, but unless Intel has secretly learned nothing from its mobile line over the years, it won’t be the flame-throwing embarrassment the current Xeon is.
Well, I was disappointed with Yonah’s fp performance. I was expecting a major boost there and it just didn’t happen. The overall impression was pretty good though.
As far as Woodcrest, I don’t think anyone except Intel engineers really know how well its going to do. My bet would be that it is a strong performer, given that it was developed at a time when the Opteron is kicking Intel’s ***. But who knows.
I’ve hearing this old song since the first K7 was release:
“Once the new P3 Tualatin core is released”
“Once the new P4 northwood core is released”
“Once the new P4 Prescott core is released” (OK, this /was/ the best of the lot)
“Once the new Yonah core is released”
And now:
“Once the new Woodcrest core is released”
Truth is, Woodcrest with all of it’s new core glory and it’s dual FSB backbone, will still use the age-old shared bus design; will still lack any time of fast glue-less inter-connect transport (now that CSI is officially dead); and even if it does outperform AMD on a core to core basis, (And considering Intel’s past record, I wouldn’t bet the farm on Intel delivering on schedule and on spec) on >2 socket configuration (AKA server configuration) the Opteron will continue to outperform the Woodcrest, even if Intel stuff each core with 10MB of L3 cache.
Call me an AMD fanboy, but I’m in the market for 2/4 socket servers *now*; why should I care if Intel may/may-not release a new uber-CPU in half-a-year’s time if AMD gives me what I need *now*? (Speaking from experience, both the IBM e325/6 and the HP DL385 are fine machines that stamp all over their Xeon [e345, DL380] counter-parts)
More to the subject, when AMD offered to dual their dual core Opteron/Athlon64s, Intel caved in.
Dell will do the same.
For now, neither of them have any reason to get their ass publicly ass-whipped. They’ll cave in and wait for Intel to deliver something competitive.
G.
What, are you seriously contending that Intel has not been performance competitive with AMD at any point since the release of the K7? Nice. Yes, that would definitely earn you an “AMD fanboy” label.
Dell obviously receives a financial incentive to remain an Intel-only shop. Dell has withstood the large majority of any harm it would see from Intel’s Xeon performance. That is, they’ve already paid the majority of the cost and losing the financial incentive by adopting Opterons because they’ve been “shown” to have a disadvantage by staying exclusively with Intel would be pointless. Woodcrest will probably bring Intel’s offerings back to being competitive with the Opteron. That’s all that Dell needs. Well that and Intel’s fabrication capacity.
What, are you seriously contending that Intel has not been performance competitive with AMD at any point since the release of the K7? Nice. Yes, that would definitely earn you an “AMD fanboy” label.
Nope. I /did/ say the every time Intel solution got under-performed, all major sites began the usual “you just wait for the next Intel <insert_name> core”, like dead-man’s reflex.
Dell obviously receives a financial incentive to remain an Intel-only shop. Dell has withstood the large majority of any harm it would see from Intel’s Xeon performance. That is, they’ve already paid the majority of the cost and losing the financial incentive by adopting Opterons because they’ve been “shown” to have a disadvantage by staying exclusively with Intel would be pointless.
I agree.
I venture to guess that Dell will use AMD processors if, and only if, Intel becomes too cozy with Apple.
Given the relatively low quality of Dell servers (at least by my own experience) we only use Dell if our the client demands Dell. (Which is beyond rare.)
At least to me, Dell using Opterons means little.
Woodcrest will probably bring Intel’s offerings back to being competitive with the Opteron. That’s all that Dell needs. Well that and Intel’s fabrication capacity.
“Probably” is the right word in the context.
Intel has made severe design mistakes in all recent cores (see my previous post) that have yet to be fixed.
/If/ Intel delivers on time and /If/ Intel delivers on spec (I’m still waiting for the 4Ghz P4 that was supposed to be released a year ago…), the Woodcrest /should/ be competitive.
G.
Nope. I /did/ say the every time Intel solution got under-performed, all major sites began the usual “you just wait for the next Intel <insert_name> core”, like dead-man’s reflex.
That’s fine. There were several times along that road that the P4 led, or reached performance parity with AMD, until the K8. Where they often lost consistently while they could still remain competitive was in performance/dollar.
I’m going to assume that we read different sources of information. While I often see forthcoming competing technologies mentioned in articles about either company’s products, I can’t recall any off of the top of my head that takes the tone of “you just wait!” Either I should read more, or you should read less, stupid journalism.
I venture to guess that Dell will use AMD processors if, and only if, Intel becomes too cozy with Apple.
I think Dell will only use AMD processors if Intel’s forthcoming Conroe/Woodcrest are so blatantly bad that Dell will make more money diversifying their offerings than Intel saves them by not selling Intel products. That is unless Dell could be especially harmful to Intel in AMD’s anti-trust case, in which case Dell could probably sell Opterons without worrying. I don’t think that Apple ships enough units to be a big threat to Dell with respect to Intel.
Intel has made severe design mistakes in all recent cores (see my previous post) that have yet to be fixed.
/If/ Intel delivers on time and /If/ Intel delivers on spec (I’m still waiting for the 4Ghz P4 that was supposed to be released a year ago…), the Woodcrest /should/ be competitive.
Intel has made a lot of compromises with its designs to move aging architectures well outside of their planned design space. They’ve done a respectable job of converting the Pentium 3 into the Pentium M. They’ve done a reasonably decent job of pushing NetBurst into all sorts of places they never planned. They were dragged into a chipset with support for DDR. They were dragged into adding EM64T. They were dragged into making dual-core processors. Their designs with regard to all of these were clearly compromised. But those pesky laws of nature changed their plans abruptly for having 5+GHz Pentium 4s and they’ve simply been riding it into the ground.
You can look at it this way, if Woodcrest isn’t competitive with the Opteron then AMD’s going to have a lot more good quarters at Intel’s expense, because they’re going to look like clowns.
I’m going to assume that we read different sources of information. While I often see forthcoming competing technologies mentioned in articles about either company’s products, I can’t recall any off of the top of my head that takes the tone of “you just wait!” Either I should read more, or you should read less, stupid journalism.
It was a general remark. Even Anadtech and Inq (which I usually enjoy) tend to be over-optimistic when it comes to Intel. Maybe it’s because people tend to assume that FAB capacity & Money == quality and production yields; In real like things are far more complicated.
I think Dell will only use AMD processors if Intel’s forthcoming Conroe/Woodcrest are so blatantly bad that Dell will make more money diversifying their offerings than Intel saves them by not selling Intel products. That is unless Dell could be especially harmful to Intel in AMD’s anti-trust case, in which case Dell could probably sell Opterons without worrying. I don’t think that Apple ships enough units to be a big threat to Dell with respect to Intel.
Don’t forget that Dell mostly lives in the low-end server segment. I doubt that they are loosing to much cache for not having the Opteron aboard. I’d assume that Intel’s marketing money easily offsets this loss.
However, now that Intel spends marketing money on apple, I doubt that they can offer the same deals to Dell, increasing the size of Dell lack-of-Opteron problem.
Intel has made a lot of compromises with its designs to move aging architectures well outside of their planned design space. They’ve done a respectable job of converting the Pentium 3 into the Pentium M.
I don’t doubt it. The PM and the subsequent Centrino are fine core by themselves, but they still carry the old basic PPro design. When it comes to the server space and FPU intensive applications (HPC), it still doesn’t cut it.
They were dragged into a chipset with support for DDR. They were dragged into adding EM64T. They were dragged into making dual-core processors. Their designs with regard to all of these were clearly compromised. But those pesky laws of nature changed their plans abruptly for having 5+GHz Pentium 4s and they’ve simply been riding it into the ground.
To be honest, at least in my view the Netburst architecture was flawed by design. When the first Tualatin P3 were released, they literally stumped all over their far-more-expensive P4 cousins. Intel should have reconsidered the Netburst move back then when it could change course without opening the door to AMD.
However, the Netburst offered fast and easy access to the magical >3Ghz PR party, even if their relative performance, clock to clock, were far slower then both the K7 and the P3/M.
You can look at it this way, if Woodcrest isn’t competitive with the Opteron then AMD’s going to have a lot more good quarters at Intel’s expense, because they’re going to look like clowns.
Intel has huge marketing war-chest and huge fabrication capacity. I don’t see they die soon. Even if the K8L tears the Woodcrest to pieces…
The Pentium M isn’t a server line. As for NetBurst, its intention was to sacrifice IPC for clockrate to maximize throughput. NetBurst had considerable successes in media applications and games. It really wasn’t until they couldn’t scale their clockrate as planned that they fell behind. As for Intel dying, no I don’t think that’s entirely likely. Intel is incredibly profitable and has production capacity far greater than AMD. On the consumer-side of things, Intel has guaranteed business.
AMD on the other hand was circling the drain and reached its production capacity. If Intel hadn’t taken NetBurst this far off course, they probably could have bankrupted AMD. If Intel flubs Woodcrest, AMD’s financial will continue to improve.
For anyone interested in blogs by Sun engineers involved in performance technologies, here’s a quickie “Sun performance blogger linkroll” that I just threw together:
http://blogs.sun.com/roller/page/eric_boutilier?entry=solaris_and_t…
Eric
This reminds me of the challenge AMD put forth to Intel not long ago. And Dell will try to ignore it for the same reasons.
This is a very entertaining PR stunt by Sun. I approve. It would be so fun if more competitors had challenge/smack-downs like this.
What kind of hardware is Sun pitching? Its Niagara processors or AMD Opterons? Dell is easy to figure out because they would be going exlusively with Xeons. If that is the case, AMD will have quite an upperhand except in those tests where the 1066 mhz FSB and cache are very important.