People prefer Opteron over Woodcrest
Woodcrest is for ultra low end uncertified use only, Rackable report indicates. Customers want proven Opteron solutions, instead of flaky Woodcrest.
Rackable always offered both Xeon and Opteron servers.Before the launch of Woodcrest, Rackable bet a lot of people will switch, and Rackable CEO pumped Woodcrest at IDF. Guess what? Companies don't buy the crap, and Rackable posted a loss.
"There have been more barriers to change than we had initially anticipated." Rackable CEO said.
What he doesn't understand is, software compatibility is the deciding factor. People who have standardized on Opteron will not switch to Xeon with ancient FSB. The only thing good about Woodcrest is that it's faster than Netburst. But that's not good enough for enterprises which need dependable computing power.
On performance, Intel's Woodcrest is only good for workstation applications. For server applications which need a lot of bandwidth, Intel's FSB is a major bottleneck. On floating point, Woodcrest is about 30% slower than Opteron.
For Rackable to make a profit, it should stop doing Xeons. It's simply stupid to spend 50% of development and marketing effort on about 10% of its sales.
posted by Sharikou, Ph. D. @ 10/30/2006 04:17:00 PM
27 Comments:
How about performance? Customers don't want that now because AMD doesn't have it?
Your posts against Intel seems to be getting shorter and shorter and not getting repetetive. Why is that? Are you running out of arguments against Intel now that it has started outperforming AMD on all levels?
"Opetron is a proven solution". Is this all it can show for? Prove that it works and customers will be happy. That is such a high standard. I wonder how Intel can ever beat that. ROFL!
Rackable states that 84% of its sells are AMD, and its Opteron customers are sticking with AMD, then ad the comment “go figure”. (What an ass-hole)
Aren’t reporter/writers suppose to report the news, not make it.
Pssst. Rackable LOST money....not a good story.
You are pathetic. Quit while you're behind.
While this is relevant and shows how much faith the higher end and mid-level users have in Intel based systems, this is only one company's report, and not the entire industry. Let's just wait till we have numbers on volume share before we make any broad statements.
Rackable states that 84% of its sells are AMD, and its Opteron customers are sticking with AMD, then ad the comment “go figure”. (What an ass-hole)
Aren’t reporter/writers suppose to report the news, not make it.
Sharikou said...
"Woodcrest is for ultra low end only, Rackable says."
Why do you insist on lying, you must be aware that people can actually read the links you quote right?
No where in the article does it say this, I did see where Intel took back 6% of sales, granted not alot but it is still 6% that AMD isn't selling.
Since you choose to quote Rackable, lets do it again...
-[Link]-
""It is clear to us now that Intel has leadership in performance and in performance per watt," Barton said..."
or...
Barton suggested that Xeon server shipments would likely nudge just a few percentage points higher in the coming months.
Which it did, but still no mention of Woodcrest being ultra lowend.
As it turns out, large customers that have already certified their Opteron-based systems are staying with the products rather than moving over to new Xeon-based kit.
Opteron = high end enterprise proven.
Woodcrest = low end flaky.
First of all, the issue of certification of Woodcrest is not news. We were talking about this when Conroe was first previewed. It simply takes time to certify new systems. This won't be an issue soon.
Secondly, it is very naiive to think that a 10-20% improvement in speed is a determining factor in server purchases.
Companies look for support from their favorite vendors which is something that hurt AMD last year but has turned around this year with HP, DEll, IBM, and Sun all expanding their AMD server offerings. Remember, X2 came out last year so it has already been certified as have the P4 based Xeons and Itanium.
Remember too that with socket F the dual core Opteron is upgradable to quad core on all systems, which effectively removes any current advantage that Intel has. Intel will have quad core 4-way C2D systems when Caneland is released but AMD will have quad core 4-way before then. Simarly, Intel will have quad 2-way a bit ahead of AMD. By third quarter of 2007 there won't be much different in performance or power consumption between AMD and Intel. Although I realize that some are still hoping for a huge jump for Intel from 45nm. We'll see.
I wouldn't describe Woodcrest as low end. It's a DP processor so it should cost more than single processors like Opteron 1xxx and Xeon. I imagine it goes for more than P4 Xeon DP and about the same as Opteron 2xxx. It should go for less than P4 Xeon MP and Opteron 8xxx.
I think Scientia is a couple of notches above sharikou in the intelligence front. He's definately an Amd fanboy like sharikou, but unlike sharikou he know how to present an opinion without comming across as a cartoon caricature, which sharikou has now become.
Although this is interesting and shows that speed is not the only relevant factor for company cpu purchases I think you went too far when you claimed Rackable said one thing when actually they obviously did not and it was actually you.
Saying things like that can get you in trouble with the lawyers and for heavens sake don't give them chance to make even more money !
Are AMD quad cores coming sooner or is Intel just devaluing AMD's product? Please comment
AMD: Opteron CPUs now up to 47% cheaper
"Although this is interesting and shows that speed is not the only relevant factor for company cpu purchases"
Exactly.
AMD problem is quantity off products in OEM vendors.
For example the server I just bought is based on an Intel PD820. Why?
The other processors P4 630, P4 5xx and celeron are much worst. Between crap and crap I choose the better crap.
If the same server existed with an AMD Opteron inside would I go for the Opteron? Surely.
Was the processor speed a decision factor? Surely not or I wouldn’t buy it, especially if there are much better options like Woody and Opteron.
Well this just proves nobody wants woodmistake and everybody wants a opteron for its higher relibility and proven quality in the server market for years. Intel isn't so good at that. Performance is nothing without trust in your system and a longer life span.
AMD for many years have made systems that last year after year. AMD is 4P Intel only 2P. Nobody wants extra rack space. Thats why the companies will not waste time and money on a limited solution with a dead upgrade path.
Off topic but it would seem that AMD is only aiming for 35% growth in production next year over this.
-[Link]-
You Intel fanboys don't get it do you.... AMD/Sharikou WON and you LOST. Now go home.
Sharikou said...
"For Rackable to make a profit, it should stop doing Xeons. It's simply stupid to spend 50% of development and marketing effort on about 10% of its sales."
How do you know they spend 50% of development and marketing on Xeon?
If you think a company that sells 84% AMD is doing so then I must question your source.
So do you have a source?
How do you know they spend 50% of development and marketing on Xeon?
I think Rackable probably spent more than 50% to develop the Xeon products. Its Opteron line was pretty much stable, but it had to do frequent changes for Intel platform due to all sorts of changes all the time, chipset, memory...
sharikou wrote:
I think Rackable probably spent more than 50% to develop the Xeon products.
You also think Intel CPUs physically explode (1 per core!) thereby causing batteries to explode.
I'll take a verifiable source thanks. Your idle speculation has been proven to be the ramblings of a madman.
Sharikou also compares Intel to Nazis. He must of forgot to study history when he was going for his Ph .D. How can this clown compare the most evil organization of modern times to a computer manufacturer. Millions died because of the Nazis, you idiot, don't you ever belittle their memories again by comparing their struggle to a damb computer chip.
Anonymous said...
You Intel fanboys don't get it do you.... AMD/Sharikou WON and you LOST. Now go home.
7:59 AM, October 31, 2006
Sharikou, please shut up!
"The phenomenon that we are seeing in Q3 and Q4 is that there is a fair amount of stickiness (around Opteron)," Barton said.
This is absolutely not surprising. It has taken 3+ years for AMD to get a decent share in the server market. And now people are expecting Intel to being able to regain that lost market share in the couple of months preceding the release of K8L ?
This is just not going to happen. The inertia in the server market is so big that Woodcrest is just too late. Intel will *not* be able to regain any significant market share before the release of K8L.
Anonymous said...
7:59 AM, October 31, 2006
"Sharikou, please shut up!"
Some who is not Sharikou said...
GET YOUR OWN DAMN BLOG AND YOU SHUT YOUR PIE HOLE. I'm sick of reading you mealy mouth Intel zealots bitching in this blog. IF YOU DON'T LIKE THE BLOG, DON'T LET THE DOOR HIT YOU IN THE ASS ON THE WAY OUT! GET YOUR OWN BLOG AND BITCH ABOUT SHARIKOU THERE, THE REST OF US DON'T GIVE A SHIT ABOUT YOU OR HOW MUCH YOU HATE THE MAN!!!
C'mon Sharikou! Put a stop to this incessant, seemingly endless bullshit dewd! Some of us who read this blog and actually find it interesting and/or thought provoking are sick of having to wade through the Intel fanboys blog crapping to do so! Start doing a little better job of moderation for these idiots who do nothing but go off-topic and/or "thread crap", thanks in advance!
I'm not really an AMD fan. I've bought both Intel and AMD systems in the last year. In all honesty if someone is buying something in the Celeron/Sempron range there really is no difference between Intel and AMD. Intel currently has better battery life in mobile but AMD may catch up in 2007 with 65nm and better chipsets.
Kentsfield and 4X4 are very similar in that both are selling two dies for the price of one. This is why neither Intel nor AMD will be rushing to produce lots of these. I know that 1 Million Kentsfields might sound like a lot but this is 1 Million in by Q2 07. This means 1 Million chips in 18 for a company that will produce 54 Million chips. This is then less than 2%. 4X4 will be less than 1% of production for AMD.
Frankly, I am still uncertain what 4X4's market is. This seems like too much for just a game machine but would be pretty good for a low end workstation. Let' be honest, two quad processors and up to four graphic cards. No game exists that would tax this system nor even be likely to within the next two years. However, I could imagine workstation loads that that would make users very happy with a system like this. It isn't clear though if this hardware will be applicable to workstations or if AMD is trying to keep this separate from Opteron 2xxx. If it is intended for workstations then this changes things since this could easily pull back 2-way share in AMD's favor. However, it has been suggested that these boards wouldn't have room for HTX and won't use registered memory. 4X4 is much less important if it is just a high end game system.
My guess is that by 2009 Intel and AMD will look a lot more alike than they do now. However, several fundamental problems remain that threaten to bring processor progress to a halt. 64 bits is the practical limit so we won't see 128 bit Integer operations. 8 cores on one chip is not practical unless the memory bandwidth can be doubled. So, this rules out expanding cores and expanding to 128 bits. This doesn't leave a whole lot for continuing expansion. Desktop systems could go to 2-way but 4-way systems are much more expensive. Perhaps we will see 2-way move down to the desktop and 4-way begin to displace the upper 2-way market. SSE processing with a separate GPU is possible. The practical limit for SSE is about 64 parallel 64 bit calculations. There is still some room to expand on this if memory bandwidth can keep up. We really need to change directions though from DDR3 and DDR4. The increasing latency from these will begin to nullify any benefit in speed. If memory continues to drag we may see higher performing memory come in at a higher price level. This could include memory cells based on Z-RAM or the use of TTRAM or SRAM buffering. It could include different fanout topologies as I suggested with my HTDIMM article. This could include narrower buses as we saw with RDRAM and FBDIMM. With increased complexity we could even see memory that was linked to predictive signaling or where predictive buffering was used. These fundamental problems will be faced by both Intel and AMD. They may be too complex to solve without more cooperation.
BTW, in terms of changes in marketshare from what I've observed markets tend to be elastic for up to 2 quarters. If a company cannot satisfy demand within six months then they will lose sales and marketshare. If the market is truly demanding higher SSE performance then AMD could lose share before Q2 07. However, if it's a different issue like lower power consumption or exandability then AMD should remain competitive.
While this is relevant and shows how much faith the higher end and mid-level users have in Intel based systems, this is only one company's report, and not the entire industry. Let's just wait till we have numbers on volume share before we make any broad statements.
This is very relevant because Rackable sells to MS, Yahoo, Google, and others.
I would say that some of these folks have been privy to Barcelona info and realize that when it comes out AMD will retake the lead in perf.
Just adding two cores should almost catch them up with Clovertown because adding a core adds bandwidth which servers so dearly love.
Hey Sharikou, you are going to love this.Intel fanboy JumpingJack at THG has been caught spreading FUD about AMD!See the thread "INQ says rev. A1 K8L ES Demo close" in the Hardware CPU section.
http://forumz.tomshardware.com/hardware/INQ-rev-A1-K8L-ES-Demo-close-ftopict206444.html
Post a Comment
<< Home