Barcelona hint via Cray warning

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
The OP simply noted that there's more evidence implying a delayed barcelona ramp-up.

All this bickering benefits no one.

Good find Periander, and thanks.
 
I would like to defend Baron on one point though... pretty much everything AMD has been saying jives. 3 months ago AMD stated that this year would not see volume shipments of Barcelona... so this news story should be of no surprise to anyone (IF it's alluding to Barcelona).

However, what Baron refuses to acknowledge is by the time Barcelona goes to volume shipping it'll be too late; Penryn will already be available... and at this stage it's doubtful Barcelona will be able to fend off Penryn.
Click to expand...


I didn't mention anything about competition. And Penryn will have to have the Quad FSB which will make it a more expensive platform. You can see that C2Q scales for shit with the FSB.

Anyway, please don't defend me. It's kind of weird. You can agree if you want. Also, please don't put words in my mouth.
 
The OP simply noted that there's more evidence implying a delayed barcelona ramp-up.

All this bickering benefits no one.

Good find Periander, and thanks.
Click to expand...


There's no bickering. People are just offering their reasoning for whether they believe there will be a slow ramp or not. I will see if I can find the article where Ruiz says exactly that.
 
AMD said repeatedly Barcelona will be around but just like it took several quarters for C2D to become a large portion of shipments the same thing will happen here.
Click to expand...

Several? Baron, C2D hasn't even been out for 4 quaters, much less seven. But regardless of your grammatical error, C2D only took 2 quarters to ramp up.


Several can't get to seven. Several is just a plural descriptor of amount. I could have said a few or a couple and it would have meant the same thing.

I wish I did have info as to the mix at Fab 36. That would make it easier to see if they will have volume in Q3. Also, it depends on where Chartered is with 65nm. I always thought they should put Turion/Brisbane in Chartered, leave older Turions and Optern at Fab 30 and reserve most of Fab 36 for Barcelona derivatives.

But at any rate there will be more info tomorrow and the rest of the week. I have yet to hear mobo makers complain that they can't get chips for demos and testing.
 
I would like to defend Baron on one point though... pretty much everything AMD has been saying jives. 3 months ago AMD stated that this year would not see volume shipments of Barcelona... so this news story should be of no surprise to anyone (IF it's alluding to Barcelona).

However, what Baron refuses to acknowledge is by the time Barcelona goes to volume shipping it'll be too late; Penryn will already be available... and at this stage it's doubtful Barcelona will be able to fend off Penryn.
Click to expand...


I didn't mention anything about competition. And Penryn will have to have the Quad FSB which will make it a more expensive platform. You can see that C2Q scales for **** with the FSB.

Anyway, please don't defend me. It's kind of weird. You can agree if you want. Also, please don't put words in my mouth.

How is what I said putting words into your mouth??? I'm stating an opinion, not fact. I would have quoted you otherwise. What you're not getting is that Barcelona is already irrelevant! It doesn't matter if it kicks arse because Penryn will match it by the time AMD is in volume shipping.

Barcelona/Phenom=R600

It's over buddy.

You can quote me on that...
 
I would like to defend Baron on one point though... pretty much everything AMD has been saying jives. 3 months ago AMD stated that this year would not see volume shipments of Barcelona... so this news story should be of no surprise to anyone (IF it's alluding to Barcelona).

However, what Baron refuses to acknowledge is by the time Barcelona goes to volume shipping it'll be too late; Penryn will already be available... and at this stage it's doubtful Barcelona will be able to fend off Penryn.
Click to expand...


I didn't mention anything about competition. And Penryn will have to have the Quad FSB which will make it a more expensive platform. You can see that C2Q scales for **** with the FSB.

Anyway, please don't defend me. It's kind of weird. You can agree if you want. Also, please don't put words in my mouth.

How is what I said putting words into your mouth??? I'm stating an opinion, not fact. I would have quoted you otherwise. What you're not getting is that Barcelona is already irrelevant! It doesn't matter if it kicks arse because Penryn will match it by the time AMD is in volume shipping.

Barcelona/Phenom=R600

It's over buddy.

You can quote me on that...


Saying I refuse to acknowledge something is putting words in my mouth. This post and my response had nothing to do with whose chip will be faster. I would say we'll see, but I wouldn't say I care either way.

I want it to defeat both so you can all shut up. Day after day of these posts is crazy. We need a duopoly. Embrace before it embraces you.
 
http://www.theregister.co.uk/2007/06/04/barcelona_delays_cray/


Cray had been enjoying a recent return to profitability. CEO Pete Ungaro noted that he was "disappointed" with the delays that put both growth and profitability "in jeopardy."
 
We need a duopoly. Embrace before it embraces you.
Click to expand...

I don't think anyone would argue against that. But "wishing upon a star" is a bit naiive.

As for the couple, several, etc you know it's B.S. Instead of
"a couple" to describe the 2 quarters it took for Intel to make wide availability of C2D, you used several. Once again, you're putting a spin on your words.

I don't care what anyone says. Couple does not equal several. And a few doesn't equal several. My momma taught me this:

Couple: 2
Few: 3-5
Several: 6-~
 
Anyone remember when Quad FX aka 4x4 launched on November 30th? Oh, wait, that's when AMD said it launched but no one could even get them until the next year!

Oh how the seasons have changed! It amazing how long its taking them to get product off the ground, whether it be CPU or GPU. If Intel stays with it's "Tick Tock" strategy this might stay in Intel's favor for quite a while.
 
"Fend off" means preventing Intel from taking additional Enterprise marketshare. If Penryn is just as good as Barcelona, AMD loses. Many OEM's have switched back to Intel and I doubt they're going to switch their volume 1 & 2U servers to Barcelona if Penryn is just as good.

Everyone likes to tout performance per watt, but I can tell you that the majority of companies buying care very little about that. Performance per watt doesn't really come into play until you hit the larger server farms and datacenters (google, MSN, akami, yahoo, etc) Are they big customers? Sure. Do they make up the majority of the enterprise market? No. The rest care about performance and price... plain and simple. I'm not seeing how AMD is going to compete profitably in those sectors.
Click to expand...

While server farms need per watt performance, I'd agree a lot of customers will look simply at performance and/or performance/price ratio. Here's what I'm thinking: even in a virtualized server (which uses a cpu heavily), you'd still need quite a performance advantage margin (more than 10% and perhaps more than 15% for example) to be a meaningful difference if the other variable are roughly the same, as regards switching brands I bet. BUT....except for virtualized servers and scientific computing and other tasks that routinely peg a processor at 100% (or over 90%) for extended time periods, the ratios are more important than simple performance. And since overall operating cost is really the bottom line for servers, and also important for any business also that buys a lot of units, then for very large areas of the market, AMD would be an attractive choice if their prices are correct (and they'll simply be set by the market). This is true even if their margins are less than Intel's due to manufacturing. Therefore if AMd has reasonably competitive performance (not necessarily better), it can be expected to be competitive for a broad array of applications, with a sufficiently large pool of interested customers.
 
"Fend off" means preventing Intel from taking additional Enterprise marketshare. If Penryn is just as good as Barcelona, AMD loses. Many OEM's have switched back to Intel and I doubt they're going to switch their volume 1 & 2U servers to Barcelona if Penryn is just as good.

Everyone likes to tout performance per watt, but I can tell you that the majority of companies buying care very little about that. Performance per watt doesn't really come into play until you hit the larger server farms and datacenters (google, MSN, akami, yahoo, etc) Are they big customers? Sure. Do they make up the majority of the enterprise market? No. The rest care about performance and price... plain and simple. I'm not seeing how AMD is going to compete profitably in those sectors.
Click to expand...

While server farms need per watt performance, I'd agree a lot of customers will look simply at performance and/or performance/price ratio. Here's what I'm thinking: even in a virtualized server (which uses a cpu heavily), you'd still need quite a performance advantage margin (more than 10% and perhaps more than 15% for example) to be a meaningful difference if the other variable are roughly the same, as regards switching brands I bet. BUT....except for virtualized servers and scientific computing and other tasks that routinely peg a processor at 100% (or over 90%) for extended time periods, the ratios are more important than simple performance. And since overall operating cost is really the bottom line for servers, and also important for any business also that buys a lot of units, then for very large areas of the market, AMD would be an attractive choice if their prices are correct (and they'll simply be set by the market). This is true even if their margins are less than Intel's due to manufacturing. Therefore if AMd has reasonably competitive performance (not necessarily better), it can be expected to be competitive for a broad array of applications, with a sufficiently large pool of interested customers.

Well if youve got actual numbers showing either Penryns power consumption or K10s power consumption, why dont you go ahead and share with everyone, since otherwise no one knows, and your argument could just as easily apply to Intel....if Penryn uses less power than K10.
 
Hey, I'm not saying there's something wrong with mighty Intel! I'm saying it looks like AMD has a good chance to remain viable. And no, I don't know what the real world average power draws of new server platforms circa 2008 will be! :)
 
Hey, I'm not saying there's something wrong with mighty Intel! I'm saying it looks like AMD has a good chance to remain viable. And no, I don't know what the real world average power draws of new server platforms circa 2008 will be! :)
Click to expand...

I didnt say you comented on either Intel or AMD in anyway. What I pointed out, in a round about way, was that unless you have some secret information, noone knows which uses less power at idle or load; a penryn or a K10. As such, your argument applies to which ever one uses less.
 
If there is a significant difference on average power draw, that would be an nice advantage to help win those bigger cusomers, yes.
 
AMD would be an attractive choice if their prices are correct (and they'll simply be set by the market).
Click to expand...

Exactly.

The biggest threat to AMD isn't Penryn or C2D, it's Intel aggresive pricing. If AMD can't make a decent margin on a CPU, it's wasting its time. And we all know Intel (right now) is out to squash AMD.

It'll be interesting to see if Ruiz sticks to his guns about not giving up market share. Though, even AMD's current tactic of aggressive pricing doesn't seem to be working.

AMD is in quite the conundrum.
 
If there is a significant difference on average power draw, that would be an nice advantage to help win those bigger cusomers, yes.
Click to expand...

And until someone starts producing actual numbers for both CPUs, not Intel said this or AMD said that, noone will know which uses less.
 
Q4 starts in October. Q3 is right now. Tey said mid to late Q3 for shipments to start. Dell, HP, Sun will get the first batches. The Cray volume is good but they are niche where one server will have a few thousand CPUs, but the real volume is in 2P/4P.
Click to expand...

Actually, Q3 starts in July. Right now, it's Q2. :wink:

D'oh. I must be working too hard. That's even better since AMD did say shipments in June and machines by Aug with some maybe debuting in July.Nice way of saying "i was wrong"....without actually having to say it... Damn, you dodged a bullet on that one. :roll:


Don't tell me....you're GOJDOs replacement. Comment on the topic or go elsewhere.
Yeah because GOJDO was the only person that could tell that you have a huge superiority complex and will never openly admit that you're wrong. :roll:
 
It doesn't work that way sometimes. In most corporate environments, corporate buyer drones typically have the final say so... and they usually don't have all the facts (such as the energy bill!). They see a low priced OEM offer from Dell or HP w/ great performance and buy it.

That's what I mean about power consumption not being such a big issue... the people who care don't matter. That's not universally true, but it's certainly been my experience.

*shrugs*
 
If there is a significant difference on average power draw, that would be an nice advantage to help win those bigger cusomers, yes.
Click to expand...

And until someone starts producing actual numbers for both CPUs, not Intel said this or AMD said that, noone will know which uses less.

Damned if we don't agree. Guess it must be true.
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom