Barcelona and R600 to be a double release.

visaris

Gawd
Joined
Apr 14, 2005
Messages
842
February 28, 2007 07:03 PM
http://www.informationweek.com/news/showArticle.jhtml?articleID=197700288&subSection=Breaking+News
AMD also demonstrated working versions of its next-generation graphics chip the R600 to be released by the end of June.


1 March 2007 12:00 AEST
http://www.itnews.com.au/newsstory.aspx?CIaNID=46286&src=site-marq
CRN: When will Barcelona ship?

RICHARD: It's slated for introduction at the end of the second quarter and will be in the market in the third quarter.

Published Wednesday 28th February 2007 22:59 GMT
http://www.theregister.co.uk/2007/02/28/amd_690g_launch/
"R600 is doing very well and there's a reason we're going to launch it when we're going to launch it...I'll take the blame", Richard said.

AMD has decided to stretch the R600 delay into a dual release with barcelona... It's an interesting idea, if true.
 
I think the benches will be way too skewed, by both a new CPU, and a new GPU.

It will be very hard to determine what performance is gained by CPU, and what is gained by GPU, but it aught to be an impressive combination to say the least.
 
Well, you have to remember they are probably most interested in selling the platform. 'Twill be awesome non-none the less, and certainly explains the delays. I don't like it but it sure as heck makes scene from their point of view. They're not a video card company anymore.
 
I think the benches will be way too skewed, by both a new CPU, and a new GPU.

It will be very hard to determine what performance is gained by CPU, and what is gained by GPU, but it aught to be an impressive combination to say the least.
..until someone reviews it on a C2D or a Athlon X2, which will probably happen the first hour that reveiws are available.
 
AMD shows off Teraflop in a Box:

http://theinquirer.net/default.aspx?article=37937
Thursday 01 March 2007, 14:45
The single-system, Accelerated Computing computer runs Windows XP Professional on an AMD Opteron dual-core processor combined with two AMD R600 Stream processors. The box is capable of performing more than 1 trillion floating-point calculations per second using a general "multiply-add" (MADD) calculation, the firm boasted.

This source claims the CPU in the teraflop box was Barcelona:

http://www.internetnews.com/ent-news/article.php/3663121
March 1, 2007
The demonstration involved an early version of a forthcoming Barcelona CPU and two R600 graphics cards, which are massive floating point accelerators. Using a custom calculation program, the three chips ran in parallel to produce over a teraflop of computation, or 100 billion calculations per second. However, AMD admitted practical applications for the new technology are still considered to be years away.

Possible influence for simultaneous release?
 
My biggest complaint is that the terflop box doesnt demonstrate average performance, but instead a costom made calculation that was designed to run best on CTM.

The question becomes how often will this type of calculation be used in real world applications? I doubt it will be often.

What I'd like to see is a well rounded bench using industry standard applications, with a few synthetics thrown in to show peak performance. This type of testing hasnt been done since the K7 days. And now we have AMD coming out with meaningless bench results like this too.

I'm a big AMD fan myself, but I think this result is totally worthless. I want to see a well rounded bench test. I dont want to see two new products benched at the same time becouse that will undoubtedly screw the results.
 
Within the past week or so, i'd say AMD has definately done a good job of getting their word out, and making lots of people hold onto their money. They have been extremely vague, but judging by the mood im sensing on the forums here and elsewhere, it seems that they did a good job of getting at least a few people to hold off on C2D/8800 *despite* the delays of their products....

so far so good.....but *goddamn*, they better deliver in june, or tons of people will never forgive them!
 
http://www.internetnews.com/ent-news/article.php/3663121[/url]
March 1, 2007
It wasn't a Barcelona in the box.

AMD's own press release calls it an "AMD Opteron™ dual-core processor." That's the name of the current product. If it were Barcelona, they would call it that or at the very least, call it Next Generation AMD Opteron.
 
so far so good.....but *goddamn*, they better deliver in june, or tons of people will never forgive them!
QFT.

AMD's own press release calls it an "AMD Opteron™ dual-core processor." That's the name of the current product. If it were Barcelona, they would call it that or at the very least, call it Next GenerationAMD Opteron.
I doubt AMD will abandon the Opteron name any time soon. After all, they have been using Athlon for a number of product generations. While I can't claim to know for sure (and it doesn't really matter anyways), the second page I linked too seems confident that the K10 was used.
 
"R600 is doing very well and there's a reason we're going to launch it when we're going to launch it...I'll take the blame", Richard said.

I smell marketing BS. AMD should be fully aware that the longer they push back the release the less chance they have on the market. I am still giving them the benefit of the doubt - though I am starting to wonder why.
 
Eh, as a Linux user, ATI's cards (R600) don't make much sense anyways (poor drivers), so I suppose I'm not really passionate about the issue.
 
I don't even think AMD should release roadmaps anymore. They obviously can't even keep to any type of timetable, so why bother?
 
I didn't think that ATI/AMD could foul things up worse than they did last time (7800 series out for 6 months before the X1800 came to market and when it did come to market it was only marginally better), but it looks like they did it again with the R600, except this time even worse than the last.
 
Eh, as a Linux user, ATI's cards (R600) don't make much sense anyways (poor drivers), so I suppose I'm not really passionate about the issue.

lol, why would you buy a R600 for linux? or a geforce 8800.. or anything dedicated, for any reason..
 
lol, why would you buy a R600 for linux? or a geforce 8800.. or anything dedicated, for any reason..
Linux makes a great OS for servers (and I do use it for such). In this case, on board graphics are a great solution, 2D open source drivers are great, and everyone is happy.

On the other hand, I run a Linux box as my workstation and gaming box. There are times when I need to view 3D graphs of various data. These often involve OpenGL aware visualization software, and it helps to have a discrete card to handle transparencies and any odd texture requirements that pop up now and then. Even minuscule 3D scenes with a touch of an alpha channel can bog down a 2D card and a software OpenGL library.

Then there is gaming. While I don't do anything crazy, I want to be able to play Half-Life classic, NWN, DII, HL2, Morrowind, etc. It would be nice if I could play the Oblivion I paid for, but I haven't had a chance to try Oblivion through Cedega yet. Anyways, the point is that ATI's binary drivers are not really all that great. While they may be able to handle my work-related requirements, Cedega and a number of games tend to require (or benefit from) features that are missing from ATI's cards/drivers in Linux. Speed is also often a factor. Until ATI starts putting out better Linux drivers, there's no way a desktop Linux user is going to pick one of their cards over the competition's.
 
I didn't think that ATI/AMD could foul things up worse than they did last time (7800 series out for 6 months before the X1800 came to market and when it did come to market it was only marginally better), but it looks like they did it again with the R600, except this time even worse than the last.

Yup, if this is just pure marketing bullshit, it will be a very expensive mistake for AMD/ATI.

Launching a "platform" is not important for a high-end graphics card, because most of the sales of a high-end card are retail. The OEM market for high-end graphics cards is actually pretty small.

I only hope for ATI's sake that this doesn't delay their mid-range R600 derivatives. It's one thing to hand the high-end market to Nvidia for 6 months, but it's another thing entirely to give up the lucrative sub-$200 market. If AMD/ATI lets Nvidia release the 8600 and 8500 series without rebuttal, that's practically a death-blow.
 
Hopefully, Barcelona will be everything AMD needs to stay competative with Conroe. As for R600, I dont really agree they should wait, however it may work to their advantage if you think about it, Nvidia is plagued with the worst drivers in the companies history and Not just for vista, look at the platform drivers that have not veen updated for close to a year. Ati could use this time to really optimize their drivers and give nvidia a bloody nose this round.
 
This combo better be something big. On one hand amd finally got dell to start pimping their goods and the ati 19xx series is no slouch. On the flip side, amd has a problem meeting supply, they're getting spanked by c2d, and ati is almost an entire product cycle behind nvidia 8800.
I really hope for amd/ati's sake, barcelona/r600 are stellar performing parts or we may be seeing them go the way of cyrix/s3. It's really hard to stay in business continually losing money producing inferior parts with a dwindling market share.
 
It's comments like this that really get me...

First AMD has hands down the better architecture., and platform.
Second AMD is still gaining market share...

Two of your points are flat out wrong. AMD aint going anywhere. People have been layin down the same line of BS for years and here we are in 2007... AMD is still here... History tends to repeat itself...
 
First AMD has hands down the better architecture., and platform.

That's been true several times in the past and I fully expect it will be true again at some point in the future but I just don't see how you can say what AMD is selling today is better than what the competition is offering. AMD has some solid offerings, especially at certain price points but if I'm building today it's going to be Conroe and a G80. In another few months who knows, but today that's some pretty stiff competition.
 
Naw, I just buy and sell their stock from time to time and try not to lose my ass :p

I'm trying to stay optimistic but with the current product line-up, I honestly can't recommend my brand of choice for a complete new setup :(

Enthusiasts like us may be a small % of the market, but start adding up all the family and friends you brag and/or build systems for, and the numbers start going way up.

I'm rooting for amd with my heart and my wallet....for now. If they can't do a leap frog in the next 12 months (like the original a64 and opteron) then it may be doom & gloom in their future. Let's keep our fingers crossed and wait and see.

Hopefully in 12 months you can poke me with a stick and laugh :D
 
That's been true several times in the past and I fully expect it will be true again at some point in the future but I just don't see how you can say what AMD is selling today is better than what the competition is offering. AMD has some solid offerings, especially at certain price points but if I'm building today it's going to be Conroe and a G80. In another few months who knows, but today that's some pretty stiff competition.

Very well said...........way better than my ramblings!!
 
That's been true several times in the past and I fully expect it will be true again at some point in the future but I just don't see how you can say what AMD is selling today is better than what the competition is offering. AMD has some solid offerings, especially at certain price points but if I'm building today it's going to be Conroe and a G80. In another few months who knows, but today that's some pretty stiff competition.

I'm looking at it from an architectural standpoint, not a performance standpoint. True enough Intel has slightly better performance with its current line.. But architecturally AMD has the better product.

Lets look at the platform. Intel is getting their asses handed to them on the platform.. That should change somewhat when they finally get CSI out the door, but that still puts them 5 years behind AMD. 5 years, and is the primary reason why they lost so much market share in the server market. It is not about performance as much as it is about scalability.

If I put x number of dollars into my system how much performance can I get? If I double it to xx number of dollars what will the performance increase be? You see becouse of the dated archaic platform Intel uses they simply cant compete on the platform...
 
I'm looking at it from an architectural standpoint, not a performance standpoint. True enough Intel has slightly better performance with its current line.. But architecturally AMD has the better product.

Lets look at the platform. Intel is getting their asses handed to them on the platform.. That should change somewhat when they finally get CSI out the door, but that still puts them 5 years behind AMD. 5 years, and is the primary reason why they lost so much market share in the server market. It is not about performance as much as it is about scalability.

If I put x number of dollars into my system how much performance can I get? If I double it to xx number of dollars what will the performance increase be? You see becouse of the dated archaic platform Intel uses they simply cant compete on the platform...

To some degree, yes. Overall though I've heard they're not even saturating most of the I/O avail to a Conroe. (IE - FSB.)

Intel's large caches also hide a lot of the bus issues they have on larger scale systems. :)
 
hes talking about the server arena, and true enough, once you scale past 4 cores, AMD's Opteron is still the better option, even with Woodcrest out....
 
I'm waiting till the R600 is released and then I'm buying. I would love to get a R600 and Barcelona, but if Nvidia's 8900 is better and Core 2 duo is still leading I will have to get them. Brand loyalty can only go so far.
 
Even on the desktop AMD has the better platform. Well let me rephrase, nVidia has a better platform, made possible by HTT....


Not to mention the benefit of having an integrated on die meshed crossbar switch for ondie communication. The list goes on, and in the end AMD will not need nearly as much hacked engineering to further product lines of tomorrow... Intel on the other hand has some major engineering challenges ahead of them that AMD has already accomplished.
 
.. AMD will not need nearly as much hacked engineering ...

Well, the majority of people care about getting stuff done, not whether or not the tool is ``well engineered''. I really could care less if someone built my CPU out of a 300x400 array of toothpicks, if it is faster than a C2D at a comparable price, I would buy it.
 
Is there something wrong with using a simple fix to resolve a problem?

USA spends Millions of dollars for a pen to write in space, Russia uses *drum roll*.....a pencil.

If Intel has a problem in their architecture that is simply resolved by adding more cache, isn't that exactly what they should do to to fix it? why is that a hack. :rolleyes:

After reading about Barcelona at anandtech, I am very excited about whats on the horizon. You also have to read into it, AMD is improving things, a lot of the new features are already implemented to some degree in C2D. There are MANY nifty things to come with Barcelona, I wish I did not have to wait to see how Intel will fight back.
 
Adding cache doesnt "fix" the problem. Sure it hides it somewhat, but the problem still exists, and sooner or later they will run into a wall. AMD has been steadily imnproving architecture and creating solutions. Intel has been adding cache.

When the day comes that Intel needs 8MB or even 16MB of cache to do the same thing that AMD 2MB accomplish says a whole lot.

Lets keep cost down and actually innovate something.

There are often more then one solution to a problem... At one point in time Humans were using parchment scrolls to keep track of data, until oneday someone came up with the idea to bind separate sheets together Of course they could have just kept making the scrolls longer and longer, but in the end binding into a book was the better solution..... A simple solution is not always the right solution.

The printing press is an incredibly complex machine for its day but it revolutionized society. Same thing with the steam engine, the first electronic computers, the first transister radios and so on and on. Like I said the simplest solution is not always the right solution
 
Adding cache doesnt "fix" the problem. Sure it hides it somewhat, but the problem still exists, and sooner or later they will run into a wall. AMD has been steadily imnproving architecture and creating solutions. Intel has been adding cache.

When the day comes that Intel needs 8MB or even 16MB of cache to do the same thing that AMD 2MB accomplish says a whole lot.
But it hasn't come to this day, in fact presently a 2MB L2 cache Core 2 Duo is faster than the same clocked 2MB L2 cache A64.

And in the high-end servers, the lack of big-cache Opterons is one major reason for its poor scalabilty past 4 sockets.
 
But it hasn't come to this day, in fact presently a 2MB L2 cache Core 2 Duo is faster than the same clocked 2MB L2 cache A64.

And in the high-end servers, the lack of big-cache Opterons is one major reason for its poor scalabilty past 4 sockets.

Opterons scale very well. Not sure where you got that information. The opterons still scale better than Intel's Woodcrest.
 
Opterons scale very well. Not sure where you got that information. The opterons still scale better than Intel's Woodcrest.
It's well known that the scalability of Opterons is poor past 4 sockets due to rapidly increasing cache coherency traffic.
 
Back
Top