Sign in with
Sign up | Sign in
Your question

So now we know (tentatively) how Barcelona performs

Last response: in CPUs
Share
September 10, 2007 6:02:14 AM

So now we know (tentatively) how Barcelona performs, and the performance is nowhere near the claims, from either AMD or its Horde.

What are everyones thoughts on that?
September 10, 2007 6:14:10 AM

i think i will find a new hobbie cause this one will be too expensive when amd closes next year. buy that lift kit for my b2 mabee.
September 10, 2007 6:24:06 AM

turpit said:
So now we know (tentatively) how Barcelona performs, and the performance is nowhere near the claims, from either AMD or its Horde.

What are everyones thoughts on that?


[sounds of crickets chirping from the AMD fanboy camp] :lol: 

Seriously though, I think it will be just enough to keep Intel on it's toes innovating for the next couple of years ...let's hope so anyways. As for AMD, that's a tough call. The next 3 quarters will say a lot. Since these chips are larger and they will have to traverse the typical learning curve of a new ramp (meaning yields are a question) they will cost more to produce. How can this be a good thing when AMD is already bleeding cash?
Related resources
September 10, 2007 6:34:19 AM

Yeah, but then, yet another AMD claim is high yields.
September 10, 2007 6:38:57 AM

I hope AMD is getting high yields. For the price they're gonna sell them for ($209 for the slowest one), it will be hard to make money if they're not using every piece of silicon that traverses the plant.
September 10, 2007 6:39:47 AM

i hope private investment saves ati.
September 10, 2007 6:57:16 AM

I'm short on time to evaluate all the numbers. Anyone care to share a brief overview on the news we have received recently? I assume we will have quite a bit more stuff to come this week none the less.

Best,

3Ball
September 10, 2007 7:18:59 AM

turpit said:
So now we know (tentatively) how Barcelona performs, and the performance is nowhere near the claims, from either AMD or its Horde.

What are everyones thoughts on that?



We do??
linkage please...
September 10, 2007 7:35:11 AM

albundy2 said:
i hope private investment saves ati.


Or if AMD gets picked up by IBM :ouch: 
September 10, 2007 7:51:04 AM

Well nvm...I was able to take a quick glance around the web to see some numbers and with an average ~15% increase over K8 it isnt as good as promised, but this is my take on it. Lets say that the respected clock for clock average of C2D over K8 as ~20% and penryn is shaping up to be somewhat ~5% or so better than C2D clock for clock. We will see somewhere around a 5 - 10% difference between performance of phenom against a penryn, in penryns favor on a given app (this is because I am giving the benefit of the doubt of about 5% more over the next 2 - 3 months that they will be spending on phenom putting them possibly 5% clock for clock behind intel) This could be great for us, because it makes them competitive again and could keep the price war going. Unfortunately AMD may not be able to sustain a price war with the larger size of the chip and the obvious easy ability of clock speed scaling that Intel will have with penryn. These are some early looks, but it would seem that Intel may be staying on top performance wise by a small margin and AMD will help keep them down price wise while also supplying another great product...regardless of any false claims that were made with "simluated" processors im sure people will buy a few of their chips...lol (I know I will prolly pick up a phenom proc for some testing when they come out as I like to do some of my own personal comparison testing.), hopefully they manage to do well in the server market in the coming months since that is where they are most likely going to be trying to make the money anyways.

Just curious is to if anyone else is someowhat perturbed at the outrageous claims that we did see early on from the company, which most people tout to have such great morals? I mean...I am not one of the peopel who has ever worshiped AMD for being the helping friend and Intel being the big bully, but for a company that pretty much lives and dies by word of mouth and good rep because of a lack of advertising doesnt it seem that they may not want to lose the trust of their customers by state/publishing such false numbers all over the introweb?!?...Just a thought!

Best,

3Ball
a b à CPUs
September 10, 2007 7:58:53 AM

From Anand...
This sums up AMD best for me.

To put it plainly: Phenom/Barcelona make this price war more difficult on AMD, while Penryn makes it easier on Intel. What's the end game? Is there a solution? We're not sure...

Thing this p!$$e$ me off most... AMD were sneaky f***s about everything. Oh we cant let u bench our chips cos of competitive blah blah. Just be men about it and say we aint got it ready.

U play games, i'll spend my $ elsewhere.
I think MOST people feel the same way i do. Were not Intel fanboys, we just want the best pruduct from the best company.

PS. I honestly do hope AMD can come back and give Intel a hiding. I also wish to procreate with Jessica Alba. Hey, we see who gets lucky first.
September 10, 2007 9:00:09 AM

turpit said:
So now we know (tentatively) how Barcelona performs, and the performance is nowhere near the claims, from either AMD or its Horde.

What are everyones thoughts on that?


Personally, I think this is bad news, even for Intel fanbois. Be honest: can you see any reason for Intel to accelarate the Wolfdale ramp if Phenom x4 / x2 comes up short? I can't. With the pressure off, Intel delays the introduction of 45nm desktop parts for as long as possible. How that benefits the Intel-minded consumer is beyond my powers of comprehension.
September 10, 2007 9:08:08 AM

easyg said:
Personally, I think this is bad news, even for Intel fanbois. Be honest: can you see any reason for Intel to accelarate the Wolfdale ramp if Phenom x4 / x2 comes up short? I can't. With the pressure off, Intel delays the introduction of 45nm desktop parts for as long as possible. How that benefits the Intel-minded consumer is beyond my powers of comprehension.

There's one massive reason, lower manufacturing cost.
September 10, 2007 9:53:11 AM

turpit said:
There ya go....a 15% improvement over K8, which puts it 3% behind C2D
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3092&p=5

http://img393.imageshack.us/img393/5853/barcy2wt8.jpg


Most people understands that the benchmark of Barcelona vs K8 is "simulated", and to have a constrainted benchmark score like that vs K8 is not bad.

Have you seen the server side benchmark of how a 2.0GB Barcelona performed?

I think you either cant read or too much of a pro-Intel troll, everything from you is negativity about AMD. just how much intel shares do you hold?
September 10, 2007 10:01:39 AM

turpit said:
There ya go....a 15% improvement over K8, which puts it 3% behind C2D
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3092&p=5

http://img393.imageshack.us/img393/5853/barcy2wt8.jpg


A gross simplification of facts Turpit.
In the 3d studio max and winrar benchmarks you can see that Barcelona does much better than 15% increase over K8.

http://www.anandtech.com/IT/showdoc.aspx?i=3091&p=10
http://www.anandtech.com/IT/showdoc.aspx?i=3091&p=8

Also you are conveniently ignoring the excellent scaling of the Barcelona.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=309...

It's this scaling that is supposed to make Barcelona perform better compared to K8 and Intel cpu's as clockspeeds increase...

Also:

"Barcelona is currently limited to DDR2-667, we were unsuccessful with attempts to run the memory any faster. Like all other MP Opterons, Barcelona requires the use of registered DDR2 memory, which is inherently slower than the unbuffered stuff we use on desktops."

Let's not get too hysterical about the desktop benchmarks yet, shall we?
September 10, 2007 10:07:25 AM

pete4r said:
Most people understands that the benchmark of Barcelona vs K8 is "simulated", and to have a constrainted benchmark score like that vs K8 is not bad.

what? what simulated?

Have you seen the server side benchmark of how a 2.0GB Barcelona performed?

I think you either cant read or too much of a pro-Intel troll, everything from you is negativity about AMD. just how much intel shares do you hold? said:
Have you seen the server side benchmark of how a 2.0GB Barcelona performed?

I think you either cant read or too much of a pro-Intel troll, everything from you is negativity about AMD. just how much intel shares do you hold?

I believe you're the one who has no clue about what happened.

The server side doesn't look good at all. Although Barcelona @ 2.0Ghz can outperforms Xeon E5345 under memory intensive programs (i.e zVisuel), it is generally slower than Clovertown in most benchmarks.

For some very odd reason, Anand used more memory bandwidth sensitive applications to benchmark K10. I wonder if this is a part of NDA.

This is from Anand regarding zVisuel:
The LINPACK and zVisuel benchmarks make it clear that [b said:
Intel and AMD have about the same raw FP processing power (clock for clock), but that the Barcelona core has the upper hand when the application has to access the memory a lot.]The LINPACK and zVisuel benchmarks make it clear that Intel and AMD have about the same raw FP processing power (clock for clock), but that the Barcelona core has the upper hand when the application has to access the memory a lot.
[/b]

Not to mention that Intel has the advantage of clockspeed. Intel had already release 3.0Ghz Clovertown, with 3.16+ Ghz more to come in Q4. In other words, if AMD can't increase their clockspeed over 2.3~2.5Ghz range in the next 3 months, how can they fight Penryn? How is that not a bad thing for AMD?

Hey, a question for you. Are you here for forum, or are you here for trolling?
September 10, 2007 10:16:16 AM

i think we can all agree that AMD will not die after this. should it not stop the bleeding of cash, it will lessen it ( it should anyway ). other than that.. anyone else see the problem of releasing a $1000 server chip @ 2.ghz? what will happen when 2ghz + comes out? price cuts? huh....

well anyways.. i suggest we all invest in AMD for a quick buck or two.. cause hey.. the shares cant possibly go down further than it already has right? ( im guessing $15 at least... )
September 10, 2007 10:18:06 AM

Jakc said:
A gross simplification of facts Turpid.
In the 3d studio max and winrar benchmarks you can see that Barcelona does much better than 15% increase over K8.

http://www.anandtech.com/IT/showdoc.aspx?i=3091&p=10
http://www.anandtech.com/IT/showdoc.aspx?i=3091&p=8

....and C2 is about 20% better than K8 clock for clock. Do you think AMD created Barcelona to compete against its own products?

Quote:

Also you are conveniently ignoring the excellent scaling of the Barcelona.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=309...

It's this scaling that is supposed to make Barcelona perform better compared to K8 and Intel cpu's as clockspeeds increase...

There are two types of scaling: 1. clockspeed, 2. socket

For clockspeed, Barcelona does appear to gain additional performance upon reaching 2.5Ghz and beyond. But the question is, can AMD do that? Although Pheonom already scaled up to 3.0Ghz, Barcelona has a more stringent requirement, as well as additional features, which are disabled on Phenom. As a result, the ability for Barcelona to reach 2.5Ghz with acceptable yield has yet to be seen.

For socket scaling, AMD has always been the best in this area due to IMC. In conclusion, AMD can still use Barcelona to capture markets of 4-way server or more. However, those only represent a very small percentage of total server market share. For 2 way or single socket, Xeon is still superior.


Also:

"Barcelona is currently limited to DDR2-667, we were unsuccessful with attempts to run the memory any faster. Like all other MP Opterons, Barcelona requires the use of registered DDR2 memory, which is inherently slower than the unbuffered stuff we use on desktops."

Let's not get too hysterical about the desktop benchmarks yet, shall we? said:

Also:

"Barcelona is currently limited to DDR2-667, we were unsuccessful with attempts to run the memory any faster. Like all other MP Opterons, Barcelona requires the use of registered DDR2 memory, which is inherently slower than the unbuffered stuff we use on desktops."

Let's not get too hysterical about the desktop benchmarks yet, shall we?


The point he tried to illustrate was that, Phenom, as a desktop processor, wouldn't be as competitive as Hector once said. In reality, it wouldn't be competitive with C2D at all. Although the gap has closed, Penryn is going to null that advantage again.

Please don't confuse desktop benchmark with server benchmark.
September 10, 2007 10:25:26 AM

yomamafor1 said:
what? what simulated?


I believe you're the one who has no clue about what happened.

The server side doesn't look good at all. Although Barcelona @ 2.0Ghz can outperforms Xeon E5345 under memory intensive programs (i.e zVisuel), it is generally slower than Clovertown in most benchmarks.

For some very odd reason, Anand used more memory bandwidth sensitive applications to benchmark K10. I wonder if this is a part of NDA.

This is from Anand regarding zVisuel:


Not to mention that Intel has the advantage of clockspeed. Intel had already release 3.0Ghz Clovertown, with 3.16+ Ghz more to come in Q4. In other words, if AMD can't increase their clockspeed over 2.3~2.5Ghz range in the next 3 months, how can they fight Penryn? How is that not a bad thing for AMD?

Hey, a question for you. Are you here for forum, or are you here for trolling?


TBH I completely disagree with you.
The benchmarks used by anand were used because they represent a number of applications that would also be used in real life situations, also the desktop benchmarks were in a way "simulated" even though "simulated" is not exactly the right word.

"Barcelona is currently limited to DDR2-667, we were unsuccessful with attempts to run the memory any faster. Like all other MP Opterons, Barcelona requires the use of registered DDR2 memory, which is inherently slower than the unbuffered stuff we use on desktops."

I would call those benchmarks "crippled".

What I do not understand is that Anand used benchmarks optimized for Intel to benchmark the K10.
What point is there to that?

And yes, turpit seems to focus entirely on the bad parts of the desktop benchmarks. Some benchmarks are very impressive in favor of the Barcelona, and those should also be taken into consideration.
In a way everyone is guilty of "selective reading" when it comes to complex benchmarks like these.

I am just happy I will be able to switch my Athlon 64 X2 for a Phenom X4 and see a more than double performance when it comes to VST performance in cubase.

September 10, 2007 10:32:57 AM

yomamafor1 said:
....and C2 is about 20% better than K8 clock for clock. Do you think AMD created Barcelona to compete against its own products?


I was responding to what turpid says.
If you want to make replies that make sense, you should read the entire thread.

Quote:

There are two types of scaling: 1. clockspeed, 2. socket

For clockspeed, Barcelona does appear to gain additional performance upon reaching 2.5Ghz and beyond. But the question is, can AMD do that? Although Pheonom already scaled up to 3.0Ghz, Barcelona has a more stringent requirement, as well as additional features, which are disabled on Phenom. As a result, the ability for Barcelona to reach 2.5Ghz with acceptable yield has yet to be seen.

For socket scaling, AMD has always been the best in this area due to IMC. In conclusion, AMD can still use Barcelona to capture markets of 4-way server or more. However, those only represent a very small percentage of total server market share. For 2 way or single socket, Xeon is still superior.


If you read the desktop review (again?) you will see a part about clockspeed scaling between the 2.0 and 2.5 "crippled" phenom. The scaling seen here is quite good. That is all I was saying.



Quote:
The point he tried to illustrate was that, Phenom, as a desktop processor, wouldn't be as competitive as Hector once said. In reality, it wouldn't be competitive with C2D at all. Although the gap has closed, Penryn is going to null that advantage again.

Please don't confuse desktop benchmark with server benchmark.


I was not confused at all, I just thought it would be nice to look at the entire picture rather than being selective about it.
I couldn't really care less about what Hector said, the Barcelona shows very promising performance in some areas, and in other areas it performs the same or less than Intel.
It would be very interesting to see 3Ghz phenom vs penrym benchmarks. The way I see it, the increased memory bandwidth could lead to some interesting results in favor of AMD at higher clockspeeds.
September 10, 2007 10:36:09 AM

Jakc said:


Also you are conveniently ignoring the excellent scaling of the Barcelona.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=309...

It's this scaling that is supposed to make Barcelona perform better compared to K8 and Intel cpu's as clockspeeds increase...

The C2Ds scales very well in the same benchmarks.


Also:

"Barcelona is currently limited to DDR2-667, we were unsuccessful with attempts to run the memory any faster. Like all other MP Opterons, Barcelona requires the use of registered DDR2 memory, which is inherently slower than the unbuffered stuff we use on desktops."

Let's not get too hysterical about the desktop benchmarks yet, shall we? said:

Also:

"Barcelona is currently limited to DDR2-667, we were unsuccessful with attempts to run the memory any faster. Like all other MP Opterons, Barcelona requires the use of registered DDR2 memory, which is inherently slower than the unbuffered stuff we use on desktops."

Let's not get too hysterical about the desktop benchmarks yet, shall we?

And when it does, it'll face C2Ds freed from FB-DIMMs and a mediocre 1st generation FB-DIMM chipset.
September 10, 2007 10:51:45 AM

LETS ALL HAVE A REALITY CHECK GUYS!!!

The period of dominance that AMD enjoyed with K8 seems to have made everyone forget that prior to that architecture Intel and AMD were pretty much neck and neck. Intel won some benchmarks while AMD won others.

It seems now that Barcelona is finally out the door, i believe we will return to this scenario. Ok K10 isn't as good as we were led to believe but its definately good enough to restore performance parity with Intel.

1. The current Benches that you see are limited to DDR 667. AND The current motherboards do not support the split power planes. With a capable motherboard the memory controller will automatically be clocked another 200Mhz higher.

So in a desktop config clock for clock we will see performance parity with Intel, while having a lower thermal envelope. Clock speed will not be an issue as the B2 revision is already running at 3ghz. So clock speeds will be high by the time Phenom ships.

In a server config, the superior architechure will be enough to remain competitive. What we are seeing is a linear scaling. So as more sockets are added the advantage will increase.

Its by no means a knockout punch for AMD. But its by no means a clear win for Intel wither. i.e The status quo has been returned. And were back to Pentium 3 / Athlon XP days.
September 10, 2007 11:18:36 AM

As a server CPU:

Of course Barcelona first and foremost is for servers. Lets not forget that (some will use them for workstations, but really optimised for servers).

In the comparison on Anand.com with Xeon 5345 (2.33 Ghz) vs. Barcelona 2350 (2.0 Ghz) it holds it's own and in some cases bests the faster Xeon.

This is the best Apples to Apples (server CPU vs. Server CPU) comparison that can be made.

Best Summary of the results (Actual results with no excuses, from Anandtech.com):

The Opteron 2350 (2 GHz) vs. Xeon "Clovertown"
General applications Opteron 2350 (2GHz) equates to Xeon clock speed of:

WinRAR 3.62 2.7 GHz
Fritz Chess engine 1.8 GHz

HPC applications
Intel optimized Linpack 1.9 GHz

3D Applications
3DS Max 9 2 GHz
zVisuel 3D Kribi Engine 2.33 - 2.4 GHz
zVisuel 3D Kribi Engine (AA) 2.4 GHz

Server applications
Specjbb 2.4 GHz
MySQL 2.33 GHz


So, this 2.o Ghz part holds it's own against the 2.33 Ghz Xeon - this is Apples to Apples comparison. When Tigerton comes out expect a 5-10% performance boost for Xeon curtousy of faster FSB (translates to more power!). So this is shaping out to be pretty interesting.

$389.00 for this CPU and $455.00 for 2.33 Ghz Xeon.

The Barcelona CPU seems to scale better than the Xeon the more cores you add.

So I would say these processors are on PAR! Which can be a good thing or a bad thing. One would say, well if AMD had a 2.3Ghz part going against the Intel Xeon part - it would wipe the floor clean. If AMD had a 6.0 Ghz part they would wipe the world free of all poverty, Osama Bin Laden and the axis of evil. The point is, "THEY DON'T HAVE THOSE PARTS." But anyway, if they did - clock for clock it seems AMD would have a slight edge over Clovertown.

Enter Tigerton and Intel will be back up with that slightedge again. Barcelona will not be CRUSHED by Tigerton, especially when it comes to performance/watt - Intel's FSB can only boost performnce so much, and die shrink to 45nm will improve TDP, but unbuffered dimms will still need power. So even on 45nm, expect Barcelona to have the performance/watt advantage - just like 90nm Opteron did over 65nm Xeon - but it will be close as well.
September 10, 2007 12:04:55 PM

For a first time poster i found that sedaine's post is the best well balanced assentment I've read in this thread.
a b à CPUs
September 10, 2007 12:30:25 PM

As usual turpit is first to fan the fanboi flames ... no guesses which side he bats on.

I find it difficult to rubbish a mod but in this case it is justified ... because the flame baiting attempt is just too sweet to ignore Hal ... Dave must have partially disabled your mainframe already it seems?

Look at the scaling of subsequent reviews before dribbling further ... its a server chip.

I'd say AMD has caught up on most fronts and when you take into account the 4S space and up, then It's clear who the superpooter and datacentre people are going to continue to invest in.

The want realistic power savings with performance gain - Barcy delivers there.

I think most of us didn't think the gaming crown would be taken back because the Core2 is just to well designed as a single cpu ... top points to those Israeli engineers.

Now when you take the FSB into account ... well as a server chip ... its flawed from the ground up.

It's saving grace is its cache architecture and awesome integer perf.

I have both types of PC's ... but I would like to see AMD survive ... to continue to keep Intel on it's toes.

I am looking forward to what Intel will do next.

Hopefully quad core prices will be soon in the realm of Joe public ... even if he doesn't really need it yet.
September 10, 2007 12:48:58 PM

Reynod said:
As usual turpit is first to fan the fanboi flames ... no guesses which side he bats on.

I find it difficult to rubbish a mod but in this case it is justified ... because the flame baiting attempt is just too sweet to ignore Hal ... Dave must have partially disabled your mainframe already it seems?

Look at the scaling of subsequent reviews before dribbling further ... its a server chip.

I'd say AMD has caught up on most fronts and when you take into account the 4S space and up, then It's clear who the superpooter and datacentre people are going to continue to invest in.

The want realistic power savings with performance gain - Barcy delivers there.

I think most of us didn't think the gaming crown would be taken back because the Core2 is just to well designed as a single cpu ... top points to those Israeli engineers.

Now when you take the FSB into account ... well as a server chip ... its flawed from the ground up.

It's saving grace is its cache architecture and awesome integer perf.

I have both types of PC's ... but I would like to see AMD survive ... to continue to keep Intel on it's toes.

I am looking forward to what Intel will do next.

Hopefully quad core prices will be soon in the realm of Joe public ... even if he doesn't really need it yet.


Turpit's point is that AMD set the bar very high and did not deliver. When Intel come out with Core 2 it thrashed any existing product.

AMD made the same claims, but AMD didn't thrash Intel, Barcelona is neck and neck with Intel with a 25% clock speed bump coming next year (yes, they said Q4, but you know those procs won't be available until January at the earliest.

Bottom line, Sharikou and Baron had a large serving of humility today. Sharikook will still come out shooting but Baron's reaction will probably be to tuck his tail and disappear for a few days.
September 10, 2007 12:50:20 PM

I dont see how this K10 arch is going to be the cash cow that keeps AMD alive. It only has a significant advantage in one small market segment: Multi-socket FP intensive computing requiring significant core-to-core communication. That segment is not going to generate the billions AMD needs to prepare its next gen arch.
September 10, 2007 12:54:25 PM

TechnologyCoordinator said:
Bottom line, Sharikou and Baron had a large serving of humility today. Sharikook will still come out shooting but Baron's reaction will probably be to tuck his tail and disappear for a few days.


Not surprisingly, Baron is nowhere to be found :lol: 

September 10, 2007 1:06:57 PM

Well, they got a decent chip that is pretty innovative on power consumption, and it scales well, but it looks like linear scaling with no shining sweetspot after a certain clock where the chip "shines", but no biggie.

They would have had a moneymaker on their hands IF the chip wasn't so big, they just have a product that will barely keep them in the game from what I looked at over at Anand. And their test server chip stopped posting. Bad news on a release product.

If they can sell these chips cheap, decent performance, they will be ok. The only thing that concerns me at this point is their die sizes and capacity, and if they could fight Intel on margin alone, which I seriously doubt. Intel knows it is a margin game at this point, and Intel has performance and capacity on their side too.

Bottom line, they needed this chip a year ago at 2.4Ghz.

Well, it is really clear WHY all those AMD execs jumped ship at this point. Now we should look how the stock performs over the next few days.
September 10, 2007 1:36:40 PM

TechnologyCoordinator said:
Turpit's point is that AMD set the bar very high and did not deliver. When Intel come out with Core 2 it thrashed any existing product.

AMD made the same claims, but AMD didn't thrash Intel, Barcelona is neck and neck with Intel with a 25% clock speed bump coming next year (yes, they said Q4, but you know those procs won't be available until January at the earliest.

Bottom line, Sharikou and Baron had a large serving of humility today. Sharikook will still come out shooting but Baron's reaction will probably be to tuck his tail and disappear for a few days.


Well as some1 mentioned, if AMD didn't have the delays, its is more or less on that 40% margin they have claimed... but things didn't turn out to be what they have predicted, and Intel is not going to stand still and get ready to be shot at...

Could you imagine if these benchmark has surfaced 2 or more months ago? It just pointing out AMD do not have the men power or huge resources to recover from something didn't go with the plans. We've got to thank both Intel and AMD for the competitiveness and value the customers,

10 years ago, Intel would proudly announce a 15 - 33MHz increase over its Pentium and Pentium MMX, and consumers would have to pay $100 for that extra 33MHz. :pt1cable: 

I remember there were an article about how AMD killed the plan for K9, and focused on "what customers want" and they have delivered that, like Anandtech pointing out, Barcelona have won the benchmarks that are used or close to real world server applications.

Servers People don't care if the processor can finish a SuperPi 8M in 30 secs, they care if the there are 50,000 queries comes in from 50 users the chip would be able to handle it and finishes the request faster than its competitor.

Looking forward to see Tigerton and Higher Clocked Barcelona, its just getting interesting :) 

Peace
September 10, 2007 1:40:53 PM

Quote:
Seems to me the claims were right on. 40%. However since the two delays, now Intel has caught up so now its like 10%.


Given that nLogic...Intel can now announce Nehalem will be 70% faster than anything AMD currently has on the market and 40% faster than Barcelona. :lol: 
a b à CPUs
September 10, 2007 1:49:22 PM

Hey wombat2 ... go join Jack if you miss him so much.

You know where he went .... it wasn't hard to follow.

Didn't have the nads to go BaronMatrix head to head so I don't miss him.

The Baron is a bit odd at times but at least I know where he stands.

I respect that.

September 10, 2007 2:35:45 PM

The one thing that's coming out that seems frustrating is the AMD benchmark of ACP (Average CPU Power, instead of TDP aka Thermal Design Power). How loaded your CPU is pretty subjective and the idea of coming out with products that aren't as awesome as you projected while at the same time changing the playing field by saying "this is the better benchmark" sounds like shenanigans.

The big positive is that this is not a failure for AMD. It's not a fantastic company saving success, but it's definitely no 4x4. As was pointed out though, this is a HUUUUUGE die and that, along with "Native quad" (which seems to not have that much of a win), will force yields down and costs up for AMD while Intel is looking to severely reduce costs soon with the Penryn launch. What we DO need is more benchmarks. I trust anand, but we need a larger variety.

Add in the fact that there's a lot of time before Phenom actually is released (so AMD may be able to squirm some performance out of it), so we're back to the waiting game :( .
September 10, 2007 3:30:41 PM

Wombat2 said:
I dont see how this K10 arch is going to be the cash cow that keeps AMD alive. It only has a significant advantage in one small market segment: Multi-socket FP intensive computing requiring significant core-to-core communication. That segment is not going to generate the billions AMD needs to prepare its next gen arch.



Barcelona will obviously still rule in multi-socket segament. Also because Intel Xeon is now only 20% advantage over Barcelona on INTEGER (was 100% before Barcelona!!!) it suddenly gives Barcelona a new market segment if they can keep their price competitive.

It made no sense before to shell out money for an Opteron for interger intensive applications since Intel had twice the performance. NOW it's close. Intel is not BLOWING AMD out of the water in any of the benchmarks. It's all very close, in some instances AMD besting Intel.

Intel has Tigerton coming out, so this will certainly change the field, but not to the magnitude that most people hope. Both AMD and Tigerton will be competitive at similar clock speeds.

One thing that I am sure of is this - Barcelona is a step in the right direction, IF they can get higher clock speeds then it will be very competitive, but Intel as well all know is a manufacturing giant and they will use this to crush AMD - especially with Tigerton coming out - 'read price cuts for Clovertown' - it's going to be a fight for AMD unless they get those clocks up.
September 10, 2007 4:15:17 PM

Let's see how the desktop variant perform first, altough it really doesn't look so gook for AMD right now.

It'll probably be enough to put them on the same playing field as Intel, but not to gain back market share I think. Take a guy like me with a decent PC based on Intel. I'll pobably upgrade next year for an Intel 45nm quad-core cpu since AMD will offer me nothing good enough to switch my mobo at the same time to go with them. My current mobo will support Intel 45nm cpu, so Intel it will be.

Still, I think AMD is in a better position today then they were last week...
September 10, 2007 4:26:03 PM

pete4r said:
Most people understands that the benchmark of Barcelona vs K8 is "simulated", and to have a constrainted benchmark score like that vs K8 is not bad.

Have you seen the server side benchmark of how a 2.0GB Barcelona performed?

I think you either cant read or too much of a pro-Intel troll, everything from you is negativity about AMD. just how much intel shares do you hold?



NO NO NO NOoooooooo, this is not a simulation. If you bothered to actaully read the article, you would have read the following, from Anand Lal Shimpi of Anandech:

Quote:

We got a call earlier in the week asking if we'd be able to turn around a review of AMD's Barcelona processor for Monday if we received hardware on Saturday. Naturally we didn't decline, and as we were secretly working on a Barcelona preview already, AMD's timing was impeccable

AMD shipped us a pair of 2U servers a day early, we actually got them on Friday but being in Denver at CEDIA we couldn't begin testing until Saturday. Luckily, Johan had Barcelona in Europe for over a week by this point and was already hard at work on server benchmarks. I augmented Johan's numbers with some additional results on these servers, but I had other plans in mind for the Barcelona system that AMD was sending me.


These are real benchmarks of a real processoer, not mearly another regurigitation of AMDs now blatently optimistic "simulations". You didnt bother reading the article, did you?
September 10, 2007 4:35:08 PM

Have to say that both AMD and Intel have brought us products that are far more advanced than anything offered 2years ago.

Now its up to the software developers, cable/dsl/fiber optic/dial up companies to catch up to them lol.
September 10, 2007 4:40:16 PM

Intel doesn't have Tigerton coming out. Tigerton is out (as much as Barcelona is "out"). Intel has Penryn coming out in the near future. AMD should have higher clocked Barcy's in the near future. Just clarifying.
September 10, 2007 5:56:50 PM

turpit said:
So now we know (tentatively) how Barcelona performs, and the performance is nowhere near the claims, from either AMD or its Horde.

What are everyones thoughts on that?


I think they did exactly what they said. Right now I believe K10 will be handcuffed until the HT3 chipset is released. All of the places where K8 was totally outclassed are gone. Even in Intel optimized code, AMD has made up the lost ground. Only clockspeed is holding them back as they seem to be scaling better than Xeon.

The true beauty of the release is the improvement over Opteron and the greater power efficiency. Asus even released a board with the Split plane so even people buying new can take advantage of the price\perf ratio.

An interesting test would be to pop the 2350 into a QFX board and OC it to see how high it can go.
September 10, 2007 6:13:34 PM

Jakc said:

A gross simplification of facts Turpit.
In the 3d studio max and winrar benchmarks you can see that Barcelona does much better than 15% increase over K8.

http://www.anandtech.com/IT/showdoc.aspx?i=3091&p=10
http://www.anandtech.com/IT/showdoc.aspx?i=3091&p=8


No, Jakc, it was not a gross simplifications.
From Anand:
Quote:
Here's where things get complicated; we knew Phenom/Barcelona would be faster clock for clock, it was only a matter of how big of an improvement we'd get. If we are to believe that 15% is the best we'll get on average, taking into account that Penryn is around 5% faster than Conroe, the updated architecture from AMD alone isn't enough to really compete with Intel. In other words, price matters.


If you are going to cast accusations, then certainly you will not be upset if I cast them back.

Now, it is funny that you would accuse me of a "gross simplification" for refering to the average (net sum of ALL the benchmarks) performance increase when you have focused on only 2 benchmarks, (curiously, the very benchmarks in which the Barcelona performs very well), 3d studio max and winrar. This itself could be viewed as a "gross simplification" or worse and more accurate, a gross misdirection on your part. Or, perhaps, just a gross error. Personally, I am curious at to which.

15% is the overall perfromance increase. Are you really going to play the same game AMD played earlier this year, of "dont look at these other benchmarks, look only at the few benchmarks that make us look good", If thats really what you want to do, key on specific benchmarks, excluding the others for the overall estimate, then allow me to present these benchmarks (taken from the correct test series on page 5):
In Cinebench Barcleona only shows a 4.6% increase, and worse, in Light Wave, it only acheived a 1.8% increase.

AMD has gone from publically slamming the benchmarks up Intels nose when AMD was brilliantly thrashing them in performance, to desperately inventing the phrases "platformance", "megatasking" creating something ambiguous that they could claim dominance in, to keying on individual benchmarks in power, to now creating a new, ambiguous power benchmark, all to present themselves in the best possible light. That is business, to be expected, and nothing to be too terribly upset about, but it does not mean the individual consumer should be foolish enough to bite on that bait.

Jakc said:

Also you are conveniently ignoring the excellent scaling of the Barcelona.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=309...

It's this scaling that is supposed to make Barcelona perform better compared to K8 and Intel cpu's as clockspeeds increase...?


I dont think you understand how scaling works, do you Jakc? It is not some magical perfromance multiplier that is going to miraculously boost Barcelonas perfromance exponentially as clockspeeds increase. Barcelona scales well, and the only real coment to make there is thank god that it doesnt scale poorly. If it did, there would be little point in pursuing faster clocks. As it is, it means they will not inccur a performance penalty as they improve clock speeds, meaning they should retain (relatively) roughly the same perfromance clock for clock at the higher clockspeeds. This is a good thing, but again, they cannot 'create perfromance' from nothing. Niether can Intel.
From Anandtech
Quote:
Even at 2.6GHz, a 15% boost in overall performance won't be enough to dethrone Intel, but with competitive pricing AMD could actually be a real alternative again.


Jakc said:

Also:

"Barcelona is currently limited to DDR2-667, we were unsuccessful with attempts to run the memory any faster. Like all other MP Opterons, Barcelona requires the use of registered DDR2 memory, which is inherently slower than the unbuffered stuff we use on desktops."?

Let's not get too hysterical about the desktop benchmarks yet, shall we?


First, no one is "hysterical". AMD lied. Pure and simple. It is a statement of fact. There is nothing hysterical in that, unless you are a fanboy who is upset now that Barcelonas performance is approaching conclusive certainty.
Second, while you do not specifically accuse me of being "hysterical", the implication is certainly there, so please show me where you feel I am being "hysterical". If you cant do that, then in rebuttal, at least show me where AMD claimed a 3% perfromance deficit over C2D, or only a 15% gain over K8 to prove that AMD hasent lied.

As for the RAM, you know, its funny, here I thought I was being nice by not pointing out the blatently obvious, (yet another individual acuses me of being biased), and then you choose to publicize this little memory speed tidbit. So, lets discuss it. Since you chose to bring it up, we should all hope that the inability to run the faster RAM is a motherboard or chipset problem, and not a CPU problem. It would not look good for AMD, if after so many months of developement (on an improved K8 Uarch mind you, not a totaly new uarch, just a modification) they were unable to run DR800. Fortunately this does appear to be the case, however, until we see the newer motherboards, we wont know for sure.

Quote:
These are standard Socket-1207 servers, meaning they don't take advantage of the split power-planes of Barcelona. Newer motherboards (such as the one Johan used in his tests) will support split power-planes, allowing Barcelona to run its North Bridge at a higher clock frequency, thus improving memory performance.



DDR 800, like scaling, is not going to push Barcelona to any great relative extremes of stellar perfromance. If its perfromance improves similarly to K8s with faster RAM, going from DDR 667 to DR 800 should give it up to an 8% further increase over K8, (15~20% going from 533 to 800) putting it 5% over C2D. And thats is a big if. Nowhere near AMDs or the fanboys 'estimations'

Finally,
Quote:
Then, at the very end (literally two hours before publication) of our benchmarking, the AMD server stopped POSTing. As of now the system will simply sit there and spin its fans without actually putting anything on the screen. A number of things could have happened, but thankfully the Barcelona system decided to die after we ran all of our tests.

The take away point here is that AMD is working as quickly as possible to push out this Barcelona release. While AMD states to expect availability as early as today, we're concerned about the maturity of these platforms. We got our test systems on Friday, if this were truly a large, widespread launch we would've had hardware long ago. Whereas everyone and their mom already has a sample of Intel's 45nm CPUs, Barcelona continues to be tough to come by and performance hasn't been anywhere near final until now.


People need to read the article before running their sucks. For that fact of the matter they should read the threads post.

Quote:
So now we know (tentatively) how Barcelona performs, and the performance is nowhere near the claims, from either AMD or its Horde.

What are everyones thoughts on that?


You didnt answer the question. How do you feel about AMD's lies this year?
September 10, 2007 6:46:23 PM

Bottom line, Sharikou and Baron had a large serving of humility today. Sharikook will still come out shooting but Baron's reaction will probably be to tuck his tail and disappear for a few days.

So Turpit can you see now why people here make me sick. I have a full-time job and have been reading reviews this morning. The scaling is superior, the power efficiency is superior. Anand HAS a 2.5GHz Opteron as did another site linked through AMDZone.

They show that K10 is what it was said to be. We still don't know which stepping these chips were since B0\B1 were supposedly the shipping revs and no one said any different.

The gaming tests DO NOT COUNT because no on ehad 1066DDR2 or HT 3 mobos. I hope ASUS has an upgrade BIOS for QFX so we can see the scaling of that platform. It has been said that one Phenom will kick the butt of two QFXs. I'd like to see it happen.

Hear that Tomz. Give it a try. The 2350 should work since the 22xx chips work. Now that would be a better test for desktop perf.
September 10, 2007 7:05:48 PM

turpit said:
There ya go....a 15% improvement over K8, which puts it 3% behind C2D
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3092&p=5

http://img393.imageshack.us/img393/5853/barcy2wt8.jpg


Jakc said:
A gross simplification of facts Turpit.
In the 3d studio max and winrar benchmarks you can see that Barcelona does much better than 15% increase over K8.

http://www.anandtech.com/IT/showdoc.aspx?i=3091&p=10
http://www.anandtech.com/IT/showdoc.aspx?i=3091&p=8

Also you are conveniently ignoring the excellent scaling of the Barcelona.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=309...

It's this scaling that is supposed to make Barcelona perform better compared to K8 and Intel cpu's as clockspeeds increase...

Also:

"Barcelona is currently limited to DDR2-667, we were unsuccessful with attempts to run the memory any faster. Like all other MP Opterons, Barcelona requires the use of registered DDR2 memory, which is inherently slower than the unbuffered stuff we use on desktops."

Let's not get too hysterical about the desktop benchmarks yet, shall we?


turpit said:
No, Jakc, it was not a gross simplifications.
From Anand:
Quote:
Here's where things get complicated; we knew Phenom/Barcelona would be faster clock for clock, it was only a matter of how big of an improvement we'd get. If we are to believe that 15% is the best we'll get on average, taking into account that Penryn is around 5% faster than Conroe, the updated architecture from AMD alone isn't enough to really compete with Intel. In other words, price matters.


If you are going to cast accusations, then certainly you will not be upset if I cast them back.

Now, it is funny that you would accuse me of a "gross simplification" for refering to the average (net sum of ALL the benchmarks) performance increase when you have focused on only 2 benchmarks, (curiously, the very benchmarks in which the Barcelona performs very well), 3d studio max and winrar. This itself could be viewed as a "gross simplification" or worse and more accurate, a gross misdirection on your part. Or, perhaps, just a gross error. Personally, I am curious at to which.

15% is the overall perfromance increase. Are you really going to play the same game AMD played earlier this year, of "dont look at these other benchmarks, look only at the few benchmarks that make us look good", If thats really what you want to do, key on specific benchmarks, excluding the others for the overall estimate, then allow me to present these benchmarks (taken from the correct test series on page 5):
In Cinebench Barcleona only shows a 4.6% increase, and worse, in Light Wave, it only acheived a 1.8% increase.

AMD has gone from publically slamming the benchmarks up Intels nose when AMD was brilliantly thrashing them in performance, to desperately inventing the phrases "platformance", "megatasking" creating something ambiguous that they could claim dominance in, to keying on individual benchmarks in power, to now creating a new, ambiguous power benchmark, all to present themselves in the best possible light. That is business, to be expected, and nothing to be too terribly upset about, but it does not mean the individual consumer should be foolish enough to bite on that bait.



I dont think you understand how scaling works, do you Jakc? It is not some magical perfromance multiplier that is going to miraculously boost Barcelonas perfromance exponentially as clockspeeds increase. Barcelona scales well, and the only real coment to make there is thank god that it doesnt scale poorly. If it did, there would be little point in pursuing faster clocks. As it is, it means they will not inccur a performance penalty as they improve clock speeds, meaning they should retain (relatively) roughly the same perfromance clock for clock at the higher clockspeeds. This is a good thing, but again, they cannot 'create perfromance' from nothing. Niether can Intel.
From Anandtech
Quote:
Even at 2.6GHz, a 15% boost in overall performance won't be enough to dethrone Intel, but with competitive pricing AMD could actually be a real alternative again.




First, no one is "hysterical". AMD lied. Pure and simple. It is a statement of fact. There is nothing hysterical in that, unless you are a fanboy who is upset now that Barcelonas performance is approaching conclusive certainty.
Second, while you do not specifically accuse me of being "hysterical", the implication is certainly there, so please show me where you feel I am being "hysterical". If you cant do that, then in rebuttal, at least show me where AMD claimed a 3% perfromance deficit over C2D, or only a 15% gain over K8 to prove that AMD hasent lied.

As for the RAM, you know, its funny, here I thought I was being nice by not pointing out the blatently obvious, (yet another individual acuses me of being biased), and then you choose to publicize this little memory speed tidbit. So, lets discuss it. Since you chose to bring it up, we should all hope that the inability to run the faster RAM is a motherboard or chipset problem, and not a CPU problem. It would not look good for AMD, if after so many months of developement (on an improved K8 Uarch mind you, not a totaly new uarch, just a modification) they were unable to run DR800. Fortunately this does appear to be the case, however, until we see the newer motherboards, we wont know for sure.

Quote:
These are standard Socket-1207 servers, meaning they don't take advantage of the split power-planes of Barcelona. Newer motherboards (such as the one Johan used in his tests) will support split power-planes, allowing Barcelona to run its North Bridge at a higher clock frequency, thus improving memory performance.



DDR 800, like scaling, is not going to push Barcelona to any great relative extremes of stellar perfromance. If its perfromance improves similarly to K8s with faster RAM, going from DDR 667 to DR 800 should give it up to an 8% further increase over K8, (15~20% going from 533 to 800) putting it 5% over C2D. And thats is a big if. Nowhere near AMDs or the fanboys 'estimations'

Finally,
Quote:
Then, at the very end (literally two hours before publication) of our benchmarking, the AMD server stopped POSTing. As of now the system will simply sit there and spin its fans without actually putting anything on the screen. A number of things could have happened, but thankfully the Barcelona system decided to die after we ran all of our tests.

The take away point here is that AMD is working as quickly as possible to push out this Barcelona release. While AMD states to expect availability as early as today, we're concerned about the maturity of these platforms. We got our test systems on Friday, if this were truly a large, widespread launch we would've had hardware long ago. Whereas everyone and their mom already has a sample of Intel's 45nm CPUs, Barcelona continues to be tough to come by and performance hasn't been anywhere near final until now.


People need to read the article before running their sucks. For that fact of the matter they should read the threads post.

Quote:
So now we know (tentatively) how Barcelona performs, and the performance is nowhere near the claims, from either AMD or its Horde.

What are everyones thoughts on that?


You didnt answer the question. How do you feel about AMD's lies this year?


Ladies and gentlemen of this supposed jury, Chef's attorney would certainly want you to believe that his client wrote "Stinky Britches" ten years ago. And they make a good case. Hell, I almost felt pity myself! But, ladies and gentlemen of this supposed jury, I have one final thing I want you to consider. Ladies and gentlemen, this is Chewbacca. Chewbacca is a Wookiee from the planet Kashyyyk. But Chewbacca lives on the planet Endor. Now think about it; that does not make sense!

Why would a Wookiee, an eight-foot tall Wookiee, want to live on Endor, with a bunch of two-foot tall Ewoks? That does not make sense! But more important, you have to ask yourself: What does this have to do with this case? Nothing. Ladies and gentlemen, it has nothing to do with this case! It does not make sense! Look at me. I'm a lawyer defending a major record company, and I'm talkin' about Chewbacca! Does that make sense? Ladies and gentlemen, I am not making any sense! None of this makes sense! And so you have to remember, when you're in that jury room deliberatin' and conjugatin' the Emancipation Proclamation, does it make sense? No! Ladies and gentlemen of this supposed jury, it does not make sense! If Chewbacca lives on Endor, you must acquit! The defense rests.
September 10, 2007 7:41:58 PM

screw AMD and Intel......who's for Nvidia jumping into the CPU market and really giving everyone a run for their money?? and the complete destruction of price scales as we know it....whats better for us than having 3 competetors instead of 2..hehe.....(wishful thinking) lol
September 10, 2007 8:06:43 PM

i believe AMD made a mistake by introducing L3 Cache; the latency is so high. this is why K10 isnt that overwhelming. if AMD made K10 without L3 but with larger L2 cache (2MB each core), it will no doubt pawn C2D hands down.
September 10, 2007 8:20:19 PM

edwuave said:
i believe AMD made a mistake by introducing L3 Cache; the latency is so high. this is why K10 isnt that overwhelming. if AMD made K10 without L3 but with larger L2 cache (2MB each core), it will no doubt pawn C2D hands down.



Well more work is done per cycle - so you need cache to shuffle all that data being processed between L3 cache. Remember, AMD brought about the idea of doing more with lower clocked CPU. Same thing here. Even though it has higher latency - it's doing more work as a result of having it there.

Put better, if he latency was 10ns and it only processed 1/20th of the job because it didn't have enough cache to shuffle between the CPU's it would hurt the processor. So you are better of with a CPU with 70ns latency but with the ability to shuffle as much data between the cores. Intel does this - albeit with a memory controler, but they have large caches for this very reason, this used to hurt AMD.
September 10, 2007 8:46:47 PM

turpit said:
NO NO NO NOoooooooo, this is not a simulation. If you bothered to actaully read the article, you would have read the following, from Anand Lal Shimpi of Anandech:

These are real benchmarks of a real processoer, not mearly another regurigitation of AMDs now blatently optimistic "simulations". You didnt bother reading the article, did you?


If you would bother to read the article and think for one sec about what you read, you would find out that the desktop benchmarks are not the performance one can expect from a desktop system with phenom chips in them.
It's a preview, and that is all it's ment to be.

This is also the reason they compared it to K8 only and not intel cpu's.
September 10, 2007 9:01:47 PM

The End of line

Turpid is blindly believing his tunnel vision over other's views backed by major and respectiful hardware and tech sites such as anandtech.

he find it hard to understand that the "simulated" or "preview" of the benchmark of Barcelona server chip doing a desktop gaming job is NOT is actual same thing as an actual desktop processor doing the job it designed to be doing.

It is not that hard for most of us to understand comparing Barcelona to K8 X2 Athlon is actually the same thing as comparing Clowertown to E6xxx series...

Turpid could defend himself "they are using the same architecture" but why the heck would both INTEL and AMD bother creating 2 completely different market segment?? - There are difference, and a lot difference... Im not here to explain the difference, if you don't know them Turpid, you shouldn't be here.

Posting 100 messages a day does not make you or represent that you have more knownledge, in fact, most of your posts are waste of time, defending your misleading statements.
September 10, 2007 9:21:05 PM

turpit said:


If you are going to cast accusations, then certainly you will not be upset if I cast them back.


Not at all, discussing (and accusing) is what we do best around here right?
Quote:

Now, it is funny that you would accuse me of a "gross simplification" for refering to the average (net sum of ALL the benchmarks) performance increase when you have focused on only 2 benchmarks, (curiously, the very benchmarks in which the Barcelona performs very well), 3d studio max and winrar. This itself could be viewed as a "gross simplification" or worse and more accurate, a gross misdirection on your part. Or, perhaps, just a gross error. Personally, I am curious at to which.


I am sure you understand I was trying to give examples of situations in which your "15% over K8" makes no sense anymore. I do not think it is fair to summ up these complex and comprehensive reviews by saying Barcelona is 15% faster than K8 hence will be slower than penryn.

Quote:
15% is the overall perfromance increase. Are you really going to play the same game AMD played earlier this year, of "dont look at these other benchmarks, look only at the few benchmarks that make us look good", If thats really what you want to do, key on specific benchmarks, excluding the others for the overall estimate, then allow me to present these benchmarks (taken from the correct test series on page 5):
In Cinebench Barcleona only shows a 4.6% increase, and worse, in Light Wave, it only acheived a 1.8% increase.


Very good turpit. Shall I again explain that I was giving examples of situations in which your statement was wrong? I was not giving a total verdict of the barcelona based on those few benchmarks, neither was I making ridiculous 4 word statements that attempt (but fail miserably) to capture the entire array of benchmarks like you did.

Quote:
AMD has gone from publically slamming the benchmarks up Intels nose when AMD was brilliantly thrashing them in performance, to desperately inventing the phrases "platformance", "megatasking" creating something ambiguous that they could claim dominance in, to keying on individual benchmarks in power, to now creating a new, ambiguous power benchmark, all to present themselves in the best possible light. That is business, to be expected, and nothing to be too terribly upset about, but it does not mean the individual consumer should be foolish enough to bite on that bait.


If you think that Barcelona using less power to render a 3d max scene than any intel chip, is bait only fools bite on, then I guess you are just not interested in this segment of the market.
Reality is that whatever best possible light AMD is putting their processors in, is possibly a light that at least a part of the market is interested in, and has use for.

Quote:
I dont think you understand how scaling works, do you Jakc? It is not some magical perfromance multiplier that is going to miraculously boost Barcelonas perfromance exponentially as clockspeeds increase. Barcelona scales well, and the only real coment to make there is thank god that it doesnt scale poorly. If it did, there would be little point in pursuing faster clocks. As it is, it means they will not inccur a performance penalty as they improve clock speeds, meaning they should retain (relatively) roughly the same perfromance clock for clock at the higher clockspeeds. This is a good thing, but again, they cannot 'create perfromance' from nothing. Niether can Intel.


And where did I claim AMD was going to dethrone Intel?
To me the good scaling is a merely a sign that in some time, when the barcelona goes beyond 3 Ghz, we may see limiting factors on intel systems that do not exist on AMD systems because of the different architecture and bustype. It would be very interesting to compare 2.5 Ghz vs 3.2 Ghz scaling on Barcelona to the intel offerings, and I think that after what I have seen from the server benchmarks, AMD will do at least a little bit better than intel, but I can not be sure.

Quote:
First, no one is "hysterical". AMD lied. Pure and simple. It is a statement of fact. There is nothing hysterical in that, unless you are a fanboy who is upset now that Barcelonas performance is approaching conclusive certainty.
Second, while you do not specifically accuse me of being "hysterical", the implication is certainly there, so please show me where you feel I am being "hysterical". If you cant do that, then in rebuttal, at least show me where AMD claimed a 3% perfromance deficit over C2D, or only a 15% gain over K8 to prove that AMD hasent lied.


Look I don't care what AMD has said. Really I don't, I never said I did either.
And your negative focus on the preview of the phenom(with the crippled MB and memory), shows to me a rather hysterical desire to point out bad things about Barcelona.
If you want to come across like that, then fine, I don't really care.
I just see something else I guess, when I look at those benchmarks. I see a processor with strong points and weak points, too complex to capture in a 5 word sentence. I also see that performance as in these first benchmarks may not yet be best possible. (remember the first athlon 64 benchmarks?)
And even if this is the best possible, and AMD will never beat intel again, then I still see things about barcelona in those server benchmarks that make me think, wow, they changed process technology AND architecture in one go, and it resulted in something competetive in some areas. Impressive for an underdog company.


Quote:
As for the RAM, you know, its funny, here I thought I was being nice by not pointing out the blatently obvious, (yet another individual acuses me of being biased), and then you choose to publicize this little memory speed tidbit. So, lets discuss it. Since you chose to bring it up, we should all hope that the inability to run the faster RAM is a motherboard or chipset problem, and not a CPU problem. It would not look good for AMD, if after so many months of developement (on an improved K8 Uarch mind you, not a totaly new uarch, just a modification) they were unable to run DR800. Fortunately this does appear to be the case, however, until we see the newer motherboards, we wont know for sure.


Quote:
These are standard Socket-1207 servers, meaning they don't take advantage of the split power-planes of Barcelona. Newer motherboards (such as the one Johan used in his tests) will support split power-planes, allowing Barcelona to run its North Bridge at a higher clock frequency, thus improving memory performance.



Quote:
DDR 800, like scaling, is not going to push Barcelona to any great relative extremes of stellar perfromance. If its perfromance improves similarly to K8s with faster RAM, going from DDR 667 to DR 800 should give it up to an 8% further increase over K8, (15~20% going from 533 to 800) putting it 5% over C2D. And thats is a big if. Nowhere near AMDs or the fanboys 'estimations'


Again, I couldn't care less about AMD or fanboy estimations! Fudzilla said Barcelona 1.9 Ghz could be clocked to 3.1 Ghz, it's all bullshit anyways and who cares?
If your best case scenario is true, then phenom is going to be a very impressive processor!
You know it is really ok to say something good about AMD, you should try it sometimes, I assure you it will do nothing to diminish the intel fanboy image you've created for yourself on these forums.


Quote:

So now we know (tentatively) how Barcelona performs, and the performance is nowhere near the claims, from either AMD or its Horde.

What are everyones thoughts on that?

You didnt answer the question. How do you feel about AMD's lies this year?


To repeat myself, I couldn't care less, about AMD's lies, about fanboy ramblings and especially not about your fanboy flamefest provoking threads.
I think AMD didn't do as bad as I feared when delay after delay after delay happened and there was total radiosilence.

I actually think phenom may prove to supply some healthy competition in the CPU market.
And I am just glad about that, whether or not I decide to buy AMD.
!