• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

AMD posts Barcelona benchmark numbers

Great, although it was with a 2.6 GHz chip. They won't launch at that speed.
 
I'll be interested in seeing if all of these improvements combine to create significant benefit clock for clock over the conroe based processors in consumer level benchmarks. If the clock speed ramp problem gets fixed, then I think they should have a winner on their hands.
 
The Barcelona 2.6GHz chip doesn't even have an estimated ship date yet.

The SPEC scores quoted for Intel are also wrong. edit: see this page for the correct values: http://blogs.zdnet.com/Ou/?p=567

This is pretty pathetic.
 
quadnad said:
I'll be interested in seeing if all of these improvements combine to create significant benefit clock for clock over the conroe based processors in consumer level benchmarks. If the clock speed ramp problem gets fixed, then I think they should have a winner on their hands.
Click to expand...
The SPEC rate tests, especially SPECfp_rate don't really correlate that well with consumer applications. For example, the Quad FX -74 would beat a QX6700 by about 40% in SPECfp_rate 2006.
 
Seeing the ZDNet article makes me sad...we really needed some good competition here.

Shame on you, AMD...
 
SLee said:
The SPEC rate tests, especially SPECfp_rate don't really correlate that well with consumer applications. For example, the Quad FX would beat a QX6700 by about 40% in SPECfp_rate 2006.
Click to expand...

Yeah, I realize that. Hopefully the other benefits they've put into the K8 core drastically improve the IPC in real world applications.
 
Yeah unfortunately Barcelona has not proven to have the potential we were hoping it would....
Of course, we are talking about hardware that hasn't been released yet, and probably won't be for a couple more months, but if the Core2Duo launch taught me anything, it's that the "leaked" samples and benchmarks were pretty darn accurate.

If that holds true for Barcelona, :(
 
quadnad said:
Yeah, I realize that. Hopefully the other benefits they've put into the K8 core drastically improve the IPC in real world applications.
Click to expand...

Agreed - Let's hope that in real world usage the Barcelona holds up its own.

But regardless of real/simulated tests, it's still shady of them to show only the results of the 2.6ghz clocked Barcelona, where in reality only the 2.0ghz is scheduled for release, all while while pitting it against older hardware.

That's like saying "This new processor, which is better than the processor that we will actually be releasing, is far better than our competitors older, more dated products."

Hrm...*hugs his 3000+* You'll still always be my favorite machine...
 
yawn.... Let's see some REAL benchmarks with a CPU that will actually be available at launch.
 
It's unfortunate, but AMD seems to have shot themselves in the foot.

I mean, they had years of lead time to make their cpu's even faster while they were on top performance wise. But all they did were little tweaks that got the current stuff to clock higher easier.

Now if they had worked to make their cpu's faster clock speed wise and performance at the same speed wise then they wouldn't be in this situation right now I'd bet. To me at least it really seems like they slacked off while they had the performance crown.
 
WintersFury said:
To me at least it really seems like they slacked off while they had the performance crown.
Click to expand...

According to one of the more recent Anandtech articles AMD had actually started work on a new processor while K8 was doing well, but while it was still in the design stage they had to scrap it. Supposedly it was an extremely parallel design, but because AMD had overestimated the adoption rate of multithreaded software the concept wouldn't have yeilded good enough performance.

EDIT: I just reread the article and it turns out that only lost them about 6 months. Considering that they released K8 in 2003, that doesn't really explain why we're where we are right now, but I think it gives us an idea that they weren't completely sitting around.
 
You can't expect AMD to kick Intel's ass every time. It's amazing that they were outrunning Intel for so long in the first place.

On the other hand, I do think that they need to shake up their upper management. They really need better marketing.
It's amazing how many people don't know about AMD or think they still make cheap copycat processors.
 
Just a few years ago, AMD was basically an engineering only company. They have come quite a way in the marketing realm of things these past few years. About two weeks ago, I went to AMD's Austin location and met with quite a few people including the head of marketing for the mobile division and another incharge of the desktop side.

Slowly but surely, they are beginning to become more of a "marketing" company than they had been for years.
 
GO BARCELONA ARCHITECTURE!!! :)

I'm going for AMD because I just want to pop in a quad core without changing my mobo or memory. So these preliminary "tests" are good news to me.
 
DeathFromBelow said:
You can't expect AMD to kick Intel's ass every time. It's amazing that they were outrunning Intel for so long in the first place.

On the other hand, I do think that they need to shake up their upper management. They really need better marketing.
It's amazing how many people don't know about AMD or think they still make cheap copycat processors.
Click to expand...

The way things are looking now, it won't be long before what those people think is actually what AMD makes... Cheap copycat processors, becuase if Barcelona isn't successful, they won't have the resources for the proper R&D. It's too bad really, finally, after many many years, AMD was finally starting to make a name for themselves and getting OEM's onboard with the success of the K8.
 
AMD pulled the bogus benchmarks, just in time, too. spec.org is back up after being down a few days. Results like these official ones could be pretty embarassing:

X5355 http://www.spec.org/cpu2006/results/res2007q2/cpu2006-20070612-01275.html
X5365 http://www.spec.org/cpu2006/results/res2007q2/cpu2006-20070528-01175.html

(note: these are not estimations) :p

SpecFP for Intel will improve with the faster FSB coming with 45nm Xeons. It will still trail behind Barcelona clock for clock (if AMD's projections are right), but it's moot since AMD will be at a 50% clock speed disadvantage
 
pxc said:
The Barcelona 2.6GHz chip doesn't even have an estimated ship date yet.

The SPEC scores quoted for Intel are also wrong. edit: see this page for the correct values: http://blogs.zdnet.com/Ou/?p=567

This is pretty pathetic.
Click to expand...

oooooohhhh George Ou, how about a link to some one who isn't a hypocrite. He thinks its fair for intel to compare quads to amds dual core, not to mention that those results are skewed since the intel numbers are used with a different more optimized compiler.

Kudos for being an intel fanboi. I for one am holding off until I see some real numbers from people who do not get paid by either amd or intel to spin for them.


pxc said:
AMD pulled the bogus benchmarks, just in time, too. spec.org is back up after being down a few days. Results like these official ones could be pretty embarassing:

X5355 http://www.spec.org/cpu2006/results/res2007q2/cpu2006-20070612-01275.html
X5365 http://www.spec.org/cpu2006/results/res2007q2/cpu2006-20070528-01175.html

(note: these are not estimations) :p

SpecFP for Intel will improve with the faster FSB coming with 45nm Xeons. It will still trail behind Barcelona clock for clock (if AMD's projections are right), but it's moot since AMD will be at a 50% clock speed disadvantage
Click to expand...

Again those numbers are with a different compiler which are more optimized and improve performance.

Intel C++ Compiler for IA32/EM64T application,
Version 10.0 - Build 20070308, Package-ID:
l_cc_p_10.0.023
Click to expand...


are you going to make the claim that an optimized compiler isn't going to improve performance. As per the original April numbers those were correct, you and george are doing the same thing, crying AMD is using outdated numbers, but at the same time doing it yourself.
 
(cf)Eclipse said:
calling it vaporware is a little extreme. it's not like infinium who promised a product and never released it. they're still working on it. ;)
Click to expand...


No its not too extreme. ;)

Main Entry: va·por·ware
Pronunciation: 'vA-p&r-"wer
Function: noun
: a computer-related product that has been widely advertised but has not and may never become available
Click to expand...

I'll take Merriam Webster over Wiki any day of the week,month or year. ;)
 
As I Lay Dying said:
are you going to make the claim that an optimized compiler isn't going to improve performance. As per the original April numbers those were correct,
Click to expand...
There was no problem with AMD publishing those figures back in April*. Publishing those numbers in late June on their web site is a problem for 2 reasons:

1) The Spec scores were no longer current. And for a chip set to be introduced in just a few weeks, AMD could do better than estimate performance. AMD instead chose to mislead in the face of superior competition.
and
2) AMD knew full well by that time that the 2.6GHz didn't even have a shipping date.

BTW, complaining about officially submitted Spec scores with full disclosure is a bit silly. I mean, c'mon officially submitted scores beat "estimated" (stalling) scores by a mile. If AMD had something that was all that, they would officially submit Spec scores soon since AMD claims it will launch next month. They aren't, because either the official scores are disappointing, or Barcelona isn't as close to shipping as they claim. :p

* I lied, there is one problem: AMD would have known by April that they had no validation samples of the Barcelona 2.6GHz chip for servers. Server chips are not just slapped into important servers at the last moment, they undergo months of validation in chip manufacturer's and customer's testing labs. Ask yourself why the 1.9-2GHz speed hasn't budged in months for the "August/September" launch. :D
 
Looks like another attempt for AMD to get some hype going on there new processors so when they lauch them in the next couple weeks more then 100 people will buy them:D :D

Of course I was kidding at the end, but seriously I think AMD really needs to stop making these claims. How things are looking they arnt going to hold true. Hopefully im wrong. Dont get me wrong, I am rooting for a good launch but it doesnt look like it.
 
Anyone have any idea if these are gonna be produced in AM2 sockets or the new AM3?
 
quadnad said:
EDIT: I just reread the article and it turns out that only lost them about 6 months. Considering that they released K8 in 2003, that doesn't really explain why we're where we are right now, but I think it gives us an idea that they weren't completely sitting around.
Click to expand...

I think you underestimated the amount of resources that goes into processor developement. Do you realize that it takes about 250-300 engineers, ~3-4 years of time, and several hundred million dollars to develop a new processor? 6 months wasted development time is QUITE significant.
 
AMD in its present form is finished if they can't make this processor sell, and I don't see how they can if it tops out at 2 GHz while Intel's bottom-end can usually hit 3. Looks like buying ATI is turning out to be a huge mistake. Had they not bought them (ATI is also recovering from a late, inferior product), they would have had enough cash on hand to weather the storm until they can develop something worthy of taking on Intel. Now it seems almost certain that they will go bankrupt and change management, perhaps exiting the high performance CPU segment altogether and going back to the low priced, low performance segment. If that happens, Intel will still have competitive pressures since AMD can always come back years later if Intel is slacking off, and IBM can always enter the desktop market, but it's still not as good for consumers as the "two giants duking it out" model we have now.
 
morfinx said:
I think you underestimated the amount of resources that goes into processor developement. Do you realize that it takes about 250-300 engineers, ~3-4 years of time, and several hundred million dollars to develop a new processor? 6 months wasted development time is QUITE significant.
Click to expand...

This was according to that anandtech article, and (if I remember correctly) an AMD employee himself said that they lost about 6 months worth of development time. Whatever resources it took, it's obvious that they weren't suddenly pushing the throttle to the floor to make up for lost time.
 
quadnad said:
This was according to that anandtech article, and (if I remember correctly) an AMD employee himself said that they lost about 6 months worth of development time. Whatever resources it took, it's obvious that they weren't suddenly pushing the throttle to the floor to make up for lost time.
Click to expand...

Yes 6 months of development time was lost. What I'm saying is, one does not simply develop a processor in a short amount of time. And I am an AMD employee.
 
They only lost 6/48 of the time or 1/8 of the total time from then till now (A64 launch to now)... And if they could use NOTHING of what they were working on during those 6 months for the new chips... That still isn't that huge of a setback time wise. Hasn't the chip already been delayed more than 6 months?

Money wise I can't say, but time wise, sure doesn't seem like anything huge.
 
pxc said:
There was no problem with AMD publishing those figures back in April*. Publishing those numbers in late June on their web site is a problem for 2 reasons:

1) The Spec scores were no longer current. And for a chip set to be introduced in just a few weeks, AMD could do better than estimate performance. AMD instead chose to mislead in the face of superior competition.
and
2) AMD knew full well by that time that the 2.6GHz didn't even have a shipping date.
Click to expand...

Ummm the original advertisement was relatively accurate when it was made. Also Intel has done the same thing, they have reports where they match quads vs AMDs dual core, so yeah, I am not saying AMD is right but this is something both AMD and Intel do all the time. ATI and Nvidia do it as well, the only thing AMD did is use ads that were really out of date considering they were not shipping 2.6GHz chips, and there were newer xeons scores.

BTW, complaining about officially submitted Spec scores with full disclosure is a bit silly. I mean, c'mon officially submitted scores beat "estimated" (stalling) scores by a mile. If AMD had something that was all that, they would officially submit Spec scores soon since AMD claims it will launch next month. They aren't, because either the official scores are disappointing, or Barcelona isn't as close to shipping as they claim. :p
Click to expand...

I am not complaining about the spec scores, I am saying that the newer xeons scores are done with a newer complier that are more optimized, of course AMD's scaled estimated scores are going to be lower considering they used an older compiler. Its like benchmarking an ATI card and an Nvidia card and saying that the ATI card won, because its 3dmark 01 score was higher then the Nvidia 3dmark 07 score.

* I lied, there is one problem: AMD would have known by April that they had no validation samples of the Barcelona 2.6GHz chip for servers. Server chips are not just slapped into important servers at the last moment, they undergo months of validation in chip manufacturer's and customer's testing labs. Ask yourself why the 1.9-2GHz speed hasn't budged in months for the "August/September" launch. :D
Click to expand...

fair enough....... but like I said this is common practice by AMD, Intel, Nvidia, ATI, to harp on it now and only on AMD is a joke, and shows how biased you are about one company.
 
This amount of time before Conroe's release, Intel was inviting Anand for some personal time with an E6600. AMD is spitting out benchmarks of estimated performance at the projected 2.6 GHz clock speed.

Not saying that Barcelona is crap, all I'm saying is that current signs don't look good.

Reminds me a lot of the eve of the Prescott release. Prescott was going to be the savior of all the kids with P4Cs when the Athlon 64 had just come out. First thing we heard was that it'd run hotter. Then there were some preliminary benchmarks showing Prescott to be slower than Northwood. Well that just couldn't be, why would Intel make a new processor slower than its old ones? Well the NDA lifted and the performance was indeed worse. Okay, fine, but it'll scale higher. Okay, fine, but it's really going to perform well on LGA775. There were murmurs that the performance gain from scaling clock speed was better than linear on Prescott's design, so it would really open up at higher frequencies - didn't I just read someone say something similar about Barcelona, in response to DailyTech's 1.6 GHz benchmarks? Anyway, let's see AMD match a P4 at 4 GHz and beyond! Except the Prescott never quite made it to 4 GHz. It was hotter, it was slower, and it was less efficient than both AMD's design at the time, and Intel's previous design. In retrospect, we can say it was a dud by comparison, despite the baseless speculation of those who said otherwise.

Is Barcelona going to be a good processor? It'd be hard for AMD to screw that up. Even Prescott had its place. If it runs Windows within a reasonable thermal envelope, OEMs will integrate it and Joe Average will buy it.

Is Barcelona going to win back the performance crown? Is it going to "frag" Intel? If not Barcelona, then maybe some future stepping/derivative with higher clockspeed/IPC? No clue. All I know is that current signs are not good.

I'll see you guys when the NDA is lifted.
 
You must log in or register to reply here.
Back
Top