AMD Says Barcelona will be 40% faster than Core 2

Status
Not open for further replies.
L

LstBrunnenG

Supreme [H]ardness
Joined
Jun 3, 2003
Messages
6,676
Link: http://news.com.com/AMD+Go+to+Barcelona+over+Clovertown/2100-1006_3-6152645.html?tag=nefd.top

I can see a bunch of you about to compare this to the early Conroe benchmarks, but those were actual benchmarks run by a third party. This, on the other hand, is from the mouth of AMD with no benches to back it up. Not saying they're lying, not saying their word is gold - simply interesting to hear nonetheless. I wonder if Intel jacks up the clock speeds to well above 3 GHz how that will affect these margins.

Thoughts?
 
I think its more important to get a bloody product out that can compete then having one that much faster than Intels.
 
The sooner they say it, the better they can back it up if Intel bumps up the clock speeds. Few months from now it might not be as much, but they'll say that was the case back when it was published. :p
 
I read that as well.

Based on information that is available today, I dont believe it..... It may be a bit faster in FP and SIMD code, but in day to day use it should be on par or slightly better if at all.

Until they have hard numbers this means absolutely nothing at all.
 
we'll just have to wait and see when AMD has something benchable available. should be interesting though; obviously AMD isn't going to make something available that is only as fast as Core2. Hopefully we'll see the pendulum swing back.
 
.....The quad-core chip also will outperform AMD's current dual-core Opterons on "floating point" mathematical calculations by a factor of 3.6 at the same clock rate, he said.
Click to expand...

Something is fishy about that sentence.
 
MetalX said:
Something is fishy about that sentence.
Click to expand...
Not fishy, probably accurate in certain types of heavy parallel floating point operations. K8L has twice as many floating point units as K8. It's still going to get crushed in integer performance by Intel's quad core CPUs.

Although it might not seem like it, a K8L will probably be memory bandwidth limited for FPU throughput even with dual channel DDR2-800 in heavy streaming applications.
 
Well remember, the Power PC G5 was the world's fastest computer. Maybe AMD took a lesson from Apple's engineering (READ: marketing) division.
 
Well good. I always worry when the underdog isn't doing well. I shudder to think of gaming with a monopoly on the CPU or GPU market.



pxc said:
Not fishy, probably accurate in certain types of heavy parallel floating point operations. K8L has twice as many floating point units as K8. It's still going to get crushed in integer performance by Intel's quad core CPUs.
Click to expand...

We know so little about his chip, how could you even begin to know this?
 
What? 10 posts and still no Intel fan chiming in with "By then Intel will have <insert codename here> which will be faster lawlz!!"

It's interesting to note that the article doesn't specify where that 40&#37; difference lies. It could be between the fastest most expensive Barcelona and the cheapest slowest Intel equivalent. Is it 40% faster per clock? Per price range? These things... I must know!
 
Sabrewulf165 said:
What? 10 posts and still no Intel fan chiming in with "By then Intel will have <insert codename here> which will be faster lawlz!!"

It's interesting to note that the article doesn't specify where that 40% difference lies. It could be between the fastest most expensive Barcelona and the cheapest slowest Intel equivalent. Is it 40% faster per clock? Per price range? These things... I must know!
Click to expand...

But AMD's "Barcelona" quad-core chip, due to arrive midway through 2007, will be a significant notch faster than the Clovertown chips expected to be on the market at that time, said Randy Allen, AMD's corporate vice president for server and workstation products.

"We expect across a wide variety of workloads for Barcelona to outperform Clovertown by 40 percent," Allen said.
Click to expand...

I would guess from that, Allen is refering to equivalent Clovertown chips on the market at that time. In other words, AMD's high end Barcelona chips will be 40% faster then Intel's high end Clovertown chips and so forth on down.
 
duby229 said:
I read that as well.

Based on information that is available today, I dont believe it..... It may be a bit faster in FP and SIMD code, but in day to day use it should be on par or slightly better if at all.

Until they have hard numbers this means absolutely nothing at all.
Click to expand...

I am currently am skeptical as well, I certainly hope AMD does produce a winner with Bareclona though.
 
Well, if it's 40% faster at the same clockspeed, then it looks like things may get interesting again. At least this time, it'll be two fucking incredible products in competition.
 
Sabrewulf165 said:
What? 10 posts and still no Intel fan chiming in with "By then Intel will have <insert codename here> which will be faster lawlz!!"

It's interesting to note that the article doesn't specify where that 40% difference lies. It could be between the fastest most expensive Barcelona and the cheapest slowest Intel equivalent. Is it 40% faster per clock? Per price range? These things... I must know!
Click to expand...

For the 2P server segment, it is not clear when Intel will be replacing Clovertown, as there is currently not time frame for the introduction of Harpertown, the 45nm Server 2P derivative of Yorkfield.

Yorkfield as well as Wolfdale itself is not too clear on introduction timeframe other then H2 2007.
 
Intels responce is clock york to 3.7Ghz and wolfdile to 4ghz to counter K8L after release. Why such a huge jump eh, at the same time K8L comes out? Look at the roadmaps, quincidence? I don't think so.
 
AMD is really stupid if they don't come with anything better to compete with C2D, they were leading Intel since Athlon XP(or even earlier), Athlon 64 and Athlon X2 until C2D came out. What did they do during the time they were still leading and making huge profits with their overpriced X2? Did they just sit back and relax? I really hope not. Now it seems to me that AMD is doing exactly what Intel did during the Athlon XP time, at that time AMD only used socket A and Intel always changed their socket. Now Intel sticks with their socket775 but AMD is always changing their platform :( .
 
alg7_munif said:
AMD is really stupid if they don't come with anything better to compete with C2D, they were leading Intel since Athlon XP(or even earlier), Athlon 64 and Athlon X2 until C2D came out. What did they do during the time they were still leading and making huge profits with their overpriced X2? Did they just sit back and relax? I really hope not. Now it seems to me that AMD is doing exactly what Intel did during the Athlon XP time, at that time AMD only used socket A and Intel always changed their socket. Now Intel sticks with their socket775 but AMD is always changing their platform :( .
Click to expand...

Athlon XP only lead Pentium 4 during part of it's lifetime. Once the Northwoods hit and especially the 800FSB parts hit, Intel had the performance lead until Athlon 64.

Intel "sticking" with socket 775 means practically nothing as just about every new chip requires a chipset change.
 
Sabrewulf165 said:
Athlon XP only lead Pentium 4 during part of it's lifetime. Once the Northwoods hit and especially the 800FSB parts hit, Intel had the performance lead until Athlon 64.

Intel "sticking" with socket 775 means practically nothing as just about every new chip requires a chipset change.
Click to expand...

Thank You.

I been sayin it over and over. It doesnt matter what socket it uses, you'll still have to change the board.
 
Newer chipset can support older CPUs, it is a real pain when your mobo is dead after the warranty period but you still don't need a CPU upgrade yet and you can't find a replacement for the mobo. Even if you can find one the price would be high. I've experienced more dead mobos than dead CPUs.
 
alg7_munif said:
Newer chipset can support older CPUs, it is a real pain when your mobo is dead after the warranty period but you still don't need a CPU upgrade yet and you can't find a replacement for the mobo. Even if you can find one the price would be high. I've experienced more dead mobos than dead CPUs.
Click to expand...

While this is true, it has already been announced that K8L will be supported on AM2, what more do you want? I doubt you'll get any such guarantees from Intel regarding future products.
 
You'll still likely need a new board to take advantage of HT3 support among other new features as well.

Backwards compatibility is the bane of technology... It is the single biggest cause of stagnation.
 
duby229 said:
You'll still likely need a new board to take advantage of HT3 support among other new features as well.

Backwards compatibility is the bane of technology... It is the single biggest cause of stagnation.
Click to expand...

Still it'll be nice for people who want to go K8L without buying new RAM (DDR3) or a new motherboard (if they already had AM2). It's also nice for 754/939 owners who can just skip DDR2 altogether and go straight to AM3. It seems to me like there are a lot of options for people anticipating K8L.
 
Hahaha! If this is true, I'll be able to buy AMD real cheap in a couple of months then sell it a couple of months after they release the new chips. Heehaw!:D

EDIT: Ack! Just noticed clock frequency will be lower. This may mean they will suck in single core applications. We need multi-threaded software now!!!
 
I really hope so. This one sided performance battle sucks. Though I'm going to treat this the same as the preliminary Conroe benchmarks. Until i see real number from a trusted page its all just a big rumor.
 
coolpurplefan said:
Hahaha! If this is true, I'll be able to buy AMD real cheap in a couple of months then sell it a couple of months after they release the new chips. Heehaw!:D

EDIT: Ack! Just noticed clock frequency will be lower. This may mean they will suck in single core applications. We need multi-threaded software now!!!
Click to expand...


I thought the AMD exec 'declined to mention' clockspeeds when asked ?
 
I think is bs. AMD is just trying to please investors after their disapointing earnings report today.
 
gravity said:
I think is bs. AMD is just trying to please investors after their disapointing earnings report today.
Click to expand...

No way. Maybe they're completely oblivious to the plummeting share prices, barrage of downgrades, and market cap that...well let's not go there.

How can anyone doubt these claims? I mean look at 4x4 and brisbane and stuff...like how C2Q really sucks because of the FSB and MCM and such! becauseitdoesiknowitdoeslalalalala!

Good thing they're not showing benchmarks. Heck, with that kind of performance coming in a few short months, who'd want to buy a k8 or a Core 2?

With performance increasing by dramatic leaps and bounds, this is exactly the definition of an evolutionary process.

That was so smart, taking the evolutionary path.
 
I would agree, considering that developing a new revolutionary architecture can take as long as 7 - 10 years to complete, and the last one was NexGens K6 architecture, with the one prior to that being Intel's PPro architecture.....

Since then ALL CPU's that have been released are based on these two systems, AMDs based on K6, and Intels based on PPro.

The time required to develop architectural changes on par with the two systems would be huge, and while I'm certain we will see it on both sides in the next few years, this aint it. It was a good choice based on time to market, and cost initiatives...

My opinion is that Fusion in conjuction with CTM will be that massive architectural shift that will revolutionize the CPU industry.
 
Why not show what benchmarks AMD refered to and end the FUD ? ;)

At the same time as the release of the 4x4 platform, AMD made the demonstration during the Industry Analyst Forum of a server using 4 Quad-Core Opteron and a total of 16 cores.....AMD also said that the Opteron will be compatible with existent Socket F platforms. About performances, AMD announces respectively 13% and 46% improvements with TCP OLTP and SPECfp compared to a Xeon 5355. These performances are however only based on estimations and that means that the chip isn't yet fully functional....
Click to expand...

There you have it.What did they do ? They simulated the performance of what they think will be the top bin of K8L when released vs. the Cloverton 2.66GHz.Where they say " Intel's current" when K8L is released they actually refer to Xeon 5355 since Intel won't release faster SKUs until Q4 ( or at least that's what the roadmaps say ).
 
duby229 said:
I would agree, considering that developing a new revolutionary architecture can take as long as 7 - 10 years to complete, and the last one was NexGens K6 architecture, with the one prior to that being Intel's PPro architecture.....

Since then ALL CPU's that have been released are based on these two systems, AMDs based on K6, and Intels based on PPro.

The time required to develop architectural changes on par with the two systems would be huge, and while I'm certain we will see it on both sides in the next few years, this aint it. It was a good choice based on time to market, and cost initiatives...
..
Click to expand...

You're talking crap as usual and please allow

AMD redesigned the Barcelona core, marking the biggest changes since the company made its 2003 transition from its 32-bit Athlon chips to the current 64-bit lineup. The magnitude of the transition is about halfway between the small tweaks AMD has made to Opteron over the years and the clean-sheet redesign Intel employed in moving from NetBurst to its current Core design, Allen said.
Click to expand...

Who said that ? Randy Allen, AMD's corporate vice president for server and workstation products.
Go eat crow again... :rolleyes:

Link http://news.com.com/AMD+Go+to+Barce...+page+2/2100-1006_3-6152645-2.html?tag=st.num
 
You can make it bold and big all you want but that doesnt make it true.....

"It's a water bridge. You cant see it becouse it's made out of water, the same as the river, but it's really there!"
 
I will take the same attitude I took when Conroe was released: I will believe it when I see readily available retail chips reach verify these claims.
 
gravity said:
I think is bs. AMD is just trying to please investors after their disapointing earnings report today.
Click to expand...
I dont think youre right..it would be a suicide move to announce a 40% faster chip...and then when it comes out have us find out its actually slower :rolleyes: and I do think the fine gentlemen at AMD value their jobs :p
 
Sabrewulf165 said:
Athlon XP only lead Pentium 4 during part of it's lifetime. Once the Northwoods hit and especially the 800FSB parts hit, Intel had the performance lead until Athlon 64.

Intel "sticking" with socket 775 means practically nothing as just about every new chip requires a chipset change.
Click to expand...

Not since Conroe though. I'm Quad compat right now, don't have to wait 7 to 10 months:)

You'll not get anywhere with trying to make the case for Northwoods, I already tried explaining that to him:D

The Socket not changing can be both good and bad! I have an HP 744v and I'm having to run a 2700+ at 2100+ because it only supports 266MHz FSB and no 333MHz support.
 
Status
Not open for further replies.
Back
Top