• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Unveils "Barcelona" Architecture

Caporegime
Joined
6 Dec 2005
Posts
37,858
Location
Birmingham
AMD Unveils "Barcelona" Architecture





AMD "Barcelona" die shot.


5918_large_4909_large_phenom.jpg









AMD's Opteron and upcoming Phenom logos.


5926_amd_opteron_phenom_logoing.png
 
I'm actually quite excited by this... been waiting for ages for AMD to make the CPU market interesting again. Athlon had gotten very long in the tooth tbh.

Some of the features of the architecture look really quite cool as well.

Just have to ask... these are 200FSB processors? And if so, how does that work with 1.7GHz, 1.9GHz etc? Half multis? And do half multi's have any kind of adverse effect on performance?
 
This is going to be the most anti-climatic launch ever. I'm sure these chips will scale well but the initial batch will be released at frequencies which won't even bring them close to the Conroes.

Read on the Inq a few days ago that there were some good-clocking samples of the second or third stepping doing the rounds, but they won't get to retail till November.
 
Each processor core can dynamically adjust its clock speed depending on load too. The new power management features allow Barcelona quad-core processors to operate with the same thermal envelope as current dual-core Opteron processors.

:eek: Sure it's been mentioned in passed threads but that's really clever - should give AMD another big edge in the corporate/server market where power/efficiency is king! Too bad it'll probably be the first thing we turn off when we get our hands on them! :D
 
^^ and thats why I'm waiting for the phenom to be released, if they stay in current opteron thermal range that means they will stay under 50C in my sugo while giving me 2 extra cores, something the current g0 q6600 would not be able too :)

The question is are they faster than penryn as they look like the temps could be lower than conroe
 
I hope that this kicks Intel's ass. They have done it before. :)

Not going to make me change to AMD though, no need at all but it will be good for everyone.
 
I hope that this kicks Intel's ass. They have done it before. :)

Not going to make me change to AMD though, no need at all but it will be good for everyone.

Hopefully, will give Intel an incentive to get going again, and actually develop Nehalems memory archetecture to be worth anything, my dad's AMD 6000+ beats my E6400 @ 3.25GHZ with quite a bit in memory bandwidth, which is just plain stupid.
 
Hopefully, will give Intel an incentive to get going again, and actually develop Nehalems memory archetecture to be worth anything, my dad's AMD 6000+ beats my E6400 @ 3.25GHZ with quite a bit in memory bandwidth, which is just plain stupid.

Simple solution get a quad or overclock it higher.
stickouttounge.gif



Then laugh at your dad.
 
Simple solution get a quad or overclock it higher.
stickouttounge.gif



Then laugh at your dad.

He did say in memory bandwidth, I'm sure his COnroe beats the Athlon in actual real-world applications!:D But in memory bandwidth the Intels just can't compete at all due to the Athlon's on-die memory controller. I read a benchmark someone linked to in another forum where they measures memory throughput and performance on C2Ds clocked at 3GHz with the memory gradually being overclocked independently of the cpu clockspeed, and they found that beyond 400MHz memory speed there was no visible difference either in bandwidth performance because the Conroe's bus was getting congested.
 
Hopefully, will give Intel an incentive to get going again, and actually develop Nehalems memory archetecture to be worth anything, my dad's AMD 6000+ beats my E6400 @ 3.25GHZ with quite a bit in memory bandwidth, which is just plain stupid.

Might beat it in memory bandwidth but it will beat it in not much/nothing else.
 
He did say in memory bandwidth, I'm sure his COnroe beats the Athlon in actual real-world applications!:D But in memory bandwidth the Intels just can't compete at all due to the Athlon's on-die memory controller. I read a benchmark someone linked to in another forum where they measures memory throughput and performance on C2Ds clocked at 3GHz with the memory gradually being overclocked independently of the cpu clockspeed, and they found that beyond 400MHz memory speed there was no visible difference either in bandwidth performance because the Conroe's bus was getting congested.

Well having the memory bus on-die is still a bit of an "ace in the hole" for Intel. When they do decide to add it, AMD are going to have to come up with something else to give them an "edge".

Then again I did hear AMD wanted to get the whole Northbride onto the CPU next - that's all the PCIe lanes, and a direct, high speed link to the Southbridge. Sounds pretty cool. :)
 
I think the Fusion is going to be a modular design that can include multiple FPUs, GPU, northbridge, and a variety of other things according to the application. It sounds cool that an OEM could ask AMD to make them a chip that can do x, y and z and they get it all on one neatly-packaged die, but I have no idea where AMD will get the manufacturing capability to have multiple versions of their chips being produced simultaneously and be able to convert the manufacturing process rapidly enough to meet many different orders for differently-specced chips in a short amount of time! The only possibility is if I completely misunderstood the concept :p

Unless of course they pull an Nvidia and produce the same chip for everyone and simply cut any features the client isn't paying for out using a laser! ;)
 
Back
Top Bottom