Sign in with
Sign up | Sign in
Your question

AMD's Website has Barcelona on it...

Last response: in CPUs
Share
March 3, 2007 7:01:33 PM

The native quad core design. This could help the launch on Monday rumor. Wonder if they will...
www.amd.com
The flash thing is here:
http://multicore.amd.com/us-en/quadcore/
Why would they put this on their site?

More about : amd website barcelona

March 3, 2007 7:15:07 PM

Quote:
You know, a recommendation to AMD --- stop advertising your product on what you competition does... the first thing that hits you is 'FSB is bottlenecked' --- the first thing you see is 'why you shouldn't buy my competitor' rather than 'why you should buy my processor'.

It makes it appear like they are chasing.


You are right... but...

This strategy worked well for AMD 3 years ago.
March 3, 2007 7:29:30 PM

Quote:
You know, a recommendation to AMD --- stop advertising your product on what you competition does... the first thing that hits you is 'FSB is bottlenecked' --- the first thing you see is 'why you shouldn't buy my competitor' rather than 'why you should buy my processor'.

It makes it appear like they are chasing.


You are right... but...

This strategy worked well for AMD 3 years ago.

I would disagree, what worked well for them 3 years ago was producing a better product.

It also contradicts their assertion that they are the industry leader.
None of what you said answered my question.
Related resources
March 3, 2007 7:31:54 PM

Quote:
I would disagree, what worked well for them 3 years ago was producing a better product.


I thought they used similar advertizing tactics 3 years ago in conjunction with a better product?
March 3, 2007 7:32:28 PM

AMD say: look hahaha, Intel have to use huge L2 cache on there chips to beat our slimline efficiant CPU

Customer says: So the Intel CPU is faster then, umm whats cache?
March 3, 2007 7:34:01 PM

Quote:
None of what you said answered my question.


Good point. :wink:

They are simply preparing the way for the introduction of Barcelona. You know... get people talking and interested.
March 3, 2007 7:35:31 PM

I see an extreme lack of diversity in that flash slideshow.
March 3, 2007 8:56:48 PM

It's not news that you are right; however, they're mistakenly trying to react in a situation of stall until Barcelona is out. They should rather do what Intel did; start delivering ESs and showing benchmarks.
March 3, 2007 9:08:34 PM

They need just one more step to freeze almost all the Clovertown buyers; Show some winning benchmarks as soon as possible. Adding to this the fact that Intel chips are going to swap socket sooner rather than later would give AMD a lot of time and valuable money.
March 3, 2007 9:30:51 PM

Quote:
They need just one more step to freeze almost all the Clovertown buyers; Show some winning benchmarks as soon as possible. Adding to this the fact that Intel chips are going to swap socket sooner rather than later would give AMD a lot of time and valuable money.


There are no new sockets until CSI comes some time mid 2008, the 45nm units this year will still use exactly the same sockets. Also AMD are not going to be able to release winning benchmarks, because Intel have an ace up there sleeve in the form of the unreleased, but very available (when Intel deem fit) Xeon 5365, 3GHz quad core part. Intel have demo’d it god knows how many times, they are just waiting till AMD makes a move, then are going to gate crash the party...
March 3, 2007 9:41:05 PM

A 3.0GHz Core2 quad might not be enough if Barcelona tops at 2.6GHz.
March 3, 2007 9:44:36 PM

Jack, seems Intel have stated that VRM 11.0 will last out until LGA775 dies off with CSI, LGA1366 Nehalem CPU's in mod 2008, at this socket is the first to come with VRM 11.1. Knowing that, the only future left for LGA775 is a slight speed boost to 1333MHz, that Asus for one already have stated that most of the current mobos will support without any issues. http://event.asus.com/mb/fsb1333/
March 3, 2007 9:47:49 PM

Quote:
You know, a recommendation to AMD --- stop advertising your product on what you competition does... the first thing that hits you is 'FSB is bottlenecked' --- the first thing you see is 'why you shouldn't buy my competitor' rather than 'why you should buy my processor'.

It makes it appear like they are chasing.


:cry:  I was having a good day until I found out my FSB was bottlenecked. :x

I can OC my FSB to 1960MHz but this doesn't seem to help any.

Oh well, should I just be content with my e6700 running at 3.5 GHz with a bottlenecked FSB or should I buy an AMD X2????

Can someone help me comphrehend the bottlenecked FSB thing?
March 3, 2007 9:51:41 PM

oooooh shiny I want one so bad...imagine if those seti numbers are real.

I love how the flash presentation rips on intel so bad... "we think fsb is GAY and your gonna have to upgrade it again later!"

Maybe the fear of loosing the datacenter segment scared them into an early release
March 3, 2007 9:53:53 PM

penryn will need bearlake! darg i don't wanna find the link
March 3, 2007 10:21:49 PM

Thanks, this would explain why when I go between 3.2GHz with a 1600 FSB vs a 1280 FSB there is no difference. This would also explain why I really see no benefit in OC my Ram from 800MHz to 1.07GHz.

At least I have never seen a difference while playing Windows Solitare. :lol: 

So if the K8s have most this stuff, why do they still suck so bad? j/k

The L3 cache and instructions should really help. Was the direct memory addressing between hardware and CPU cache aimed at Video Hardware Addressing to the CPU. Couldn't this give the K10 a big boost if games/video encoding? Or am I high on crack and misunderstood? Or just high on crack? :lol: 

edit: Could AMD be releasing a DX10 video card that can address a CPU directly??? Could they be making a technology that will only work if you have an AMD K10 and Their video card?
March 3, 2007 10:39:25 PM

LOL, the VAST, VAST majority of customers (or potential ones) don't know anything about FSB, architecture, or the like. They don't know and they don't care. They only know "Core2Duo is FAST and pretty cheap, and is waaaay better than what I got now. I'll buy that!"

We can sit here and wax philosophical about technical nuances, etc but the profits don't come from us.....we're minnows in their sales ocean. C2D is mainstream and big profit/sales. Barcelona isn't/won't be. Such a shame, I wish AMD would bring a mainstream giant-killer to the table and give consumers better choices at better prices. Looks like not.
March 3, 2007 10:43:28 PM

Quote:
I see an extreme lack of diversity in that flash slideshow.


:wink:

Very subtle and polite way of saying they empasized the same 3 points in 12 different ways.
I had to watch it twice to make sure it wasnt just a fast loop.
March 3, 2007 10:44:33 PM

Quote:
It also contradicts their assertion that they are the industry leader.
AMD the industry leader, that's a good one. :lol: 
March 3, 2007 10:51:09 PM

guess you think you're pretty cool eh talking about your overclocking and your crack euphoria. damn man, is it hard being that stupid?
March 3, 2007 11:44:20 PM

[quote="JumpingJack"To be fair.... AMD has lead the industry in areas, the interconnect fabric is well done.... the APM machinery they have prefected overtime is very good ... and the K8/Opteron lead the pack for a long time.[/quote]Being fair isn't fun. :wink:
March 4, 2007 12:41:47 AM

Quote:
they put it up there because they are not thinking straight. now people are going to go to frys and ask to buy the quad core cpu. i wonder what frys will sell them?


:D 

That's what happens when you paper launch.
March 4, 2007 12:41:53 AM

I'm excited. I want to seem some REAL 3rd party benchies very soon. I think this has a LOT MORE potential than Quad FX. I wonder how long until they have an Athlon X4 and quad core Quad FX processors.
March 4, 2007 1:32:33 AM

Hey, has anyone noticed on the flash animation, 64kb L1?
March 4, 2007 1:35:37 AM

Quote:
Hey, has anyone noticed on the flash animation, 64kb L1?


Yes... that is the L1 Data cache.
March 4, 2007 1:47:19 AM

Hmmm....the flash show did not make me think AMD is chasing Intel but...it did absolutely nothing to keep me tuned in or keep me from buying a CPU from Intel. It is a weak Flash show and they really need to release those ES's soon or everyone will start thinking that K10 is not going to be up to the task (they did too much talking about efficiency and not enough about outright performance). It is starting to look like I may switch to the dark side on my next build :( 
March 4, 2007 1:49:31 AM

Quote:
Hey, has anyone noticed on the flash animation, 64kb L1?


Yes... that is the L1 Data cache.It isn't specified if it is a data cache, just L1...
March 4, 2007 1:50:36 AM

Quote:
You know, a recommendation to AMD --- stop advertising your product on what you competition does... the first thing that hits you is 'FSB is bottlenecked' --- the first thing you see is 'why you shouldn't buy my competitor' rather than 'why you should buy my processor'.

It makes it appear like they are chasing.


:cry:  I was having a good day until I found out my FSB was bottlenecked. :x

I can OC my FSB to 1960MHz but this doesn't seem to help any.

Oh well, should I just be content with my e6700 running at 3.5 GHz with a bottlenecked FSB or should I buy an AMD X2????

Can someone help me comphrehend the bottlenecked FSB thing?

It is a perpetuated myth --- at least on the desktop --- with no data to back it up. AMD claims that performance is held back because the FSB is slow....

Here is how it works...

Memory is controlled by the northbridge chipset, at DDR2-800, this provides 12.8 GB/sec of possible BW in dual channel mode to the chipset memory controller (this is theoretical).

The CPU is connected to the chipset with with a 1067 MHz FSB that is 64 bit wide, giving 8.5 GB/s theoretical BW.... so the concept is that the FSB is bottlenecking....

What people overlook or fail to understand is that these BW numbers are just maximum capable, a program chuging along will only demand data at a certain data rate over the bus depending on how fast the processor is crunching it and requesting that data. In reality, normal applications running from memory cache up the instructions and data and only require a fraction of the total maximum theoretical. Say you run quake 4, on a 1067 bus, get 143 FPS. Now say you run Quake 4 on a 400 MHz FSB and you get 143 FPS.... is this showing a bottleneck?? Of course not.... why? Because the CPU is not demanding more data over the bus than what the bus can deliver.

The best example that 1067 is plenty of BW for now is in Tom's initial Kentsfield bench marks, they ran a 2.67 GHz kensfield at 1333 MHz FSB, then upped the multiplier on the CPU to normal and raun at 1067 FSB and did all the bench marks... guess what, no difference.... which means?? No bottlenecks....

It is a falacy perpetuated by AMD, and driven home by people who do not understand how bussing and CPUs interact with data transfer.

Now, there is evidence in server benchmarks for the server platforms that BW can saturate and become a bottleneck... STREAM on a Woodcrest or Clovertown saturates and AMD CPUs shine there.... but on the desktop... there are not many apps (none that I have seen) that run so fast as to saturate the FSB.

Jack

Jack is always right.
March 4, 2007 1:58:33 AM

Quote:
A 3.0GHz Core2 quad might not be enough if Barcelona tops at 2.6GHz.


Barcelona will not top out at 2.6 GHz, AMD doesn't even have a dual core that goes past 2.6 GHz, the 65 nm is not providing the performance boost needed to get there. But that is my speculation based on the data I have seen to date -- they may pull a rabbit out of the hat by June.

think back to september 1999..when after all the specualation over the first ATHLON 500mhz..... FIRINGSQUAD.com got a EX sample...than ran test ...and guess what ...it sucked

few months later....Athlon was released and mopped the floor with Intels KATMAI core..

AMD would love to spring that on INTEL again.... and i think they are slickly leaking out how powerful the K10 is...

i think it's a setup... we are trying to base how K10 will perform based on BRISBANE.... when AMD put a skeleton crew to develop it. The Meat of AMD engineers been working hard on the k10.

like i said i smell a setup.... and AMD might not spring it until JUNE

we been down this road before.
March 4, 2007 2:02:06 AM

Quote:
Hmmm....the flash show did not make me think AMD is chasing Intel but...it did absolutely nothing to keep me tuned in or keep me from buying a CPU from Intel. It is a weak Flash show and they really need to release those ES's soon or everyone will start thinking that K10 is not going to be up to the task (they did too much talking about efficiency and not enough about outright performance). It is starting to look like I may switch to the dark side on my next build :( 


like i told jack.... INTEL was murdering AMD in the dawn of ATHLON

Firingsquad got an ES sample of the first Athlong in september 99

and it sucked

but when released...Athlon killed Intel's Katmai

even an ES sample means little to final silicon.... but if the ES out of the box beats INTEL then that would say alot.... if its loses it really says nothing as its not final
March 4, 2007 2:07:17 AM

Quote:
even an ES sample means little to final silicon.... but if the ES out of the box beats INTEL then that would say alot.... if its loses it really says nothing as its not final


Nice spin.

"If the ES is good, then good for AMD. If it's bad, then I don't want to count it. That way it's good for AMD no matter which way it goes, which is the way Fanboy Facts work."

Doesn't work that way. An ES is pretty damn close to final. If AMD's ESes flop big time, then it won't be good.

Now, I believe that AMD might be able to retake the performance crown. However, we'll have to see how far they'll clock. Intel has PLENTY of headroom to use to beat AMD.

I'm really interested to see how the native with L3 will do. However, since no one has gotten a proc to review if Barcelona is launched soon I have a feeling it will be a paper launch.
March 4, 2007 2:46:58 AM

this paper launch could be the most taxing stress on me....if i don't see some benchmarks i'm gonna explode, and like this dude above me says if k10 doesn't stop the sh!t out of intel in the benchmarks with experimental chips that'll make it even worse cause i'll be hoping amd CAN get their act together to embarass clovertown...

so if this is the bomb shizzile that everyone hopes it to be...and we can get octofx systems with these r600's that can do like a teraflop a piece...and amd opens em up like nvidia did...how many Tflops could a beast like that do? I'm scared....Hold me!
March 4, 2007 3:19:16 AM

Quote:
Hey, has anyone noticed on the flash animation, 64kb L1?


Yes... that is the L1 Data cache.It isn't specified if it is a data cache, just L1...

actually if you read the sidetext, it says " 64KB l1 data cache"
March 4, 2007 1:01:11 PM

I understand your point on AMDs 65nm, still not topping over 2.6G for the K8, but bare in mind that the 65nm is still pretty new for them and they might have turned (almost) all the current 65nm on Barcelona, that's why they're not pushing it for the K8. Moreover, they can't afford to have the old 90nm chips hang around unsold like Intel does.
A lot of people around have OCed Brisbanes to 2.7-2.9G on air but I doubt AMD will be producing any major quantity of them or spend money on further binning because more than an alternative, they were just an experiment until they start making K10s. Their path until the 90nm dumping will logically be K8 exclusively 90nm and K10 exclusively on 65nm.
March 4, 2007 2:13:08 PM

Quote:
Can someone help me comphrehend the bottlenecked FSB thing?


If you had a core2duo notebook with a 667Mhz bus you would understand the bottleneck problem.... Remember notebooks are 50% of the market, so until intel until gets HT/CSI they will not be able to compete in this market once k10 is released.... Penryn will also be riding the short yellow FSB....
March 4, 2007 2:25:42 PM

Bottle necks on Notebooks come not from the processor, but from the weaker RAM, video card and hard drive. Put a Pentium M in a PGA478 motherboard, and you'll see a the difference. Put a Core Duo or Core 2 T Series processor in a motherboard designed for such, and you'll see a difference. Many sites tested the potential of the Core chips this way. They didn't say anything about FSB bottlenecks.
March 4, 2007 2:28:35 PM

quadcore laptops...wow...i haven't even thought of that... with the new super energy efficient and way faster flash drives that will be powering our laptops soon....i think F@H should be required by law to have on your computer...like all new dells get shipped with it...gonna get alzheimers crotchety bed wetting @ss!

think a octofx setup and a r600 will be able to pull above a 6.0 on windows vista index? :lol:  :oops:  :x

maybe amd went to bill gates and told him to build a system that would whore resources like a mofo and gates told him "but i already do that quite well!"
March 4, 2007 2:33:50 PM

i'm sure those chip designers know more that we do about FSB and direct connect...I think the whole point is when you read "intel reaches for the sun" it talked about 32 cores at 32nm on a single die they mentioned clogging the FSB and the various avenues by which to overcome this...

this is another one of those "we computer users can't see any slowdown but servers sure can!"

c2d is a streamlined chip...well designed but its not as super future as k10 is (supposed to be).
March 4, 2007 2:34:16 PM

Quote:
Quote:


Can someone help me comphrehend the bottlenecked FSB thing?

Well it's not that bad in 1 cpu desktop systems but in 2-4+ cpu systems have ram linked to each cpu is better then haveing all cpus use one chip set for it. even more so when each quad-core is 2 duel cores shearing the fsb link.
Also amd system lets you have more them link the to i/o and chip sets look at the nforce pro chips. And you can use left over links for HTX slots.
March 4, 2007 2:41:31 PM

Ninja & jack,

You huys are both right - the intel 667 Mhz FSB does not bottlerneck the cpu per say - I am looking at a notebook as a whole system.... You can tell from everyday use that not having a n integrated memory controller, and HT, etc. makes a notebook slugish.... multitask with things like disk i/o, downloading, burnings dvd's, etc.... And you will be saying to yourself.... yeah the core2duo processor is faster, but I miss my HT1.0s multitasking ability....
March 4, 2007 3:24:11 PM

Its no rumor u TWIT.

AMD is working on even bigger and better things far from which u could conceive.
March 4, 2007 3:25:28 PM

Turion vs Pentium M. Pentium M still wins.
March 4, 2007 3:27:15 PM

"what are we going to do tonite brain?" "the same thing we do everynite TRY AND TAKE OVER THE WORLD!"
March 4, 2007 3:44:20 PM

Quote:
Turion vs Pentium M. Pentium M still wins.


Turion X2 vs core2duo

core2duo better for games/crunching/single apps
Turion X2 better for heavy multitasking

go buy a dell 1505(core2duo) and a dell 1501(Turion X2) and configure them as close as possible.... You will be scrathing your head, wondering why the core2duo isn't kicking the Turion X2s but in everyday use.... Is it the intel chipset?, bus/ht, 64bit operating system, or what?

you(speaking figuratively) can keep your penryn with 667/800Mhz bus/intel chipset and I will keep my k10(2 or 4 core).... let me know when intel dumps it FSB for something better(perhaps nehelum(sp))....
March 4, 2007 3:59:11 PM

i liked when they plugged that raptor into the turion notebook...i freakin wish my laptop had a 10k hd! i mean i probably don't but geez...
March 4, 2007 4:01:29 PM

Thanks Jack,

I am just trying to get people to think about notebooks systems more, since they are 50% of the market now.... I always see people make blanked statement on this forum, like FSB doesn't matter, intel is better, penryn will beat k10, etc...

those statements may or may not be true for desktops - but notebooks are big now too.... And for people who configure 4p servers, they care about integrated memory, bus speeds, power saving, and chipsets just like notebook people....
March 4, 2007 4:32:25 PM

hmmmm, but the server space and mobile are diffrent beasts. No server is going to have a gpu to boost performance. Also,pick to laptops one turion,one c2d,and use the same gpu...now bench them...what do you get?the intel system destroys the amd.
March 4, 2007 4:48:43 PM

You missed the whole conversation.... The point was, there is alot more to a computer than just raw cpu benchmark speed.... chipsets, integrated memory, and fsb, power savings, etc make a big difference when multitasking.... And make a big difference when building a notebook or server...
March 4, 2007 4:49:57 PM

amd should just shrink k10 to 45nm asap and stick a apu on it!! then those 4p tweakers would loose their minds...
March 4, 2007 4:57:27 PM

Quote:
amd should just shrink k10 to 45nm asap and stick a apu on it!! then those 4p tweakers would loose their minds...


I can't wait to see what the fusion APUs can do VS amd/intel CPUs.... I also can't wait to find out if intel will be sooner or later to the APU market....
!