45nm Deneb samples out to partners

P

POPEGOLD

[H]ard|Gawd
Joined
Dec 18, 2005
Messages
1,117
Warning FUDZILLA RUMOR WARNING

FUDZILLA is reporting 45nm Phenoms are already in the hands of its partners.

H2 08 is the release date....so thats a pretty big window...

Inq is also saying it clocks well over 3.0ghz....

again FUDZILLA WARNING.... take with a grain of salt... just posting this for speculative discussion purposes....

edit: I MEANT FUDZILLA....
 
That nice for quad cores if its true. If this is true the duel cores should be right up with the E8XX0 series.
 
xBanzai89 said:
That nice for quad cores if its true. If this is true the duel cores should be right up with the E8XX0 series.
Click to expand...


its hard to imagine AMD dropping the ball on this too.... i think it be the turnaround we all been looking for.
 
Let's hope these rumors are true, and let's hope that 45nm is as much of a success as the 55nm is for their video cards. High yields and high clockspeeds here we come! Even if they can't compete with the performance of Intel, the improved yield rate should allow AMD to become a lot more profitable.
 
pxc said:
I don't see it on the inquirer, but I did see it at fudzilla: http://www.fudzilla.com/index.php?option=com_content&task=view&id=5791&Itemid=1

Oh boy, does it get 30,000 3DMarks and have reverse HT too? :D

If there's a ball to drop, AMD will do it even it if involves finding a ball first. :p
Click to expand...

I think AMD having samples of 45nm Phemons that clock over 3ghz... is more believable on face value than RHTT or 30K 3dmarks

EDIT i meant FUDZILLA... i get the 2 rumor mills confused
 
Fudzilla didn't say the samples clocked over 3GHz, but it is with 45nm that AMD expects to take K10 over 3GHz (snicker).

There are a couple of points in that article that make it suspect, as in either a non-story or as a typical truth impaired posting.

Out of the fingers of Fuad comes the electronic diarrhea of the Internet. :p
 
pxc said:
Fudzilla didn't say the samples clocked over 3GHz, but it is with 45nm that AMD expects to take K10 over 3GHz (snicker).

There are a couple of points in that article that make it suspect, as in either a non-story or as a typical truth impaired posting.

Out of the fingers of Fuad comes the electronic diarrhea of the Internet. :p
Click to expand...

Yup, he's right. To original poster, all they said was that Deneb was AMD's hope to get the k10 to 3+ Ghz. Didn't actually say it was clocking that fast.
 
I'm hoping for AMD to get this working right. We need competition back in the market. I love my current pc, but I still remember my first Pc's Athlon XP 1800+. Was what made me love pcs =]
 
Well if some of the 65 nm cpus can OC to 3.0 it should be very possible they can introduce 3.0 45nm parts question is only if those will be ultra high end AMD chips or more of middle end quads at those speeds.
 
pxc said:
If there's a ball to drop, AMD will do it even it if involves finding a ball first. :p
Click to expand...

:D AMD has all their hands full of balls right now. Whether they drop them or to squeeze and suck it up, and get back on track is key.
 
Hey, alteast you didn't post a link to the site so he couldn't get ad impressions from this forum, your learning well.

I really want that news to be true however. but its faud, he makes shit up on the fly some days.
 
After the disappointment of Phenom, I ain't holding high hopes for this. Surprise me, AMD. :)
 
pxc said:
Fudzilla didn't say the samples clocked over 3GHz, but it is with 45nm that AMD expects to take K10 over 3GHz (snicker).
Click to expand...

while i agree with the rest, i'm tired of your trolling. 3ghz is a much more realistic goal than 10ghz with netburst :D
 
I really hope AMD can pull on this time. I also fell in love with buiding pcs back with the Athlon Series. Love my AM2, but I hope amd can get back with Intel (Also an Intel fan :D).
 
(cf)Eclipse said:
3ghz is a much more realistic goal than 10ghz with netburst :D
Click to expand...
It all depends on the goal. Shipping 45nm would be a pretty big step for AMD at this point. :p

Just to point expectations in the right direction, AMD had K10 samples back at the end of 2006 and we know how it's doing so far. AMD just got back 45nm samples 3 weeks ago. You do the math. ;)

You might be on to something comparing K10 to Netburst, but I won't say it. :) Does it have something to do with running hot and underperforming?
 
(cf)Eclipse said:
3ghz is a much more realistic goal than 10ghz with netburst :D
Click to expand...

lol

while I agree with you there, I still think we're all missing the point here. Why don't we assume the best case scenario, and that is 45nm yields are excellent for AMD and nothing is holding Deneb back from hitting 3-4ghz consistently.

By the time this happens, who will have the cost advantage? Will it be enough to conquer the basic 45nm quads Intel will have available in the next month or so, let alone Nehalem which will be on the horizon?

At least for me, those are the questions I really want answered.
 
pxc said:
It all depends on the goal. Shipping 45nm would be a pretty big step for AMD at this point. :p

Just to point expectations in the right direction, AMD had K10 samples back at the end of 2006 and we know how it's doing so far. AMD just got back 45nm samples 3 weeks ago. You do the math. ;)

You might be on to something comparing K10 to Netburst, but I won't say it. :) Does it have something to do with running hot and underperforming?
Click to expand...

I read somewhere that AMD was having alo of problems with the k10 on the 65nm process....and that 45nm would fix alot of said problems. If thats true ... i could see AMD would have shifted resources to getting 45nm done.... thus the B2 core TLB bug slipped through...and B3 is slow to hit the market....

I think AMD has a skeleton crew pushing the 65nm K10 just to have something shipping... but the bulk of its limited resources are on 45nm development.

makes sense to me...AMD knows it has no shot against WOLFIE,YORKIE and NEHALEM with 65nm K10's...... with 45nm Denebs it can stay in the game.
 
POPEGOLD said:
makes sense to me...AMD knows it has no shot against WOLFIE,YORKIE and NEHALEM with 65nm K10's...... with 45nm Denebs it can stay in the game.
Click to expand...

Easy there tiger, whose getting ahead of themselves? ;)

I'd be shocked if Deneb matched a Kentsfield clock for clock. Current Phenoms are ~10% slower per clock. When was the last die shrink that yielded a 10% gain in IPC? Like, never. :p

You have to learn to crawl before you learn to walk. :p
 
harpoon said:
Easy there tiger, whose getting ahead of themselves? ;)

I'd be shocked if Deneb matched a Kentsfield clock for clock. Current Phenoms are ~10% slower per clock. When was the last die shrink that yielded a 10% gain in IPC? Like, never. :p

You have to learn to crawl before you learn to walk. :p
Click to expand...

It's not just a die shrink. Amongst other things, the L3 cache is tripled (on Phenom, no sure what AMD is doing with Deneb).
 
POPEGOLD said:
its hard to imagine AMD dropping the ball on this too.... i think it be the turnaround we all been looking for.
Click to expand...



If they do,then the end is certainly nigh.I really,really hope that AMD gets thier shit together and gets these 45nm chips out on time,and in large quantity with decent yield.

:)
 
AJL said:
It's not just a die shrink. Amongst other things, the L3 cache is tripled (on Phenom, no sure what AMD is doing with Deneb).
Click to expand...

I don't think that a larger cache would do much for AMD CPU. They are going with a larger cache because it will look good to an average joe who looks for a system at BestBuy. AMD need to attract more mainstream sales to be competitive again.
 
alg7_munif said:
I don't think that a larger cache would do much for AMD CPU.
Click to expand...

I didn't really want to point this out because I thought it might get some fans defensive, but I agree. The architecture just doesn't need that kind of cache simply because of the IMC.
 
Every day performance the cache probably wont do much, but in benchmarks it will help (who cares right?). Will also help them look better on paper compared to intel. You dont wanna be the guy with the little cache.
 
harpoon said:
Easy there tiger, whose getting ahead of themselves? ;)

I'd be shocked if Deneb matched a Kentsfield clock for clock. Current Phenoms are ~10% slower per clock. When was the last die shrink that yielded a 10% gain in IPC? Like, never. :p

You have to learn to crawl before you learn to walk. :p
Click to expand...

i didnt say deneb would fly out the gate and smite intel... i am saying they have to get the die size down to 45nm to compete with INTEL's powerful lineup.

once they get decent clocking denebs out...and the line is running at good yields...then AMD can focus on Bulldozer aka the k11

AMD needs a new arch to beat back intel.....
 
Jakalwarrior said:
Every day performance the cache probably wont do much, but in benchmarks it will help (who cares right?). Will also help them look better on paper compared to intel. You dont wanna be the guy with the little cache.
Click to expand...
There seems to be many things in K10 that are sub-optimal. It will be interesting what tweaks AMD managed to put into the 45nm shrink. Interesting because if anything major is tweaked, AMD would have known the K10 weaknesses long ago while it was still talking smack, full well knowing it wasn't a competitive part.
 
POPEGOLD said:
AMD needs a new arch to beat back intel.....
Click to expand...

They dont have to beat them on the top end ;) if they are on the same process and getting good yields, they can compete in the mid - low range dollar per dollar and watts / performance. Not many people buy top end chips, being top dog is just marketing. This could preserve AMDs existence. Maybe possibly even help them pay off some debt.
 
Bigger cache to woo the average Joe? Most people don't even know what Ghz their chip is running at, let alone how much CACHE it has. I doubt it, I think there would have to be another reason. That being said, I know very little about cpu architecture and even less about the k10.
 
quadnad said:
Will it be enough to conquer the basic 45nm quads Intel will have available in the next month or so, let alone Nehalem which will be on the horizon?

At least for me, those are the questions I really want answered.
Click to expand...

most likely not. anything competitive, performance-wise, from amd would be a shock to me, especially when the last year is considered. it's really just an utter disgrace. :rolleyes:

that said, i'm positive amd will have no problems breaking 3ghz with K10 soon :cool:
 
Yeah, just look what happened when AMD went from 90nm to 65nm, sure the processors got more efficient, but they weren't any faster, and they still don't clock as high as the 90nm counterparts.

What saying that this is going to be any different?
 
Simpson5774 said:
and they still don't clock as high as the 90nm counterparts.
Click to expand...

*cough*

i'd like to contest this. on average, the G2 stepping 65nm's are slightly better than the F3 stepping 90nm's... so let's see, 2nd gen 65nm is edging out the 5th or 6th generation 90nm parts :p
 
Simpson5774 said:
Yeah, just look what happened when AMD went from 90nm to 65nm, sure the processors got more efficient, but they weren't any faster, and they still don't clock as high as the 90nm counterparts.

What saying that this is going to be any different?
Click to expand...

well if history always dictates the future...then its a done deal... AMD is cooked
 
pxc said:
There seems to be many things in K10 that are sub-optimal. It will be interesting what tweaks AMD managed to put into the 45nm shrink.
Click to expand...
I haven't heard of anything significantly new other than the larger cache, higher frequency and HT3. If there is any additional tweaking to the core, AMD is being quiet about it. If the core architecture was tweaked to a significant degree we would probably have heard something about it already.
 
(cf)Eclipse said:
most likely not. anything competitive, performance-wise, from amd would be a shock to me, especially when the last year is considered. it's really just an utter disgrace. :rolleyes:

that said, i'm positive amd will have no problems breaking 3ghz with K10 soon :cool:
Click to expand...

"most likely not" is being extremely generous IMO...that, and 3ghz isn't nearly enough..

It's so frustrating seeing these kinds of companies flailing about
 
You must log in or register to reply here.
Back
Top