New 45nm Phenom FX Deneb spotted ~ 4.4GHz...?

All of you [H]s are missing the most important thing here. %$@ clock speed, what was really interesting is the claim that 4ghz phenom fx was as fast as a higher clocked 5ghz penryn. Hopefully this 20% faster per clock magic applies to both single threaded and multi threaded apps but no benchies yet from reputable sites on engineering editions of these "miracle" chips so unti then this could easily be a lot of bs.
 
Mav451 said:
Regardless, Ghz as a benchmark alone is obsolete (nice seeing you Netburst :D) - the A64 saw to that. IPC relative to Ghz, that's what actually matters.
Click to expand...

So false. Both ipc and ghz are needed.

What good is a cpu that can do 2000 ipc but can only scale up to 1mhz, if you put it up against a cpu that does 1 ipc but can scale up to 5ghz?

People took the amd logic and went way way way to far to the left side of the spectrum. You need both the raw hp mentality that Intel brought to us along with the value of ipc that AMD brought us.

To much of anything, in regards to any argument, is always bad.
 
DeadSkull said:
All of you [H]s are missing the most important thing here. %$@ clock speed, what was really interesting is the claim that 4ghz phenom fx was as fast as a higher clocked 5ghz penryn.
Click to expand...
The site claimed 5GHz Kentsfield, not Yorkfield.

Anyways, the Deneb has already been benchmarked and clock for clock it was only about 7% faster on average than Agena. That's not bad for what even AMD calls a "tweak." Bumping up the NB speed could increase that a little more, but you're not going to see a 20% bump in clock for clock performance vs Agena. A 20% improvement (if met) would need to include the clock speed increases over the top Agena model (9950).

That site is shens anyways.
 
DemonDiablo said:
People took the amd logic and went way way way to far to the left side of the spectrum. You need both the raw hp mentality that Intel brought to us along with the value of ipc that AMD brought us.
Click to expand...

Actually it's not AMD's logic. Since the 8086, Intel has always strived to improve IPC as well as clockspeed.
It's just that nobody noticed until the Pentium 4 for the first time did not improve IPC.
AMD's Athlon was more or less matched in IPC with Intel's Pentium 2/3. So while Athlon looked good in terms of IPC compared to the Pentium 4, it's not like it was new.
 
Scali2 said:
Actually it's not AMD's logic. Since the 8086, Intel has always strived to improve IPC as well as clockspeed.
It's just that nobody noticed until the Pentium 4 for the first time did not improve IPC.
AMD's Athlon was more or less matched in IPC with Intel's Pentium 2/3. So while Athlon looked good in terms of IPC compared to the Pentium 4, it's not like it was new.
Click to expand...

True, I probably should've worded that better for historical accuracy but when on these forums there arent many whose memories tend to reach that far back or history of the tech for that matter.

Most forum browsers equate the whole industry trend of ipc over raw power as AMD's doing from the AXP days and as such I was just dummying down my post to that level.
 
And as pxc said, nobody is talking about the 4ghz > 5ghz because that was obviously a load of crap. If any part of that rumor has any, if even a small tinkle in the eyes, chance of being true its that AMD has a 4ghz proc. Whether that be stock or oc, its about the only part of that quote worth debating on.
 
Welp here's a current agena Phenom at roughly 4Ghz, AMD debuted the engineering samples originally at 3Ghz a long time ago (if I recall correctly) before they launched them at a much lower Ghz.

http://www.xtremesystems.org/forums/showthread.php?t=199643

So maybe they just took some time to play it safe and stabilize the platform/corner the market at the lower price segment. We all know with overclocking or processors on the "outer edge" sometimes it just takes the right combo of hardware/settings/voltage to unlock some more headroom. We'll see what comes.
 
DemonDiablo said:
So false. Both ipc and ghz are needed.

What good is a cpu that can do 2000 ipc but can only scale up to 1mhz, if you put it up against a cpu that does 1 ipc but can scale up to 5ghz?

People took the amd logic and went way way way to far to the left side of the spectrum. You need both the raw hp mentality that Intel brought to us along with the value of ipc that AMD brought us.

To much of anything, in regards to any argument, is always bad.
Click to expand...

You did notice I said "relative" to Ghz? As in "in respect to". It's like you just skimmed over my post.
 
I seriously doubt a die shrink could be responsible for making that much of a difference in clock speed scalability and overall performance.

This sounds like FUD to me.
 
Dan_D said:
I seriously doubt a die shrink could be responsible for making that much of a difference in clock speed scalability and overall performance.

This sounds like FUD to me.
Click to expand...

FUD is used to denigrate a competitors product. When you are attempting to make your product appear better than it is, its called BS

:)
 
OldPueblo said:
Welp here's a current agena Phenom at roughly 4Ghz, AMD debuted the engineering samples originally at 3Ghz a long time ago (if I recall correctly) before they launched them at a much lower Ghz.

http://www.xtremesystems.org/forums/showthread.php?t=199643

So maybe they just took some time to play it safe and stabilize the platform/corner the market at the lower price segment. We all know with overclocking or processors on the "outer edge" sometimes it just takes the right combo of hardware/settings/voltage to unlock some more headroom. We'll see what comes.
Click to expand...

you forgot to mention they used liquid nitrogen.
 
Monkey God said:
you forgot to mention they used liquid nitrogen.
Click to expand...

I didn't forget at all, I linked the thread with all the info. That's not the point, the point is processors are being pushed further and further all the time. Sure the info in it's current state can't be trusted, but it's not an automatic impossibility. Change is constant in this industry, and surprises do happen.
 
OldPueblo said:
I didn't forget at all, I linked the thread with all the info. That's not the point, the point is processors are being pushed further and further all the time. Sure the info in it's current state can't be trusted, but it's not an automatic impossibility. Change is constant in this industry, and surprises do happen.
Click to expand...

Shoving a rocket engine on the back of pickup truck doesn't mean next years model pickup trucks can fly either.
 
Monkey God said:
Shoving a rocket engine on the back of pickup truck doesn't mean next years model pickup trucks can fly either.
Click to expand...

Perhaps next years model pickup trucks can hit 5Ghz using liquid nitrogen. I know it's hip to be negative, but until we know, nobody knows.
 
I completely agree that this is most likely fake,

~ BUT ~

I reserve the right to say "I told you so" just in
case it isn't... ;)

:p :p :p
 
I hope this is true, but I doubt it. The inquirer is the one who said there would be a 3GHz Phenom and that it would outperform Intel's Core 2, and that never happened.
 
I wouldn't be surprised if Deneb had a lot more going on for it. Remember how quiet AMD was with 4870?

They spread those fake rumors about 320 shaders and they totally undersold it. When it came out, everyone was taken back by the 800 shaders and the price/performance.

Although, on the other hand, rumors were flying about the GPU hitting 1Ghz, and those were false.

Seems like AMD's marketting department has new tricks, and if they do have something bad-ass, they're not going to let the competition know until it's too late for them to do anything in response. They just kinda hit them with surprise when they don't expect it.
 
look im not saying its impossible, im just saying I wont believe it till I see it.
I hope AMD can do it. I've been an AMD lover since my first AMD486 I bought way back, but the core duo was just too good to pass up. (And I owned one of the infamous 300a celerons as well)
 
You have a very good point there, sdlvx...

AMD/ATI have been doing that lately... Slipping
false info that's both too high and too low at the
same time... Keeps their competitors off guard...

At least it keeps everyone guessing and constantly
thinking/talking about AMD which is probably the
main goal from their point of view... ;)
 
I just don't think AMD would drop so many 65nm K10 parts (like Kuma) to spend time with Deneb, only to add more cache and have a die shrink. I'd imagine little tweaks like that wouldn't take so long.

But that site does look like shit. Visit the main page, it looks like some kids blog.
 
Sounds like total BS to me. Phenom clocks really badly past 3.0GHz. It has worse IPC than Kentsfield or Yorkfield do, and I am just not buying into the whole "die-shrink and more cache are going to change that" theory.
 
So some new review site found a link to an image on some other obscure site posted by a visitor? Unless I misunderstood... Lemme go home and photchop up the 5ghz version :D

Dont get me wrong, I would love for it to be true, but I would also like to win the lottery.

IF this is true and there is a STOCK Deneb at 4.0+ghz I will post a pic of my own foot in my mouth.
 
Dan_D said:
Sounds like total BS to me. Phenom clocks really badly past 3.0GHz. It has worse IPC than Kentsfield or Yorkfield do, and I am just not buying into the whole "die-shrink and more cache are going to change that" theory.
Click to expand...

Well...I do agree with most things you have said about the Phenom and AMD in the past such as that AMD hasn't even really released an answer to kentsfield. I disagree with this statement though, with the release of the 790GX and more importantly the SB750 most users have been seeing clocks between 3.2-3.4ghz on their 9950's. Some users have even been hitting 3.2ghz on SB600 with the 9950BE but @ much higher voltage than the SB750 w/ ACC has.

It's basically a waiting game at this point. I have the upgrade bug once again but I've realized that I don't really need to upgrade. I purchased and played around with a Q9450 but I don't do enough things that warrant the need for the extra 2 cores. I went back to my E8400 and sold off the quad. I'd like to wait for DDR3 to become the standard before I upgrade again much like I did when I switched from my s939 X2 3800+ to a Pentium D 930. I'd love to switch back to AMD but there's nothing but FUD and rumors at this point with Deneb. What I see myself doing personally is switching from this set-up staight to an i7 and than once AMD switches to socket AM3 I'll pick-up whatever their Phenom offering is should it at LEAST be faster than Penryn.
 
Here are some Deneb benchmarks from not too long ago:
http://xtreview.com/addcomment-id-6100-view-AMD-45NM-deneb-benchmarks.html
(Probably taken from this translated article: http://nl.babelfish.yahoo.com/trans...doc/hard/79405.htm&lp=zt_en&btnTrUrl=Vertalen)

If these benchmarks are real, then it seems rather hard to believe that they can catch a Kentsfield at 5.0 GHz. Also, it makes one wonder why the tested CPU was only running at 3.2 GHz (overclocked), when stock speeds of 4.4 GHz are planned.
 
With the introduction of Deneb, there is supposed to be a HT increase for the motherboards. Maybe that motherboard didn't have the update needed for it, and caused the CPU to slow down. Anything is possible.

The model the reviewer had, may have been a lower model, and even a stepping as well.
 
Sorry but a major leap like that warrants a lot of skepticism since the 65nms were so underwhelming. I'm just gonna cross my fingers and hope my e8400 can o/c to 4ghz instead of getting all hyped about a rumor.
 
TheBlueChanell said:
Well...I do agree with most things you have said about the Phenom and AMD in the past such as that AMD hasn't even really released an answer to kentsfield. I disagree with this statement though, with the release of the 790GX and more importantly the SB750 most users have been seeing clocks between 3.2-3.4ghz on their 9950's. Some users have even been hitting 3.2ghz on SB600 with the 9950BE but @ much higher voltage than the SB750 w/ ACC has.

It's basically a waiting game at this point. I have the upgrade bug once again but I've realized that I don't really need to upgrade. I purchased and played around with a Q9450 but I don't do enough things that warrant the need for the extra 2 cores. I went back to my E8400 and sold off the quad. I'd like to wait for DDR3 to become the standard before I upgrade again much like I did when I switched from my s939 X2 3800+ to a Pentium D 930. I'd love to switch back to AMD but there's nothing but FUD and rumors at this point with Deneb. What I see myself doing personally is switching from this set-up staight to an i7 and than once AMD switches to socket AM3 I'll pick-up whatever their Phenom offering is should it at LEAST be faster than Penryn.
Click to expand...

I've yet to have a Phenom overclocking experience as good as you've mentioned. I'm not saying that those clocks aren't possible or that they don't happen even fairly often, but I have doubts about it being the norm. Even if you take a Phenom and clock it to 3.4GHz it still won't match the Core 2 Duo G0 at 3.6GHz, and certainly won't touch the Yorkfield's hitting 4.0GHz+. So regardless of the fact that overclocking has gotten better for the Phenom they still aren't on par with Intel offerings. I still don't think a die-shrink and more cache memory will be enough to change that.

This sounds more like the statements I used to hear about B3 stepping Phenoms being worlds better than B2 stepping Phenoms and we all know how that turned out. While I have no doubts that Phenom is going to get better with more cache and a die shrink I seriously doubt it will close the gap between AMD and Intel significantly. Now that doesn't mean that AMD won't be able to offer any viable alternatives at certain price points with Deneb. I'm sure they will. But for those people buying into this hype that Deneb is going to slaughter Agena, I'm more than likely going to be telling you I told you so when Deneb actually hits the market.
 
lowteckh said:
Sorry but a major leap like that warrants a lot of skepticism since the 65nms were so underwhelming. I'm just gonna cross my fingers and hope my e8400 can o/c to 4ghz instead of getting all hyped about a rumor.
Click to expand...

I don't know about hitting 4.0GHz or more on the E8400. I know people can do it and do it all the time but I haven't been able to do it yet. (This is largely due to the DX48BT2 boards I'm using.) In any case I can say that 3.8GHz is really easy to do. I can almost hit that on stock CPU voltages using my E8400 and I have the older stepping.
 
Hmm, there was a post here something about suggesting that this was probably faked by AMD to pump up Deneb.
I think there is a much much better chance of this being posted by Intel workers or affiliates to over hype this chip and attempt to create a dissapointment of enthusiasts upon release. Too easy.

BTW, are there any Intel Retail Edge Program affiliates in this thread?
 
Dan_D said:
I've yet to have a Phenom overclocking experience as good as you've mentioned. I'm not saying that those clocks aren't possible or that they don't happen even fairly often, but I have doubts about it being the norm. Even if you take a Phenom and clock it to 3.4GHz it still won't match the Core 2 Duo G0 at 3.6GHz, and certainly won't touch the Yorkfield's hitting 4.0GHz+. So regardless of the fact that overclocking has gotten better for the Phenom they still aren't on par with Intel offerings. I still don't think a die-shrink and more cache memory will be enough to change that.

This sounds more like the statements I used to hear about B3 stepping Phenoms being worlds better than B2 stepping Phenoms and we all know how that turned out. While I have no doubts that Phenom is going to get better with more cache and a die shrink I seriously doubt it will close the gap between AMD and Intel significantly. Now that doesn't mean that AMD won't be able to offer any viable alternatives at certain price points with Deneb. I'm sure they will. But for those people buying into this hype that Deneb is going to slaughter Agena, I'm more than likely going to be telling you I told you so when Deneb actually hits the market.
Click to expand...

I am afread that this is true, but its nice to dream that AMD could pull this out of their ass :D

all the same if they even get within the ballpark it will a huge step forward for them
 
Hum the IBM power 6 does 5Ghz and IBM helped with this ! IBM is the BIG DOG WOOF!

Might be BS but might not be !

I expected that from K10 and was sadly disapointed !
 
Dan_D said:
This sounds more like the statements I used to hear about B3 stepping Phenoms being worlds better than B2 stepping Phenoms and we all know how that turned out. While I have no doubts that Phenom is going to get better with more cache and a die shrink I seriously doubt it will close the gap between AMD and Intel significantly. Now that doesn't mean that AMD won't be able to offer any viable alternatives at certain price points with Deneb. I'm sure they will. But for those people buying into this hype that Deneb is going to slaughter Agena, I'm more than likely going to be telling you I told you so when Deneb actually hits the market.
Click to expand...

Yes indeed. Ever since the early rumours of Barcelona I've been amazed at how much people just WANT to believe, even if it defies all logic and common sense.
After the huge disappointment that Barcelona/Agena has been so far, why do people still believe it's going to be a huge success?
 
I wonder what the pricing for the FX models are going to be? Any guesses?
 
Possibly a stupid question, but that cpu-z screenshot shows Core Name as Deneb. Would cpu-z even know how to identify a Deneb?
 
Scali2 said:
Yes indeed. Ever since the early rumours of Barcelona I've been amazed at how much people just WANT to believe, even if it defies all logic and common sense.
After the huge disappointment that Barcelona/Agena has been so far, why do people still believe it's going to be a huge success?
Click to expand...

Well the kicker is I thought I remembered Kyle being somewhat optimistic. I think that's what made me wait on upgrading to Core2, until finally seeing the final hard numbers.
 
You must log in or register to reply here.
Back
Top