Official Phenom II 4.0+ Chart

[possessed]

Looking at the pics, there is no AM3 (940) only AM2+ 940, the 945 will most likley be the AM3 equivalent of the 940, and my guess is that it will have same clock speed as well..

940 = AM2+ 3.0Ghz Phenom II
945 = AM3 3.0Ghz Phenom II


Maybe to be followed later by a 965FX = AM3 Phenom II @ 3.2Ghz

On a second look, you must be right, when comparing the 920 and the 925, same clock, but am2+ vs am3.

one 965fx sounds nice to be, but does seem to sound so nice on my wallet

another thing, why are am2+ cpu rated 125w and am3 95? less power consumption front the ddr3 imc?

[Throwed]

This slide smells fishy. BTW, what happen to the November release that was rumored to happen? AMD need to release some official new or my money will be going toward another Intel system. All this waiting with no "official" news is getting a tad lame. With the launch being so close there should be something coming from the horses mouth.

[informal]

This slide smells fishy. BTW, what happen to the November release that was rumored to happen? AMD need to release some official new or my money will be going toward another Intel system. All this waiting with no "official" news is getting a tad lame. With the launch being so close there should be something coming from the horses mouth.

Slide is not fishy and is (yet) unofficial since Phenom II hasn't launched yet.As for OCing headroom we have already seen 3.6-4.2Ghz OCs from non retail C1 steppings from MSI engineer over at coolaler's forum who is also a member here(toppc).

As for the launch date,i guess AMD wants a real launch and not slideware on the day they announce it.

[Oliverda]

another thing, why are am2+ cpu rated 125w and am3 95? less power consumption front the ddr3 imc?

I think the AM3 CPUs will get a new C3(?) stepping.

[wolverine]

Need.

Must Have

[Mats]

why are am2+ cpu rated 125w and am3 95? less power consumption front the ddr3 imc?

Because they run at different speeds.

3 Ghz, 125 W, and later 2.8 GHz, 95 W. That's not really a big drop in TDP. They just haven't included the upcoming AM3 BE in that pic yet.
Besides, BE's usually have a higher TDP than locked counterparts.

[mnemonix]

Must Have

WILL have.

[Overvolt_25]

This slide smells fishy. BTW, what happen to the November release that was rumored to happen? AMD need to release some official new or my money will be going toward another Intel system. All this waiting with no "official" news is getting a tad lame. With the launch being so close there should be something coming from the horses mouth.

Are you sure the new Phenoms were suppose to be released and you are not confusing the Shanghai release date with this one? Just curious as I know the server counterparts were just released. I have not seen anything official stating the desktop versions being released this early. Could be wrong though.

[kazgirl]

one 965fx sounds nice to be, but does seem to sound so nice on my wallet

I'll go one further on the theory of the 965FX as well, to cover AM2+ and AM3 variants of the Phenom II Black Edition, as it were:

Possibly siilar to the 7-series chipsets they may follow suit so:

Phneom II 965X = AM2+ Phenom II @ 3.2Ghz with Unlocked Multiplier
Phenom II 965FX = AM3 Phenom II @ 3.0Ghz with Unlocked Multiplier

So any Phenom II model #, with an X - would be an unlocked AM2+ Phenom II.
Any Phenom II model # with an FX on the end would be an unlocked AM3 Phenom II.
Any Phenom II with just a Model # ending with 0 would be a standard locked AM2+ Phenom II
Any Phenom II with a Model # ending with "5" would be a AM3 locked multi Phenom II.

That would work don;t you think...

Phenom II 960 = 3.2Ghz AM2+ Locked
Phenom II 960X = 3.2Ghz AM2+ Unlocked
Phenom II 965 = 3.2Ghz AM3 Locked
Phenom II 965FX = 3.2Ghz AM3 Unlocked

[Daveburt714]

kazgirl: That would just make too much sense....

You looking to fill one of those lost spots in AMD Marketing!

[Throwed]

Are you sure the new Phenoms were suppose to be released and you are not confusing the Shanghai release date with this one? Just curious as I know the server counterparts were just released. I have not seen anything official stating the desktop versions being released this early. Could be wrong though.

There is a thread right here at XS claiming a Nov. release. Though is your friend. Either way it goes, I am not waiting until Feb 09 to upgrade. I'll just go Core i7 and call it a day. AMD :banana::banana::banana::banana::banana:foots around too much for me and my fanboism toward their company has depleted.

[Voodoo²]

I think those mid-low numbers like 940 for a 3.0Ghz part are a really good sign of the potential of this new processors. AMD is leaving plenty of room for better and faster models.

What do you think about those 8xx models with just 6MB cache? How they make it? No L2? 1/3 of L3 disabled?

[Glow9]

My guesses FX 975 hehe
or
old school FX 951, 953, 955, 957 <-- I'm sure fans of AMD from a while back will pick up on these numbers

[Overvolt_25]

just so i am not missing anything, the Phenom 940 will be for socket AM2+ right? And from the looks of it is that the highest we will see for AM2+ with everything else listed being for AM3 that is for 2Q09? Also, i thought i say a post today showing am3 transition starting in 4Q08 and while i dont think that will happen when they release the new 45nm X4's is that also suppose to be the release of the new socket?

Starting to get confused with all the new naming schemes coming out so close together

EDIT: looked it up and see that the 940 is for the AM2+ so no need to answer that one. AM3 is still simply slated for 1Q09 but is there any indication of it being on the Jan8 release date as well?

[ragnarok95]

I cant wait for the 940.

[keithlm]

I cant wait for the 940.

Okay. We'll get one sent out to you tomorrow.

--AMD Tech Support



(Oh well... Well we can always wish...)

[Tony]

I have to admit this Phenom2 info has me more excited than the i7 stuff, i have a feeling AMD have been working hard to bring performance up a little..and remember its platform performance they are working on,you have to couple this CPU with the new boards featuring sb750 and 4XXXGPU to get the full effect.

I can't wait to see initial benches around 4GHZ, where is Macci when you need him?? LOL

Overall this is exciting and something the industry does need IMO

[Nedjo]

hate to be the one to ask the same question when AMD releases a new SKU, but..... I wonder how it will respond to the cold?

the chart showing overclocking under liquid makes me feel warm and fuzzy, but still nothing on the cold-bug problems

oooh you'll be surprised

[Nedjo]

I can't wait to see initial benches around 4GHZ, where is Macci when you need him?? LOL

I'd guess working hard on AOD 3.0

[flopper]

what would be fun if amd would be able to let their chips overclock higher than i7.
That be hilarious.

[Nedjo]

I'd guess that this is the secret behind Denebs high overclocability:


The PFET performance improvement is more dramatic with drive current now up to 660 µA/µm compared with 510 µA/µm on 65-nm transistors. Again, this increased output current is the result of optimized compressive strain for the p-channel device. The new design of the PFET moves the embedded silicon-germanium source/drain regions closer to the channel to maximize the transfer of stress, thereby increasing hole mobility. Although shorter gate lengths are not driving the improvements, it is a reduction in dimensions that allows increased channel stress to provide the performance scaling. AMD 45-nm PFET design reduces the space from embedded silicon-germanium to the channel edge by half.

The transistor drive current for AMD's 45-nm devices is much lower than that of the Intel HKMG transistors. But power consumption is quickly becoming a high priority for server chips. AMD's transistors exhibit very low channel leakage. Our transistor benchmarks indicates that leakage current is less than one-third of the value measured on AMD's 65-nm process. It's also significantly lower than the Intel 45-nm HKMG process. In fact the Ion/Ioff ratio for AMD's PFET is nearly 10 times better than that for the Intel PFET.

[possessed]

Because they run at different speeds.

3 Ghz, 125 W, and later 2.8 GHz, 95 W. That's not really a big drop in TDP. They just haven't included the upcoming AM3 BE in that pic yet.
Besides, BE's usually have a higher TDP than locked counterparts.

I did compare AM2+ 920 2.8GHz vs am3 925 2.8GHz, but the slide says that the am2+ is 125w and am3 95w, that IS the difference I was referring to.


generally from what what I see we've got: XYZ

for X is the series
9 quad
8 quad 6MB total cache (L2+L3), think only 4MB L3
7 tri
6 quad without l3
4 tri without l3
2 dual no l3

Y is performance indicator, the higher the better

Z is socket
0 AM2+
5 AM3

[Hornet331]

I'd guess that this is the secret behind Denebs high overclocability:


The PFET performance improvement is more dramatic with drive current now up to 660 µA/µm compared with 510 µA/µm on 65-nm transistors. Again, this increased output current is the result of optimized compressive strain for the p-channel device. The new design of the PFET moves the embedded silicon-germanium source/drain regions closer to the channel to maximize the transfer of stress, thereby increasing hole mobility. Although shorter gate lengths are not driving the improvements, it is a reduction in dimensions that allows increased channel stress to provide the performance scaling. AMD 45-nm PFET design reduces the space from embedded silicon-germanium to the channel edge by half.

The transistor drive current for AMD's 45-nm devices is much lower than that of the Intel HKMG transistors. But power consumption is quickly becoming a high priority for server chips. AMD's transistors exhibit very low channel leakage. Our transistor benchmarks indicates that leakage current is less than one-third of the value measured on AMD's 65-nm process. It's also significantly lower than the Intel 45-nm HKMG process. In fact the Ion/Ioff ratio for AMD's PFET is nearly 10 times better than that for the Intel PFET.

Imho the overclockability steams for somewhere else.

Intel Pmos driver current is 1070µA/µm, 62% better then AMDs, and if this slides are correct they can reach the same oc on air as york/nehm.

Channle leakage was expected to be low due to SOI.

[Helmore]

oooh you'll be surprised

Are you pretending you know something we don't or do you simply know something we don't?

[Voodoo²]

I did compare AM2+ 920 2.8GHz vs am3 925 2.8GHz, but the slide says that the am2+ is 125w and am3 95w, that IS the difference I was referring to.


generally from what what I see we've got: XYZ

for X is the series
9 quad
8 quad 6MB total cache (L2+L3), think only 4MB L3
7 tri
6 quad without l3
4 tri without l3
2 dual no l3

Y is performance indicator, the higher the better

Z is socket
0 AM2+
5 AM3

I thought the same. But with further study on the tables that is not the case. Look the 910 should be 905 because is an AM3 model and 915 should be a 2.7Ghz AM3 proccesor. There is no consistency across the naming scheme.