Thursday, July 14th 2011

FX-Series Processors Clock Speeds 'Revealed'

On several earlier articles like this one, we were versed with the model numbers and even possible prices of AMD's next-generation FX series desktop processors, but the clock speeds stayed under the wraps, that's until a table listing them out was leaked. AMD's FX-series consists of eight-core FX-81xx parts, six-core FX-61xx, and quad-core FX-41xx parts, probably harvested out of the Zambezi silicon by disabling modules (groups of two cores closely interconnected with some shared resources). Most, if not all, FX series chips have unlocked multipliers, making it a breeze to overclock them. All chips come in the AM3+ package, feature 8 MB of L3 cache, and 2 MB L2 cache per module.

Leading the pack is FX-8150, with a clock speed of 3.6 GHz, and TurboCore speed of 4.2 GHz, a 500 MHz boost. The next chip, FX-8120, has a boost of close to a GHz, it has a clock speed of 3.1 GHz, that goes all the way up to 4 GHz with TurboCore. This will be available in 125W and 95W TDP variants. Next up is the FX-8100, with 2.8 GHz clock speed, that goes up to 3.7 GHz, another 900 MHz boost. The scene shifts to 6-core chips, with FX-6120, no clock speed numbers were given out for this one. FX-6100, on the other hand, is clocked at 3.3 GHz, with 3.9 GHz Turbo. The FX-4100 is the only quad-core part with clock speeds given out by this source: 3.6 GHz, with a tiny 200 MHz boost to 3.8 GHz. You can see that there is no pattern in the turbo speed amounts specific to models, and hence we ask you to take these with a pinch of salt.

Source: DonanimHaber
Add your own comment

412 Comments on FX-Series Processors Clock Speeds 'Revealed'

#1
TheMailMan78
Big Member
Damn_Smooth said:
So if OBR has an engineering sample, Why would he post false benches?

Am I correct that he was always an Intel fanboy? If I'm not, disregard the following statement.

If so (Deal with me here for a moment, I've been watching a lot of Scooby Doo today.) the only reason he would post false benchmarks would be to protect Intel. If they were worse, he could show them for a laugh. We already know that he is not under NDA.

Now I just need to find someone to unmask.
All engineering sample are "false" as they do not reflect the final product. Its like looking at an unfinished painting. The only reason anyone has a engineering sample is to see how the painting will fit on the wall (or motherboard). Not until the artist (AMD) is done painting will we see the full picture (Properly performing CPU).
Posted on Reply
#2
Damn_Smooth
Pestilence said:
But what is he trying to protect intel from? Sandy Bridge is Mainstream. Intel still has Sandy Bridge E and Ivy Bridge up its sleeve.
SB-E will be priced out of competition and Ivy won't be out for a year. SB and Bulldozer are direct competitors.

And, he doesn't have ES chips of SB-E or Ivy.
Posted on Reply
#3
Pestilence
Damn_Smooth said:
SB-E will be priced out of competition and Ivy won't be out for a year. SB and Bulldozer are direct competitors.

And, he doesn't have ES chips of SB-E or Ivy.
Dominator or whatever his name is has SB-E ES and IB ES chips.
Posted on Reply
#4
seronx
Damn_Smooth said:
So if OBR has an engineering sample, Why would he post false benches?

For the Lulz




Is the in the picture

Am I correct that he was always an Intel fanboy? If I'm not, disregard the following statement.

He is what you would call an Intel Evangelist, he deals in the Intel Drug...I mean Underground CPU trade

If so (Deal with me here for a moment, I've been watching a lot of Scooby Doo today.) the only reason he would post false benchmarks would be to protect Intel. If they were worse, he could show them for a laugh. We already know that he is not under NDA.

He got a pay out if you checked his blog of like 16 intel cpus

Now I just need to find someone to unmask.

The picture is up top
I typed in your posts deal with it!
Posted on Reply
#5
Damn_Smooth
TheMailMan78 said:
All engineering sample are "false" as they do not reflect the final product. Its like looking at an unfinished painting. The only reason anyone has a engineering sample is to see how the painting will fit on the wall (or motherboard). Not until the artist (AMD) is done painting will we see the full picture (Properly performing CPU).
I agree with you 100% on that. But if that ES beats the competitions product already, The finished painting is going to be beautiful.

Of course, I'm just speculating here and I really don't believe my own bullshit, but it's something to think about.
Posted on Reply
#6
Pestilence
Shit i'm too drunk to find screenshots right now BUT 2011 pricing is not set in stone. If BD kills SB intel can drop prices OR just speed up IB
Posted on Reply
#7
Thatguy
Pestilence said:
Shit i'm too drunk to find screenshots right now BUT 2011 pricing is not set in stone. If BD kills SB intel can drop prices OR just speed up IB
Its all about getting that ninja like drop on your enemy.

Stealth.
Posted on Reply
#8
TheMailMan78
Big Member
Damn_Smooth said:
I agree with you 100% on that. But if that ES beats the competitions product already, The finished painting is going to be beautiful.

Of course, I'm just speculating here and I really don't believe my own bullshit, but it's something to think about.
Better or worse its all FUD until I see a bench on TPU.
Posted on Reply
#9
Damn_Smooth
Pestilence said:
Shit i'm too drunk to find screenshots right now BUT 2011 pricing is not set in stone. If BD kills SB intel can drop prices OR just speed up IB
If BD kills SB, whats to say it won't compete performance-wise with SB-E? And how do you know that they can speed Ivy up? Maybe the delay was because they were having problems?
Posted on Reply
#10
Pestilence
Thatguy said:
Its all about getting that ninja like drop on your enemy.

Stealth.
Intel knows its going to be on top one way or another. It also domnimates the market right now with SB. If you had a 2600K and BD beat it in multithreaded apps by 5% but got killed in things like gaming. Would you trade your SB or an BD rig? I sure wouldn't.
Posted on Reply
#11
Damn_Smooth
TheMailMan78 said:
Better or worse its all FUD until I see a bench on TPU.
I couldn't agree more, good Sir.

I'm just letting myself get carried away.
Posted on Reply
#12
Pestilence
Damn_Smooth said:
If BD kills SB, whats to say it won't compete performance-wise with SB-E? And how do you know that they can speed Ivy up? Maybe the delay was because they were having problems?
Performance estimates show SB-E being alittle faster then SB clock for clock because of the increased memory bandwidth thanks to quad channel memory. We know Amd can't compete with IPC so i don't see it beating up on SB-E especially the 6 core.

Night boys. I'm going to gets some pussy from the lady.
Posted on Reply
#13
Damn_Smooth
Pestilence said:
Performance estimates show SB-E being alittle faster then SB clock for clock because of the increased memory bandwidth thanks to quad channel memory. We know Amd can't compete with IPC so i don't see it beating up on SB-E especially the 6 core.
But we don't really know that AMD can't compete with IPC, because we don't know anything at all.
Posted on Reply
#14
seronx
Pestilence said:
Performance estimates show SB-E being alittle faster then SB clock for clock because of the increased memory bandwidth thanks to quad channel memory. We know Amd can't compete with IPC so i don't see it beating up on SB-E especially the 6 core.

Night boys. I'm going to gets some pussy from the lady.
AMD Zambezi has more IPC per module compared to Intel IPC per core

and AMD Zambezi mimics Tri-channel simply do to how many predictors the IMC has
Posted on Reply
#15
TheMailMan78
Big Member
Damn_Smooth said:
I couldn't agree more, good Sir.

I'm just letting myself get carried away.
It all relies on the MC. IF they left it alone then BD will be a failure. IF they souped it up then BD will be competitive.

As its stands now my 8 ball says they tweaked it but didn't really replace it. So if I were a betting man I would say about 20% maybe 30% better then the Phenom II. Not a Sandy killer but not bad ether if the price stays low.
Posted on Reply
#16
seronx
TheMailMan78 said:
It all relies on the MC. IF they left it alone then BD will be a failure. IF they souped it up then BD will be competitive.

As its stands now my 8 ball says they tweaked it but didn't really replace it. So if I were a betting man I would say about 20% maybe 30% better then the Phenom II. Not a Sandy killer but not bad ether if the price stays low.
They didn't leave it alone though look at Llano which has a retweaked IMC it can handle 2100+MHz stock

Zambezi is a brand new IMC
30% higher IMC performance + 20% higher memory clock
Posted on Reply
#17
TheMailMan78
Big Member
seronx said:
They didn't leave it alone though look at Llano which has a retweaked IMC it can handle 2100+MHz stock

Zambezi is 30% IMC performance + 20% higher memory clock
Like I said they tweaked it. They didn't really change it however. As far as I know Llano loves the tight timings just like the Phenom II. Now who's to say BD will have a different controller then Llano? Could be? Maybe? Who knows.

Like I said maybe 20% or 30% faster over all then the Phenom II. Maybe.
Posted on Reply
#18
seronx
TheMailMan78 said:
Like I said they tweaked it. They didn't really change it however. As far as I know Llano loves the tight timings just like the Phenom II. Now who's to say BD will have a different controller then Llano? Could be? Maybe? Who knows.

Like I said maybe 20% or 30% faster over all then the Phenom II. Maybe.
JF-AMD said it will use a new IMC(Valencia/Interlagos)

Not the one in Magny Cours or Istanbul

Instead of beating the dead horse from K7 they finally are getting off that design
Posted on Reply
#19
TheMailMan78
Big Member
seronx said:
JF-AMD said it will use a new IMC(Valencia/Interlagos)

Not the one in Magny Cours or Istanbul

Instead of beating the dead horse from K7 they finally are getting off that design
That doesn't really mean anything. Like I said the Liano still handles memory pretty much the same as the Phenom II. Its no where near on par with Intel.
Posted on Reply
#20
seronx
TheMailMan78 said:
That doesn't really mean anything. Like I said the Liano still handles memory pretty much the same as the Phenom II. Its no where near on par with Intel.
http://www.xtremesystems.org/forums/showthread.php?262409-Looks-like-Bulldozer-will-have-a-new-and-improved-IMC&p=4635932&viewfull=1#post4635932

JF-AMD
Don't you think that if the memory controller could not keep the cores fed that they would caught and addressed that issue?

You should not use the knowledge of existing architectures to try to express how this new one will work.
What he said
Posted on Reply
#21
Damn_Smooth
What I want to know is if Bulldozer will split threads between modules, or if it will fill a module first.

What I mean is, if you have a 4 threaded application, will it use 1 core from each module, or will it use 2 modules?

Does anybody know the answer?
Posted on Reply
#22
TheMailMan78
Big Member
seronx said:
http://www.xtremesystems.org/forums/showthread.php?262409-Looks-like-Bulldozer-will-have-a-new-and-improved-IMC&p=4635932&viewfull=1#post4635932



What he said
Liano is the new tech and it doesn't really do that very well. Thats my point. Unless BD has a different MC then Liano then well.......failsauce.

Damn_Smooth said:
What I want to know is if Bulldozer will split threads between modules, or if it will fill a module first.

What I mean is, if you have a 4 threaded application, will it use 1 core from each module, or will it use 2 modules?

Does anybody know the answer?
You mean like HT?
Posted on Reply
#23
Damn_Smooth
TheMailMan78 said:
You mean like HT?
No, I mean if a game uses 3 threads, will that use 1 module and a core from the next module, or will it use 3 cores from 3 separate modules.
Posted on Reply
#24
seronx
Damn_Smooth said:
What I want to know is if Bulldozer will split threads between modules, or if it will fill a module first.

What I mean is, if you have a 4 threaded application, will it use 1 core from each module, or will it use 2 modules?

Does anybody know the answer?
If you have a 4 thread application it will use 4 cores

TheMailMan78 said:
Liano is the new tech and it doesn't really do that very well. Thats my point. Unless BD has a different MC then Liano then well.......failsauce
Zambezi does have a different IMC

Llano's IMC is Phenom IIs tweaked for CPU/GPU

The GPU has leeway



But, there was a improvement regardless

Damn_Smooth said:
No, I mean if a game uses 3 threads, will that use 1 module and a core from the next module, or will it use 3 cores from 3 separate modules.
Modules don't play a part, so ignore them

it will use 3 cores
Posted on Reply
#25
Damn_Smooth
seronx said:
If you have a 4 thread application it will use 4 cores
Yes, I know that. But which cores will it use?
Posted on Reply
Add your own comment