AMD PHENOM FX(deneb 45nm) VS INTEL COREi7 (nehalem 45nm)

[ultimategpu]

INTEL Corei7

The upcoming Processor "INTEL Corei7 [code name - NEHALEM ] hitting the market on 2008 december
->Special Architecture of 8 threads in the corei7 processor
->the processor is featured with the copy of AMD's phenom X4 architecture L3 cache and diminished L2 cache
->the processor speed from the basic of 2.4ghz to 2.9ghz for the extreme corei7 processor
->it has the increased FSB than core2quad
->Special Overclocking performance as based on the 45nm architecture than core2quad q9550
->Advanced memory controller to the copy of AMD's memory Bandwidth architecture

AMD phenom FX

The upcoming "AMD's Phenom FX [code name - DENEB ] will hit the market on march 2009 est.

->AMD's first 45nm architecture based processor
->Featured with 512MB L1 cache , 2MB L2 cache and 6MB L3 cache
->Special feature ----> phenom FX-80 at 4Ghz and the FX-82 at 4.4GHz (STOCK SPEED)
->featuring the same Hyper transport @2200 to 3000
->Ultra fast direct link between the RAM & the GPU
->Supports fast rendering video for the full HDTV 1080i
->AMD phenomFX-80 & FX-82 can be overclocked to 6ghz to 6.5ghz

Information listed here are from the source of the Intel and Amd engineering sample .

INTEL COREi7

*img401.imageshack.us/img401/2489/corei7940benchmarked02zu1.jpg



AMD PHENOM FX


*img293.imageshack.us/img293/4389/phenomcpuzba4.jpg

*img401.imageshack.us/my.php?image=corei7940benchmarked02zu1.jpg

 

[desiibond]

->Advanced memory controller to the copy of AMD's memory Bandwidth architecture

FYI, Hypertransport is not AMD's architecture. The technology is used by AMD and Transmeta in x86 processors, PMC-Sierra, Broadcom, and Raza Microelectronics in MIPS microprocessors, AMD, NVIDIA, VIA and SiS in PC chipsets, HP, Sun Microsystems, IBM, and Flextronics in servers, Cray, Newisys, QLogic, and XtremeData, Inc. in high performance computing, and Cisco Systems in routers.

HyperTransport Consortium is an industry consortium responsible for specifying and promoting the computer bus technology called HyperTransport. It is led by founding members Advanced Micro Devices, Alliance Semiconductor, Apple Computer, Broadcom Corporation, Cisco Systems, NVIDIA, PMC-Sierra, Sun Microsystems, and Transmeta.

Even before AMD started using this, there were companies like Nvidia and SUN that were already using this technology.

And if you say copy, you should know that AMD is still paying license fees to Intel for using Intel's X86 architecture. so, stop saying these things.

Coming back to topic, Intel's Nehalem based processors are the fastest at the moment. Deneb will be out next year and nobody has any idea how they will perform. So, it will be really stupid if you compare these two.

 

[MetalheadGautham]

this total dumbness. Stop debating about something which is not even out yet.

 

[desiibond]

Lot of members in this forum have respect for AMD and many are waiting for AMD to hit back with a product that would give them performance crown. There was no flame wars in this area and don't start one.

 

[MetalheadGautham]

Lot of members in this forum have respect for AMD and many are waiting for AMD to hit back with a product that would give them performance crown. There was no flame wars in this area and don't start one.

Its not about flame wars. You can't just compare two CPUs when they have not even released. Especially things like overclockability. Thats one thing which can be the biggest lie ever when it shows up in the market. This thread should have started in January, not now.

 

[ultimategpu]

Lot of members in this forum have respect for AMD and many are waiting for AMD to hit back with a product that would give them performance crown. There was no flame wars in this area and don't start one.

Both engineering sample have been tested and reviewed for Overclocking and stress test
->but they are not reviewed completely as you said , because some issues say that Intel brings out the L3 cache from the AMD technology and minimizing their own L2 cache , is this 8MB L3 cache makes makes the corei7 fast .

->As phenom FX is finally decided to launch first two models of FX-80 & FX-84 ,as the stock speed of 4ghz & 4.4ghz , but 4cores & 4threads

->But intel brings 4cores & 8threads, but lower clock speed 2.4ghz

Intel's Cache architecture (corei7)--------->socket 1366
L1 cache ->128Kbps to 256kbps
L2 cache ->512Kbps to 1MB
L3 cache ->6MB to 8MB

AMD cache architecture (phenom FX) -----> AM2 socket with full 3000HT
L1 cache ->512Kbps
L2 cache ->2MB
L3 cache ->6MB to 8MB


These are the specification listed for the upcoming Processor .

AMD will launch the phenom FX with in 2009/feb


By this architecture and clock speed we can assume !!!!

 

[desiibond]

^^ What we can assume. expectation is 4-7% increase in performance over 65nm Phenom's.

Testing was done by AMD only and sample's are not yet given to press for testing.

AMD shouted like anything during Athlon XP and Athlon64 days that speed is not important, architecture is and performance/watt is. I don't care if they launch 4GHz or 4.4GHz processor. I will not put a thought on it until I see the reviews from leading tech benchmarkers. Till then, for me, it's "watch the proceedings".

 

[pc_game_lover2004]

Lot of members in this forum have respect for AMD and many are waiting for AMD to hit back with a product that would give them performance crown. There was no flame wars in this area and don't start one.

but intel still rules and is the best

 

[desiibond]

yes. Intel is ruling the market now. Best example is AMD's market share down from 23% to 17% in one quarter.

It is ATI that is actually keeping AMD alive right now. Big profits are coming from 4xxx series.

 

[ultimategpu]

O yes , then Y did the INTEL want to use the L3 cache, that is already used by AMD X4 phenom????

 

[pc_game_lover2004]

O yes , then Y did the INTEL want to use the L3 cache, that is already used by AMD X4 phenom????

bhai sab ki apni apni marzi

 

[desiibond]

O yes , then Y did the INTEL want to use the L3 cache, that is already used by AMD X4 phenom????

I can't say anything if you make such nonsense arguments.

And L3 cache was introduced by Intel. You don't believe it??

*www.excaliberpc.com/553583/intel-pentium-extreme-edition-3.46ghz.html

That is one P4 Northwood core processor.

Northwood core was introduced in January 2002.

And since you cannot understand how Microprocessors work,

Intel stopped L3 cache and Hyperthreading concept because the core then was not able to fully utilize these two.

Now, Nehalem has superb core, wider bus, lot of bandwidth, they had a chance to try L3 cache and Hyperthreading and it worked wonders. Hyperthreading is giving 30% increase in performance in 3d applications.

And oh, btw, If you want to support AMD, you should mention these

1) Performance/watt
2) x86-64
3) On-die memory controller
4) on-die GPU

First three are thing of past anyways as they are incorporated by many manufacturer and have to wait and watch how Fusion is going to transform the market.

 

[ultimategpu]

I can't say anything if you make such nonsense arguments.

And L3 cache was introduced by Intel. You don't believe it??

*www.excaliberpc.com/553583/intel-pentium-extreme-edition-3.46ghz.html

That is one P4 Northwood core processor.

Northwood core was introduced in January 2002.

And since you cannot understand how Microprocessors work,

Intel stopped L3 cache and Hyperthreading concept because the core then was not able to fully utilize these two.

Now, Nehalem has superb core, wider bus, lot of bandwidth, they had a chance to try L3 cache and Hyperthreading and it worked wonders. Hyperthreading is giving 30% increase in performance in 3d applications.

And oh, btw, If you want to support AMD, you should mention these

1) Performance/watt
2) x86-64
3) On-die memory controller
4) on-die GPU

First three are thing of past anyways as they are incorporated by many manufacturer and have to wait and watch how Fusion is going to transform the market.

So , what the L3 cache is used in the Corei7 ,,, ha ha ha

 

[DigitalDude]

[...]>it has the increased FSB than core2quad[...]

its not FSB anymore its QTPI

[...]

->Featured with 512MB L1 cache , 2MB L2 cache and 6MB L3 cache[...]

[...]
Intel's Cache architecture (corei7)--------->socket 1366
L1 cache ->128Kbps to 256kbps
L2 cache ->512Kbps to 1MB
L3 cache ->6MB to 8MB

AMD cache architecture (phenom FX) -----> AM2 socket with full 3000HT
L1 cache ->512Kbps
L2 cache ->2MB
L3 cache ->6MB to 8MB

[...]

lol

So , what the L3 cache is used in the Corei7 ,,, ha ha ha

dunno what's there to laugh

a stupid thread with outdated info IMO

_

 

[MetalheadGautham]

^^+1

 

[comp@ddict]

Phenom's 4GHz in 1.168VCore feat is what puts me to awe

 

[MetalheadGautham]

Phenom's 4GHz in 1.168VCore feat is what puts me to awe

ANd THATS the feat which put me in serious doubt if Deneb really IS that good. It appears too good to be true. To displace a highly feature rich and enery efficient and high performing architecture like Nehalem, AMD's current Phenom architecture doesn't seem enough in its 65nm state. I can't help wondering how a core shrink can boost its performance to such high levels. But then again, Phenom was once quoted as being "theoritically" a great architecture, but it "practically" failed for some reason. Could this core shrink be the one thing which can make phenom as good as it should be, theoritically ?

 

[ultimategpu]

ANd THATS the feat which put me in serious doubt if Deneb really IS that good. It appears too good to be true. To displace a highly feature rich and enery efficient and high performing architecture like Nehalem, AMD's current Phenom architecture doesn't seem enough in its 65nm state. I can't help wondering how a core shrink can boost its performance to such high levels. But then again, Phenom was once quoted as being "theoritically" a great architecture, but it "practically" failed for some reason. Could this core shrink be the one thing which can make phenom as good as it should be, theoritically ?

Phenom was failure due to the TLB error (models like 9500,9600) ,, so the TLB error was cleared with the models like 9550 , 9650 , 9750 ,9850BE ,9950
but still phenom X4 cannot competite with the core2quad , cause of lower cache memory in phenom ,No overclocking capabilities in phenom X4(65nm so hot)

So , finally DENEB was to be launched with higher clock speed and higher Cache memory to strike the INTELS "nehalem & Bloomfield"

DENEB has 45nm so FX-80 can be overclocked to 6ghz to 6.5ghz & FX-84 can be overclocked to 7ghz (thats really Phenom ), but it uses more power than INTEL corei7

lets see in 2009/feb

 

[MetalheadGautham]

yup. only feb/2009 will tell the TRUE story.

 

[IronManForever]

So , what the L3 cache is used in the Corei7 ,,, ha ha ha

Duh! That is one hel of an argument.

Theres no point in discussing this as desiibond highlighted. The comparision that has been put up is highly biased.