Sign in with
Sign up | Sign in
Your question

AMD Says Intel Nehalem Microarchitecture Copied Their Design

Last response: in CPUs
Share
August 17, 2008 1:14:29 AM

AMD Says Intel Nehalem Microarchitecture Copied Their Design

Today, AMD’s top CPU and GPU executives, Randy Allen and Rick Bergman, covered a number of key industry topics that are sure to get lots of attention during IDF in a conference call with the press today. They disclosed the usual AMD news about upcoming products, but this slide took the cake. AMD included the slide as a joke and mentioned that imitation is the greatest form of flattery!!

Posted by Nathan Kirsch | Fri, Aug 15, 2008 - 12:50 PM

http://www.legitreviews.com/news/5090/
August 17, 2008 1:30:20 AM

haha...thats great.
August 17, 2008 1:41:00 AM

Intel stole something ??? My faith in corporate ethics is severely shaken.... I think I am going to have to sell my Enron stock now *sobs*




The preceding post has been a joke thank you for reading.
Related resources
August 17, 2008 2:02:57 AM

look, we've gone over this with thunderman already - intel and AMD have both copied each other at various points in their histories, get over it, go away and stop trolling
August 17, 2008 2:11:08 AM

no1 really cares anymore hahaha
August 17, 2008 2:15:58 AM

Does that mean AMD can finally release a cpu that matches Intel's line? :na: 
August 17, 2008 2:23:51 AM

I was waiting for this one. Not a suprise.

Intel provides the USA Gov with products, so they can do what they want.
August 17, 2008 2:45:57 AM

dagger said:
Does that mean AMD can finally release a cpu that matches Intel's line? :na: 



No that would mean Intel copied AMD's designs they clearly state that Intel stole them!!! I hope AMD's designers can remember at least bits and pieces of the work so they can at least make something competitive!!!

Edit: oops it does say copy LOL

Well I guess if AMD does release the same chip the benchies should be a mirror image and at that point it will break down to price ?
a b à CPUs
August 17, 2008 3:03:12 AM

Funny slide though.

a b à CPUs
a b À AMD
a b å Intel
August 17, 2008 3:42:51 AM

Funny slide indeed.

There's something I don't get though: if AMD is indeed ahead of Intel when it comes to processor efficiency, how come the Phenom 9850 is listed as 125W while the Q6600 claims only 95W? Both processors are based on 65nm, and they trade blows in benchmarks. If anybody can explain it, I'd appreciate it.



August 17, 2008 3:49:42 AM

L3 clatch at 2MB would account for some but not alot.
August 17, 2008 4:14:12 AM

@ jimmysmitty

300 dollars for a processor then will be 150 in a year. And you don't think they aren't going to screw you over on the x58 chipset? I will be shocked if it is anything less then two hundred dollars.
August 17, 2008 4:20:49 AM

only ppl with a lot of money on there hands will jump on nehalem when it first launches, as most ppl know, wait for a few months until later revisions come out then jump on board with nehalem. Its not that much different right now, processors still range into the 300 and if nehalem is about that price it isn't half bad, but ddr3 sticks will be expensive although prices are starting to drop. How exactly is Intel going to screw ppl over on the x58 chipset when it is cater to the enthusiast market and not mainstream?
August 17, 2008 4:29:14 AM

Lol slide is funny
But Don't forget AMD has also "Copied" Intel's ideas before.
This is nothing but Propaganda for AMD Fanbois
August 17, 2008 4:31:06 AM

shadowthor said:
How exactly is Intel going to screw ppl over on the x58 chipset when it is cater to the enthusiast market and not mainstream?


Maybe "screwing people over" isn't the best choice of words, but I think the point being made is that Intel will likely drastically overcharge for the chipset. Intel has somewhat of a bad record regarding performance chipset as of late. You may remember that some months ago when the X48 chipset was released several review sites found that it was little more than a rebadged X38 with the FSB speed set a little higher at the factory, and yet it cost significantly more.
a c 127 à CPUs
a b À AMD
August 17, 2008 7:07:36 AM

E3210 said:
@ jimmysmitty

300 dollars for a processor then will be 150 in a year. And you don't think they aren't going to screw you over on the x58 chipset? I will be shocked if it is anything less then two hundred dollars.


Of course it will drop in price. Everything drops in price. Look back at AMDs Athlon X2 back when it was kicking the bejebus out of Pentium D. They were priced very high. Then Core 2 came and the prices dropped. Thats alwasy how it goes and its due to it actually costing less for the company to make them.

Second part, Intel charges a certain amount for the chipsets yes but its not going to be $200 dollars. The chipsets are cheaper than that. Its the mobo makers who decide the final price of the mobo.

I don't care whats going to happen with Nehalems release cost wise because I am waiting for Westmere, the 32nm shrink of Nehalem. My reasoning is because it will allow more cores, run cooler and also improve on the original Nehalem design.
a b à CPUs
August 17, 2008 9:29:40 AM

good points there.

Actually the first lot of 939 X2s were damn expensive (never mind the FX
s were just priced in the stupid range) and didn't drop much in price for ages.

My 4400+ was $700 new ...

The core2 competition pushed the AMD prices down quickly.

And even helped them get rid of their excess inventory of D's too ... a bit of "halo" effect I imagine.

Now we need Deneb to be a halway decent product to ensure Nehalem is cheap too.

Still ... anyone with a reasonable X2 or any E6xxx series or better processor with a decent graphics card can play any games currently out there now ... at the lower resolutions anyway.

We need the applications to make the most out of the extra cores ...

Otherwise the hardware is just going to sit idle.

The programmers need to catch up.

Wouldn't it be nice if the OS made decent use of the extra cores too ... I don't see the third and fourth cores doing much at all most of the time here.

a b à CPUs
August 17, 2008 10:57:25 AM

^ must hurt if you can get 6000+ for ~$100 these days =(

That's why I always buy extremely old high performance tech =) Got my Athlon XP during the Athlon 64 age =)

jimmysmitty said:
Of course it will drop in price. Everything drops in price. Look back at AMDs Athlon X2 back when it was kicking the bejebus out of Pentium D. They were priced very high. Then Core 2 came and the prices dropped. Thats alwasy how it goes and its due to it actually costing less for the company to make them.

Second part, Intel charges a certain amount for the chipsets yes but its not going to be $200 dollars. The chipsets are cheaper than that. Its the mobo makers who decide the final price of the mobo.

I don't care whats going to happen with Nehalems release cost wise because I am waiting for Westmere, the 32nm shrink of Nehalem. My reasoning is because it will allow more cores, run cooler and also improve on the original Nehalem design.


I like waiting until the nm is half the amount of my previous processor... Good way of knowing when to upgrade... Also most programs aren't even optimised properly for dual-core let-alone quad-core...

Looks like I'll be keeping the Athlon X2 for some time =)
August 17, 2008 1:17:34 PM

anyone remember thunderman ... may be this could be his reincarnation . BTW Even they copied from amd they are doing it right at the moment, taking down all the benchies ...
August 17, 2008 4:18:45 PM

Quote:
have we come so far that we cannot even have a laugh and joke and take the slide in the manner it was meant?


i++


August 17, 2008 9:47:11 PM

aevm said:
Funny slide indeed.

There's something I don't get though: if AMD is indeed ahead of Intel when it comes to processor efficiency, how come the Phenom 9850 is listed as 125W while the Q6600 claims only 95W? Both processors are based on 65nm, and they trade blows in benchmarks. If anybody can explain it, I'd appreciate it.


I think it might have something to do with AMD using SOI and Intel went with Hi-K or something like that (I could be utterly wrong here) Also AMD has to power its on die memory controller so who knows how many watts that uses ?
August 17, 2008 10:06:29 PM

JonathanDeane said:
Also AMD has to power its on die memory controller so who knows how many watts that uses ?


Important point that is almost always ignored.
a b à CPUs
August 17, 2008 10:53:22 PM

According to Wikipedia (Whether it's accurate) Core i7 (Nehalem) will have a TDP of 130 watts.
http://en.wikipedia.org/wiki/Intel_Core_i7

And also at the Intel Pressroom whitepaper for Nehalem written there is mention of a Dual Processor 16 thread Nehalem platform(Probably another Skulltrail).
a b à CPUs
August 18, 2008 7:12:37 AM

jj463rd said:
According to Wikipedia (Whether it's accurate) Core i7 (Nehalem) will have a TDP of 130 watts.
http://en.wikipedia.org/wiki/Intel_Core_i7

And also at the Intel Pressroom whitepaper for Nehalem written there is mention of a Dual Processor 16 thread Nehalem platform(Probably another Skulltrail).

Sounds about right, given that it is a completely new architecture with a fairly large die. I'd expect the low end 2.66GHz model to use quite a bit less than that though.
a c 127 à CPUs
a b À AMD
August 18, 2008 2:12:48 PM

jj463rd said:
According to Wikipedia (Whether it's accurate) Core i7 (Nehalem) will have a TDP of 130 watts.
http://en.wikipedia.org/wiki/Intel_Core_i7

And also at the Intel Pressroom whitepaper for Nehalem written there is mention of a Dual Processor 16 thread Nehalem platform(Probably another Skulltrail).


We will probably get the full details from Intel in the next month or so. But I would think this is the TDP for the 3.2GHz version and we will have to wait and see if its like the QX9650 that also has a TDP of 130w but only hits that at about 3.6-3.8GHz.
a b à CPUs
a b À AMD
a b å Intel
August 18, 2008 2:46:08 PM

@ JonathanDeane: Thanks! I knew there was something wrong with the direct comparison between those CPUs, I just couldn't figure out what. Now it makes more sense.

TBH I don't care care if the Nehalem consumes 100W or 130W. I'll get one next year regardless of TDP :)  My video cards will most likely need close to 600W anyway, so 30W more or less won't matter. I'll let IT admins who have lots of servers worry about CPU TDPs.
a c 127 à CPUs
a b À AMD
August 18, 2008 3:20:34 PM

^You plan on running 2 4870X2s? Cuz man 600w is a lot......
a b à CPUs
a b À AMD
a b å Intel
August 18, 2008 3:29:56 PM

I have no idea what I'll run. Whatever will sound best next summer. Most likely something that doesn't exist yet. I'll finish the mortgage payments by then and I want to celebrate it somehow :) 
August 20, 2008 8:20:12 AM

enigma067 said:
AMD Says Intel Nehalem Microarchitecture Copied Their Design

Today, AMD’s top CPU and GPU executives, Randy Allen and Rick Bergman, covered a number of key industry topics that are sure to get lots of attention during IDF in a conference call with the press today. They disclosed the usual AMD news about upcoming products, but this slide took the cake. AMD included the slide as a joke and mentioned that imitation is the greatest form of flattery!!

Posted by Nathan Kirsch | Fri, Aug 15, 2008 - 12:50 PM

http://www.legitreviews.com/news/5090/



first this crap is 2 year old - who copied who? does amd not have an Intel licence to build x86?

what is a cpu but a bunch of on and off switches, amd does this to boost their stock from time to time


intel is shrinking all computer parts, if amd had a decent design they be in first not last, all things will be in the cpu. it will be apu -- all process unit

after all we all know that amd did kind of copy there design and fixed it!
August 21, 2008 6:12:23 PM

IMC
Hypertransport
Native-quad core die
x86-64 instructons
Native dual/quad core designs

As the AMD engineering team aquired the guys from DEC who designed alpha 64 fine and their designs were bought by AMD....kind of a fine line on some very key developments, but the undisputable accomplishments are

Alpha designed the first 64bit cpu in 92, the EV7 which the barton chip is largely based off of was the first chip with an IMC which was around in 99, the EV8, which never launched due to DEC being bought out was the first chip capable of similtanious multithreading, was finished and ready to launch more than a year before the P4, actually the p4 was meant to launch around the same time but strangely when DEC went under, p4 was delayed for near 2 years and showed up with multi-threading

Hypertransport was around since the launch of the 500mhz socket A chip, part of the Dec Alpha bus, they held it back until opteron since intel had motherboard manufactures in their pocket.

Was responsible for DDR memory standard

First to hit 1ghz

Barton was a 64bit RISC cpu, optimized for 32bit code

Was ready to launch a 32bit barton with IMC, but held back for opteron since intel was in no way capable of beating them to the punch

wrote the x86-64 arch instructions, which intel pay for

first to 64bit cpu...with IMC....and Hypertransport

First to produce a dual-core cpu...with 2 cores on a single die

First to produce a native DX10 gpu.

First to use a native quad core design

First to use DDR5

First to single core 1 teraflop gpu performance
between the launch of the p3 in 99 and the launch of Core arch in the end of 2006 intels biggest accomplishment was making the highest power consumption desk-top chip, with the worst memory perforamnce, worst cpu performance...but achieving this feat while managing to claim the highest clock speed. So bad that the 2.8 ghz P4's were even in performance of their own 1.8ghz P3's

What was the last innovation from intel that AMD copied? yes, intel was forced to give up their monoply of x86 cpu's since um...other wise chips aren't compatible....and AMD gave them access to x86-64, which they still couldn't make use of thus Core 2's 32bit reign began.

the only thing intel design original in nehalem is the 45nm fab process and 3 way memory grouping. It took them 10 years to get hypertransport, 5 years and change for IMC, still discount 64bit, as well as virtualization. Oh and they're more than a year away from offering a gpu capable of the 2004 x850xt PE's performance, 20% of present single gpu power.

But i'm bored and tired and intel copies, shrinks than mass produces, they're out of things to copy now though.
August 21, 2008 6:27:10 PM

dragonsprayer, I agree with everything you said but the if AMD had a decent cpu. They do have decent ones, at stock, close to Intels, and the 1% of 1% that oc, to them the AMDs arent worth it, tho theres some as well.
a c 127 à CPUs
a b À AMD
August 21, 2008 8:01:33 PM

iocedmyself said:
IMC
Hypertransport
Native-quad core die
x86-64 instructons
Native dual/quad core designs

As the AMD engineering team aquired the guys from DEC who designed alpha 64 fine and their designs were bought by AMD....kind of a fine line on some very key developments, but the undisputable accomplishments are

Alpha designed the first 64bit cpu in 92, the EV7 which the barton chip is largely based off of was the first chip with an IMC which was around in 99, the EV8, which never launched due to DEC being bought out was the first chip capable of similtanious multithreading, was finished and ready to launch more than a year before the P4, actually the p4 was meant to launch around the same time but strangely when DEC went under, p4 was delayed for near 2 years and showed up with multi-threading

Hypertransport was around since the launch of the 500mhz socket A chip, part of the Dec Alpha bus, they held it back until opteron since intel had motherboard manufactures in their pocket.

Was responsible for DDR memory standard

First to hit 1ghz

Barton was a 64bit RISC cpu, optimized for 32bit code

Was ready to launch a 32bit barton with IMC, but held back for opteron since intel was in no way capable of beating them to the punch

wrote the x86-64 arch instructions, which intel pay for

first to 64bit cpu...with IMC....and Hypertransport

First to produce a dual-core cpu...with 2 cores on a single die

First to produce a native DX10 gpu.

First to use a native quad core design

First to use DDR5

First to single core 1 teraflop gpu performance
between the launch of the p3 in 99 and the launch of Core arch in the end of 2006 intels biggest accomplishment was making the highest power consumption desk-top chip, with the worst memory perforamnce, worst cpu performance...but achieving this feat while managing to claim the highest clock speed. So bad that the 2.8 ghz P4's were even in performance of their own 1.8ghz P3's

What was the last innovation from intel that AMD copied? yes, intel was forced to give up their monoply of x86 cpu's since um...other wise chips aren't compatible....and AMD gave them access to x86-64, which they still couldn't make use of thus Core 2's 32bit reign began.

the only thing intel design original in nehalem is the 45nm fab process and 3 way memory grouping. It took them 10 years to get hypertransport, 5 years and change for IMC, still discount 64bit, as well as virtualization. Oh and they're more than a year away from offering a gpu capable of the 2004 x850xt PE's performance, 20% of present single gpu power.

But i'm bored and tired and intel copies, shrinks than mass produces, they're out of things to copy now though.


1. What ATI has accomplished has nothing to do with AMD. AMD didn't work on those ATI did.
2. AMD has not used DDR5 as it doesn't exist yet. ATI is using GDDR5 and GDDR is not the same as DDR
3. You are such a blind sheep that its funny. You follow AMD blindly believeing everything they do as if they are your best friend.

BTW HyperTRansport is NOT AMDs. Its part of the HyperTransport Consortium that TOGETHER has been developing the HTT links. Just like how AMD is in a consortium with IBM and 9 other companies JUST to get to 32nm HiK.
October 14, 2008 5:04:38 PM

it is usually a great idea to upgrade CPUs only when the manufactoring process is at half the size of your current generation, but there are exceptions: the Core 2 launch was such a massive event that i jump on the first bus , paying $220 for an e4300: and i was very happy and surprised by the performance of the entire system, and cool temps/lower fan noise of the beast at full load! But i became unhhapy pretty fast: the 975X chipset was suboptimal for the true CPU potential, and by the time a good chipset with great overcloking potential hitted the shelfes my 65nm was already old tech : 45nm parts were amazingly cooler and quieter, so instead of wayting for an overpriced and bugged nehalem i decided to jump on an e8600/P45 combo, probably the last and best dual core ever. when we see good and stable nehalem mobos/chipsets and 32nm CPus i will upgrade again, by that time most games will make use of 4 or more cores.

I really hope that AMD can put competitive products on the CPU market, but i just dont see that happening before the company is kicked out of the market...maybe ATI can survive as a pure GPU producer, but AMd is doomed by waay too many years with slower and hotter processors.
October 14, 2008 6:13:45 PM

enigma067 said:
AMD Says Intel Nehalem Microarchitecture Copied Their Design

Today, AMD’s top CPU and GPU executives, Randy Allen and Rick Bergman, covered a number of key industry topics that are sure to get lots of attention during IDF in a conference call with the press today. They disclosed the usual AMD news about upcoming products, but this slide took the cake. AMD included the slide as a joke and mentioned that imitation is the greatest form of flattery!!

Posted by Nathan Kirsch | Fri, Aug 15, 2008 - 12:50 PM

http://www.legitreviews.com/news/5090/


Less the fact that Intel released the Core uArch which didnt need a IMC to take the performance crown, this is the next logical step for the Core uArch. Next thing will be Bell Labs claiming everyone copied their transistor design, its retarded and unessary.

Word, Playa.
a b à CPUs
October 14, 2008 8:06:52 PM



Next we will be hearing that Intel will follow AMD's lead and release a bodged chip which is actually more advanced but somewhat more slower than the competition...


For god sake give it a rest already...

here we go again another round and round in circles conversation.


Open door, gazed around and walked straight out again..
October 14, 2008 8:10:14 PM

iocedmyself said:
IMC
Hypertransport
Native-quad core die
x86-64 instructons
Native dual/quad core designs

As the AMD engineering team aquired the guys from DEC who designed alpha 64 fine and their designs were bought by AMD....kind of a fine line on some very key developments, but the undisputable accomplishments are

Alpha designed the first 64bit cpu in 92, the EV7 which the barton chip is largely based off of was the first chip with an IMC which was around in 99, the EV8, which never launched due to DEC being bought out was the first chip capable of similtanious multithreading, was finished and ready to launch more than a year before the P4, actually the p4 was meant to launch around the same time but strangely when DEC went under, p4 was delayed for near 2 years and showed up with multi-threading

Hypertransport was around since the launch of the 500mhz socket A chip, part of the Dec Alpha bus, they held it back until opteron since intel had motherboard manufactures in their pocket.

Was responsible for DDR memory standard

First to hit 1ghz

Barton was a 64bit RISC cpu, optimized for 32bit code

Was ready to launch a 32bit barton with IMC, but held back for opteron since intel was in no way capable of beating them to the punch

wrote the x86-64 arch instructions, which intel pay for

first to 64bit cpu...with IMC....and Hypertransport

First to produce a dual-core cpu...with 2 cores on a single die

First to produce a native DX10 gpu.

First to use a native quad core design

First to use DDR5

First to single core 1 teraflop gpu performance
between the launch of the p3 in 99 and the launch of Core arch in the end of 2006 intels biggest accomplishment was making the highest power consumption desk-top chip, with the worst memory perforamnce, worst cpu performance...but achieving this feat while managing to claim the highest clock speed. So bad that the 2.8 ghz P4's were even in performance of their own 1.8ghz P3's

What was the last innovation from intel that AMD copied? yes, intel was forced to give up their monoply of x86 cpu's since um...other wise chips aren't compatible....and AMD gave them access to x86-64, which they still couldn't make use of thus Core 2's 32bit reign began.

the only thing intel design original in nehalem is the 45nm fab process and 3 way memory grouping. It took them 10 years to get hypertransport, 5 years and change for IMC, still discount 64bit, as well as virtualization. Oh and they're more than a year away from offering a gpu capable of the 2004 x850xt PE's performance, 20% of present single gpu power.

But i'm bored and tired and intel copies, shrinks than mass produces, they're out of things to copy now though.


Umm unless I am mistaken, Barton was a 32bit socket A chip. The first 64 bit chip was the S754 Clawhammer. They were not the first DX10 either. That was Nvidia. The 8800's came out before the first DX10 ATI which was the @2900's.
October 14, 2008 9:19:58 PM

To argue that Intel copied AMD on native die design, IMC, and hypertransport is like arguing Toyota copied Mercedes on putting four wheels on a car, and two headlights in the front.

Get over it. AMD copied Intel, and Intel copied AMD.
!