fidgewinkle

Distinguished
Feb 27, 2007
162
0
18,680
I couldn't post to the comments on the article, because I apparently can't pass the character recognition test. Since I spent a bit of time on that response, I decided to post it here. It is more or less a rebuttal of the idea that Intel adding graphics processing to Nehalem is a sign of them copying AMD.

I can't understand why people are so surprised that Intel is integrating a graphics processor into their microprocessor. They telegraphed this move with the announcement that they had been working on an 80 core vector processing chip. They have obviously been spending R&D dollars on this for quite a while. It is very possible that AMD caught wind of the project and realized that they needed chipset and graphics expertise. Thus, they purchased ATi.

While I don't have any inside information, the evidence doesn't suggest the assertion of this article. One doesn't just throw new processing on a chip in less than two years without having the expertise all lined up ahead of time and having run test chips with experimental blocks on them. Not even Intel.

Intel didn't just pull Core 2 out of nowhere. It was derivative of Pentium M, and it was clear that Pentium M would lead them to dominate low power processing. It is obvious when one compares Core and Core 2 to Turion.

AMDs best chance to catch up with regard to placing vector processing in the microprocessor is to leverage their high-end expertise and catch up in 2010. For Intel, this will all depend upon how much Intel is learning from projects like the 80 core project and their improving graphics processors. For AMD, it will all depend upon how well they can get ATi and AMD to work together, and how well they can take the ATi expertise and translate it into an integrated solution for their microprocessor.
 

Bazukaz

Distinguished
Jul 11, 2006
43
0
18,530
For some reason , everyone has forgotten , that Intel planned to integrate graphics into pentium III processors. The project was canceled because the customers said they don't need this at that time , if i remember well.
 

fidgewinkle

Distinguished
Feb 27, 2007
162
0
18,680
For some reason , everyone has forgotten , that Intel planned to integrate graphics into pentium III processors. The project was canceled because the customers said they don't need this at that time , if i remember well.

Wasn't it supposed to be the natural evolution of the SSE functions added with MMX? I would say that would probably been too fast an evolution and not worked well at that time. Of course, we ended up with probably too little innovation between pentium2 and pentium4, and are finally getting back on track. It makes a lot more sense to have innovative new functions on the microprocessor when you're just adding extra cores to get more processing power. AMD and Intel need to avoid falling into the number of cores trap after falling into the clock speed trap.
 

Bazukaz

Distinguished
Jul 11, 2006
43
0
18,530
For some reason , everyone has forgotten , that Intel planned to integrate graphics into pentium III processors. The project was canceled because the customers said they don't need this at that time , if i remember well.

Wasn't it supposed to be the natural evolution of the SSE functions added with MMX? I would say that would probably been too fast an evolution and not worked well at that time. Of course, we ended up with probably too little innovation between pentium2 and pentium4, and are finally getting back on track. It makes a lot more sense to have innovative new functions on the microprocessor when you're just adding extra cores to get more processing power. AMD and Intel need to avoid falling into the number of cores trap after falling into the clock speed trap.

I totally agree with with you. Adding extra cores only helps if the software is well optimised for multi-cores. However , todays most apps(with some exceptions) don't utilise the multi-processing potencial properly. Maybe this will improve over time , but i doubt that programming firms would invest much more on code writing for performance(what i see , it is usually opposite - there are a lot of poorly written apps , even for single - core CPUs).
Multi - core programming adds quite much complexity - often it is difficult to split an application to several threads.
 

sweetpants

Distinguished
Jul 5, 2006
579
0
18,980
I believe you're talking about this right?

Nehalem will no longer use a FSB but a serial point to point interconnect. Even more revolutionary is the fact that Nehalem will have an integrated memory controller (IMC) and that the number of serial interconnects is variable (Intel's version of "HyperTransport"). Another potentially groundbreaking move is that some Nehalem CPUs will have a GPU integrated (Intel's version of "Fusion"). With an integrated memory controller, new interconnect, and potentially integrated graphics, Nehalem will obviously require a new socket.

- http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=2955&p=3
 

fidgewinkle

Distinguished
Feb 27, 2007
162
0
18,680
How did AMD really take off as a company? 8086 or something like that?

Your post made me wonder about this, and it seems that AMD has been producing clones of Intel chips since the 8080 in 1975. At least that is what wikipedia has to say. I'm surprised that they've been at it for essentially the entire PC era.
 
How did AMD really take off as a company? 8086 or something like that?

Your post made me wonder about this, and it seems that AMD has been producing clones of Intel chips since the 8080 in 1975. At least that is what wikipedia has to say. I'm surprised that they've been at it for essentially the entire PC era.

Yup!
 

hergieburbur

Distinguished
Dec 19, 2005
1,907
0
19,780
And Intel isnt copying anyone. Actually its the reverse.
Right, Also Intel licensed HyperTransport from AMD a few years ago, and the fact that AMD did NOT invent the IMC by any means. Most of us knew this was coming eventually.
 

InteliotInside

Distinguished
Feb 3, 2007
171
0
18,680
Typical AMD fanboys...

They claim that Intel are always copying AMD... :[ Le sigh, they talk as if Intel just started business with Core 2 Duo...

1) Intel's IGP project looks to have started before AMD's Fusion platform, as they're already in the works to produce some silicon for a 2008 projected target. DAAMiT on the other hand won't be getting this project off of the ground until 2010.

2) HyperTransport doesn't belong to AMD. HT has its own consortium and is used on devices other than AMD processors. Intel is not a member of the HT Consortium, and never has been.
The Front Side Bus system has been working fine for years, which is why Intel never dropped it until now, where it starting to become a bottleneck (especially in MP Xeon/Itanium systems).
What happens when your current system becomes a hindrance, you CREATE something new.

3) AMD weren't the first with an integrated memory controller, I can't remember who was, but it wasn't AMD.
Also, Intel already had an integrated memory controller developed in 2000/2001 for the Pentium 4.
http://news.com.com/Intel+preparing+Northwood%2C+Timna+chips+for+2001/2100-1001_3-244356.html
Unfortunately, Rambus flopped, so Intel was forced to go without a memory controller. No wonder NetBurst was always so data hungry...

Is Intel copying AMD?

No.

During the early 1990s and 1980s, AMD not only manufactured processors for Intel and then started making their own processors, which many say were just revisions of Intel's works--The Pentium and the Pentium Pro.

That's as simple as I can paint the picture, without going into more detail.

If anything, AMD have copied Intel ever since they made their own commercial processor.
 

Bluefinger

Distinguished
Mar 10, 2006
531
0
18,980
Technically, its more of a case of everyone copying each other. AMD copies Intel, Intel copies AMD, and they both take whatever they think works from everyone else. So what's the big fuss? I dunno... all that matters is that Intel is bringing out decent products for once, and I hope it continues. Now it's AMD's turn to get their asses in gear.
 

hergieburbur

Distinguished
Dec 19, 2005
1,907
0
19,780
Typical Intel fanboy, pretty much only telling one side of the story, while leaving out the whole story.

Just to clear things up, [*:8b1a902044]Intel did NOT copy IMC from AMD, and has come up with several of their own ideas in the past.[*:8b1a902044]AMD did NOT copy everything from Intel, and have had several of their own original ideas in the past.Anyone claiming that either company doesn't innovate, or that one company exists solely because they copy from the other, is pretty much an uninformed idiot. Note that I am not directing this at you, its just a general comment.
 

billystringer

Distinguished
Oct 25, 2006
95
0
18,630
I couldn't post to the comments on the article, because I apparently can't pass the character recognition test. Since I spent a bit of time on that response, I decided to post it here. It is more or less a rebuttal of the idea that Intel adding graphics processing to Nehalem is a sign of them copying AMD.

I can't understand why people are so surprised that Intel is integrating a graphics processor into their microprocessor. They telegraphed this move with the announcement that they had been working on an 80 core vector processing chip. They have obviously been spending R&D dollars on this for quite a while. It is very possible that AMD caught wind of the project and realized that they needed chipset and graphics expertise. Thus, they purchased ATi.

While I don't have any inside information, the evidence doesn't suggest the assertion of this article. One doesn't just throw new processing on a chip in less than two years without having the expertise all lined up ahead of time and having run test chips with experimental blocks on them. Not even Intel.

Intel didn't just pull Core 2 out of nowhere. It was derivative of Pentium M, and it was clear that Pentium M would lead them to dominate low power processing. It is obvious when one compares Core and Core 2 to Turion.

AMDs best chance to catch up with regard to placing vector processing in the microprocessor is to leverage their high-end expertise and catch up in 2010. For Intel, this will all depend upon how much Intel is learning from projects like the 80 core project and their improving graphics processors. For AMD, it will all depend upon how well they can get ATi and AMD to work together, and how well they can take the ATi expertise and translate it into an integrated solution for their microprocessor.

^
|
|

F@NBOY!!! :roll:
 

tamalero

Distinguished
Oct 25, 2006
1,134
140
19,470
if I remember correctly.....
Intel actually refused to join the HT consortium .
intel as always been a fan of closed technologies, and not open like HT...

intel had a IMC proyect, but they trashed it, it was TIMNA right?

AMD copied intel ( most of their processors where derivatives, like the K5 = intel's 486 on steroids )
but once AMD introduced the athlon, no way its "copiying"
AMD started their own stuff with 3dnow and other instructions too.


and if you read the toms's info, you'll see that they even mention the fact that intel was never so interested in trashing the fsb until they found that no way a fsb can handle the multicore stuff

and to those who claims that intel had IGP in mind... links please?
because as far we know.. intel favors the SELL MORE CHIPS than INCLUDE MORE ON ASINGLE CHIP.

thats how they get money, selling chipsets, mobos, integrated graphics in the mobo and cpus...
 

bumster

Distinguished
Jan 2, 2007
45
0
18,530
I don't think that AMD is copying Intel at all. Did everyone forget about 3DNOW!? 3DNOW! was AMD's step towards processing 3D graphics on the CPU way back when they came out with the K6 line. I feel that this is just another evolutionary step for all CPUs no matter who makes them.
 

CaptRobertApril

Distinguished
Dec 5, 2006
2,205
0
19,780
Don't innovate. Replicate.

It's faster and cheaper! :lol:
Sweet, I'm totally showing my girlfriend that post. Thanks, Cap! I'm also going to tell her you're a real captain with a Ph.D so she'll believe me.

Actually I only have a Master in Bation. :oops:
 

InteliotInside

Distinguished
Feb 3, 2007
171
0
18,680
InteliotInside wrote:
Typical AMD fanboys...

Lol! More like Typical "put product name here" fanboys.

Is there a meaning behind this statement?

Typical Intel fanboy, pretty much only telling one side of the story, while leaving out the whole story.

I was using AMD fanboy logic.

Also, my post wasn't directed to you, so my apologies.

AMD did NOT copy everything from Intel, and have had several of their own original ideas in the past.

Obviously. If AMD had no ideas, then we wouldn't be getting a K8L.