Understanding changes in CPU technology (P4 -> C2D)

[andru]

Can anyone direct me to an article that explains the technology difference between P4 and C2D? Back in the day it was all about Ghz... now there seem to be many different factors that affect processor preformance.

I suppose the thing I am finding most confusing is exactly how FSB works, and how a processor running at a lower frequency can preform so much better than my old p4.

Thanks in advance

 

[newls1]

Not to be a smart ass, but try google, its your friend

 

[andru]

Yep, doing self research for sure. I just thought perhaps someone in the community might have a good read they could recomend.

 

[Jakalwarrior]

To make things really simple, it does more work per clock cycle. The old pentiums were set up to run really fast without much regard for efficency. AMD took the alternate path and now intel has changed their mind and followed suit.
I cant remember the exact stuff they said about its development but its pretty much from the ground up and has more in common with the Dothans which were based on P3s than the netburst architecture P4s.

 

[Calefaction]

Basically the difference is pipeline size. The P4 used a very long pipeline, but cranked up the juice to the CPU to make it "fast". The NetBurst architecture was based around this long pipeline and Intel's engineers had some reason or another for doing it (I remember reading a detailed article on NetBurst years ago that described why it was a long pipeline...the theory was good, the execution was bad).

The Core architecture basically goes the way of AMD. Shorter pipeline, more work per cycle, less cycles. This allows a 2.4ghz C2D to pretty much destroy say a 3.4ghz P4. There are obviously deeper differences than that, but that's the basics.

 

[Donnie27]

The Core architecture basically goes the way of AMD. Shorter pipeline, more work per cycle, less cycles. This allows a 2.4ghz C2D to pretty much destroy say a 3.4ghz P4. There are obviously deeper differences than that, but that's the basics.

This invites a Flame Fest! Core Architecture went back to the Pentium Pro roots long before there was any kind of Athlon. It was born from the Pentium-M

 

[Calefaction]

This invites a Flame Fest! Core Architecture went back to the Pentium Pro roots long before there was any kind of Athlon. It was born from the Pentium-M

Don't even start dude. We don't need any !!!!!! crap in here.

 

[cleric_retribution]

http://en.wikipedia.org/wiki/Intel
http://en.wikipedia.org/wiki/Core_2_Duo
http://en.wikipedia.org/wiki/Pentium
http://en.wikipedia.org/wiki/Pentium_4#Successor

I am suprised to read on wikipedia that the "Core" architecture is historically linked to the Pentium Pro line of processors. I always liked my pentium pro's lol.

 

[Donnie27]

Don't even start dude. We don't need any !!!!!! crap in here.

Exactly, what I was warning you about. Core has NOTHING to do with anything from AMD, not an opinion but a fact.

 

[Calefaction]

Exactly, what I was warning you about. Core has NOTHING to do with anything from AMD, not an opinion but a fact.

First off, I am not a !!!!!! of either company, both make exceptional CPU's. Second off, you would be fool not to think AMD's stellar success with a short pipeline and middle of the road clock speeds didn't have some effect on Intel's design of the Core micro architecture. Yes it's loosely based on Pentium Pro, and yes it was designed by the same team in Haifa as the Pentium M, but without competitive pressure from AMD, Core wouldn't be in our laps right now.

Don't be an Intel bias tool. While Core obviously isn't a reverse engineering of any of AMD's chips, it most certainly took some competitive cues from AMD's success with a short, wide pipeline and medium clock speeds. See Prescott having a 31 step pipeline, with Conroe weighing in at only 14 steps.

 

[Donnie27]

First off, I am not a !!!!!! of either company, both make exceptional CPU's. Second off, you would be fool not to think AMD's stellar success with a short pipeline and middle of the road clock speeds didn't have some effect on Intel's design of the Core micro architecture. Yes it's loosely based on Pentium Pro, and yes it was designed by the same team in Haifa as the Pentium M, but without competitive pressure from AMD, Core wouldn't be in our laps right now.

Don't be an Intel bias tool. While Core obviously isn't a reverse engineering of any of AMD's chips, it most certainly took some competitive cues from AMD's success with a short, wide pipeline and medium clock speeds. See Prescott having a 31 step pipeline, with Conroe weighing in at only 14 steps.

I didn't call you anything, I simply warned you that what you're saying is inviting a Flame Fest. Intel went back to ITS ROOTS, WTF do you not understand about that LOL! Only a damned FOOL would try to say Conroe had anything to do with AMD=P Conroe is NOT inspired by AMD, neither was Dothan or Banias as they progressed.

The Core architecture basically goes the way of AMD.

That line is absolute Bullshit! P3 has two fewer pipes than Athlon or do you even know that?

I didn't say it wasn't brought on by competition from AMD, hell I just got through with arguing that Point with Scali2. You should practice not being an AMD tool and reading the damned AMD zone too much, hehehe! AMD's success is a figment of your AMD Green tinted imagination Bucko! How many Quarters have AMD been in the black now after 20+ Quarters of being in the red, yes, I thought so.

Intel Drew on Pentium Pro, then to Pentium3 and many, including myself, thought it sucked folks in Intel didn't support the first Pentium-M with a real desktop board. So please take your calls of Tools and Fans some frackin where else, please? Oh and that old dead assed line about who built it, oh brother. Intel is in many Countries and it doesn't matter if some African or Eskimo built it, their name tag or ID Badge say Intel, get it? It is an Intel processor, NOT an AMD one LOL! You're funny Are you anti Semitic?

 

[Donnie27]

Can anyone direct me to an article that explains the technology difference between P4 and C2D? Back in the day it was all about Ghz... now there seem to be many different factors that affect processor preformance.

I suppose the thing I am finding most confusing is exactly how FSB works, and how a processor running at a lower frequency can preform so much better than my old p4.

Thanks in advance

Read though some of the Core 2 Duo Launch material? The main differences are folks at Intel wanted a better Mobile Processor so they looked at the best laptop Processor they ever built, their P3 and then Pentium M. Other processors of that time had to use more power to run Fast or save power and run slower. The first Banias was a rebuild of this processor to absolutely save power, run faster and have a large Cache to keep it from going to main memory.

Here's a good one that will not put you to sleep.

http://techreport.com/reviews/2006q3/core2/index.x?pg=2

Core's genesis was a project known internally at Intel as Merom, whose mission was to build a replacement for the Pentium M and Core Duo mobile processors. The Israel-based design team responsible for Intel's mobile CPUs followed a distinctive design philosophy focused intently on energy efficiency, which helped make the Pentium M a resounding success as part of the Centrino platform. When power and heat became problems for Netburst-based desktop and server processors, Intel turned to Merom as the source of a new, common microarchitecture for its mobile, desktop, and server CPUs.

 

[Met-AL]

Come on Donnie, he didn't say what you are claiming he said.

During the Netburst era, Intel figured the mass population would buy PC's based entirely on clock speed, such as the OP here, and AMD went the other way. Well, the "other" way, the way AMD headed down, turned out to be the better way, and Intel reversed it's decision and went back that way as well.

He didn't say Intel copied AMD, or C2D are based on the AthlonXP/64. He just said that Intel is going down the same road with the C2D that AMD did.

Stop being so sensitive, this is why you seem like such a !!!!!!, even though you aren't

 

[Donnie27]

Come on Donnie, he didn't say what you are claiming he said.

During the Netburst era, Intel figured the mass population would buy PC's based entirely on clock speed, such as the OP here, and AMD went the other way. Well, the "other" way, the way AMD headed down, turned out to be the better way, and Intel reversed it's decision and went back that way as well.

He didn't say Intel copied AMD, or C2D are based on the AthlonXP/64. He just said that Intel is going down the same road with the C2D that AMD did.

Stop being so sensitive, this is why you seem like such a !!!!!!, even though you aren't

AMD had no choice LOL! If they could have sold something to get them out of the red, hell they would have.

I warned him, look at his stupid assed comments in replies to me about cut the !!!! crap LOL!

I didn't say he said Intel Copied AMD, I'm saying that's Bullshit that Intel went the way of AMD, meaning fewer pipes since P3 already had 2 fewer pipes than AMD already. Folks saw what Tualatin did at 1.4GHz and Pentium M did with Games. Intel leaned on the Pentium-M because Netburst Failed to reach 5GHz and faster. Intel didn't go the way of AMD LORFL! That's the Flame invitation I was warning about. I even linked to make that point. If Netburst is at about 7GHz and not burning holes through the side of Case, we don't see a Core 2 Duo.

 

[Martyr]

AMD had no choice LOL! If they could have sold something to get them out of the red, hell they would have.

I warned him, look at his stupid assed comments in replies to me about cut the !!!! crap LOL!

I didn't say he said Intel Copied AMD, I'm saying that's Bullshit that Intel went the way of AMD, meaning fewer pipes since P3 already had 2 fewer pipes than AMD already. Folks saw what Tualatin did at 1.4GHz and Pentium M did with Games. Intel leaned on the Pentium-M because Netburst Failed to reach 5GHz and faster. Intel didn't go the way of AMD LORFL! That's the Flame invitation I was warning about. I even linked to make that point. If Netburst is at about 7GHz and not burning holes through the side of Case, we don't see a Core 2 Duo.

this is what makes me think intel still has a quick answer to anything amd may put out. im sure they did alot of r&d on how to clock processors very fast before they gave up with netburst, and they could implement it on the core to ramp up speeds very quickly.

 

[Slozz]

Don't even start dude. We don't need any !!!!!! crap in here.

too late.

 

[Donnie27]

this is what makes me think intel still has a quick answer to anything amd may put out. im sure they did alot of r&d on how to clock processors very fast before they gave up with netburst, and they could implement it on the core to ramp up speeds very quickly.

I'm sicken by some of the BS posted unchallenged by refugees from the AMD sub form. Saying Intel followed AMD's lead of High IPC, low Power Draw and short fat pipes is utter BS. Or using Code words for such a thing. This, when Intel had P3, Pentium-M already having shorter pipes and running cooler than Anything AMD had that time and probably ever, sheesh! It's like some folks live in an Alternate reality or something.

Common Sense question? What would you do if you launched Centrino, then have it out sale everything in the mobile market, even when the whole market is in a slump? Millions of folks and mosts OEM begged Intel to bring some form of Pentium M to the Desktop. But we're supposed to believe Intel followed or took cues from AMD? Right.

 

[Met-AL]

I'm sicken by some of the BS posted unchallenged by refugees from the AMD sub form. Saying Intel followed AMD's lead of High IPC, low Power Draw and short fat pipes is utter BS. Or using Code words for such a thing. This, when Intel had P3, Pentium-M already having shorter pipes and running cooler than Anything AMD had that time and probably ever, sheesh! It's like some folks live in an Alternate reality or something.

Common Sense question? What would you do if you launched Centrino, then have it out sale everything in the mobile market, even when the whole market is in a slump? Millions of folks and mosts OEM begged Intel to bring some form of Pentium M to the Desktop. But we're supposed to believe Intel followed or took cues from AMD? Right.

Well if you want to be retarted about it, then I will say, Intel abandoned the short pipes after the P3 when Netburst came into being, hoping to crank up the clock speed and decieve consumers by getting them to think that a 2.2Ghz P4 was faster then a 1.8Ghz Athlon. Only when it was obvious to Intel that Netburst was NOT going to hit 10Ghz, did they change their tune and say that clock speed wasn't everything, while all along the rest of the industry knew that.

AMD didn't do that. AMD continued on with the K7 and only then did Intel decide to come back to the short pipes with the Pentium-M. But AMD never left the "right" road, so that makes Intel the dumbass that chose the wrong road, but wisened up when they got schooled and decided to copy AMD. So, how is that for a fan boy way to put it. It fits right in your type of posting.

 

[NulloModo]

Well if you want to be retarted about it, then I will say, Intel abandoned the short pipes after the P3 when Netburst came into being, hoping to crank up the clock speed and decieve consumers by getting them to think that a 2.2Ghz P4 was faster then a 1.8Ghz Athlon. Only when it was obvious to Intel that Netburst was NOT going to hit 10Ghz, did they change their tune and say that clock speed wasn't everything, while all along the rest of the industry knew that.

AMD didn't do that. AMD continued on with the K7 and only then did Intel decide to come back to the short pipes with the Pentium-M. But AMD never left the "right" road, so that makes Intel the dumbass that chose the wrong road, but wisened up when they got schooled and decided to copy AMD. So, how is that for a fan boy way to put it. It fits right in your type of posting.

Intel didn't say speed was king all along, they did abandom the P4-M chips in favor of lower-clocked Pentium Ms in their notebooks fairly early on, but that had to do with heat and power usage as much as anything else.

I wouldn't look at it as a right vs. wrong road either, Intel fully believed that the Netburst architecture could hit 10ghz+ by more than a year ago, and if it had cranked up in clockspeed as fast as their were hoping, it would have wiped the floor with everything else. However, since no one was building CPUs 2ghz, much less 3ghz+ at the time when Netburst was being designed, the engineers had no way of knowing that crosstalk and thermal limitations would hit them like they did. Once Intel did realize they couldn't do it, they took the cue from the success of the Pentium M and had the Israeli team start devising a new desktop chip based on it, but designing and proofing a new CPU architecture takes years, and since most people in the marketplace still believed clockspeed was everything, there was no choice to continue to ramp up Netburst as best they could.

Hindsight being 20/20 it wasn't the best technological decision, but marketing wise, it was correct, as Intel still sold far far more P4s than AMD ever sold A64s, and while AMD did catch on and even overtake for a bit with the performance crowd as shown in their retail processor sales, the overwhelming bias of OEM P4s and Xeons makes the tiny retail box market inconsequential.

Now Intel has the OEM sales and the retail box/performance crowd back since they have released the C2D, and AMD is doing what Intel had to do - stick with what you've got and ramp it up, as shown by the 4x4 platform and the FX70 series while they try to get this K8L chip out, and once that hits market, maybe we will see things shift again for a bit, but who knows until products are actually on the shelves.

 

[EQTakeOffense]

Exactly, what I was warning you about. Core has NOTHING to do with anything from AMD, not an opinion but a fact.

But, AMD has everything to do with Core.

AMD was a producer of Intel chips in the early 80's since Intel at that time didn't have the facilities to keep up with demand. Long story short, AMD pretty much reversed engineered Intel's processors in the late 80's and early 90's to create the Athlon XP. These reversed engineered CPU's were based on the Pentium Pro at the time, and eventually lead to the current Athlon64/X2 architecture.

There are some very intresting legal cases between AMD and Intel, least among them AMD's suit against Intel for monoploistic practices. Intel actually sued AMD for reverse-egineering their CPU's.

 

[Met-AL]

Hindsight being 20/20 it wasn't the best technological decision, but marketing wise, it was correct,

Yep, it was marketing genious. RIght down to calling it Netburst so common folk would associate the P4 with speeding up the internet experience back then. Those commercials were so full of bull-puckey.

If anyone remembers, the K7 and the P3 were neck and neck in performance per clock speed and AMD was getting pretty competitive, especially when they hit 1Ghz first. So, if you can't flat out win that way, come out with a CPU that will totaly own the competition in regards of it's clock speed. Most "common" people have no clue other then this Compaq with a 1.33Ghz AMD Athlon T-bird sure sounds a lot slower then this HP with a 1.5Ghz P4, when in fact that T-Bird was by far the better performer. It didn't matter, though, higher number means better right?

I guess it left me with a feeling that Intel was trying to scam. Some kind of indirect false advertising. It made me resist buying a P4 or letting anyone that asked me for advice buy a P4.

But hey, the Marketing department is not in charge over at Intel anymore. Intel has vindicated themselves, the Core 2 is awesome and awesome in a genuine way, at an awesome price. AMD milked us a little too much with crap like $390 X2 4200's and now they are gonna pay for it.

 

[Glyphic]

Not all about Speed? I remember a while back Intel kept promising us 4-5GHZ by xxxx date.

 

[Donnie27]

Well if you want to be retarted about it, then I will say, Intel abandoned the short pipes after the P3 when Netburst came into being, hoping to crank up the clock speed and decieve consumers by getting them to think that a 2.2Ghz P4 was faster then a 1.8Ghz Athlon. Only when it was obvious to Intel that Netburst was NOT going to hit 10Ghz, did they change their tune and say that clock speed wasn't everything, while all along the rest of the industry knew that.

AMD didn't do that. AMD continued on with the K7 and only then did Intel decide to come back to the short pipes with the Pentium-M. But AMD never left the "right" road, so that makes Intel the dumbass that chose the wrong road, but wisened up when they got schooled and decided to copy AMD. So, how is that for a fan boy way to put it. It fits right in your type of posting.

I'm "retarted" 2.2GHz P4 kicked the shit out of the 1.8GHz Athlon with streaming Apps, hell so did 2GHz LOL! I know because I had a 2400+ and a 2GHz Northwood. AMD didn't see a clear victory until X2 launched, until then, each had strong and weak points. That's not an opinion but FACT!

If AMD could have taken market by selling processors with more pipes and MHz, I'm sure they most certainly would have. They used what they had. That's moot anyphucking way. Intel went to Conroe because P4 Failed to reach heights they planned for it=P

Yet, here we have folks acting like Intel didn't have Pentium-M, saw the benefits of it and finally listened to what the hell its customers asked for, hehehe! Now I'm retarded because I believe that, oh brother I mean damn wasn't Apple Going Intel a clue?

 

[Donnie27]

Yep, it was marketing genious. RIght down to calling it Netburst so common folk would associate the P4 with speeding up the internet experience back then. Those commercials were so full of bull-puckey.

I agreed 100% and often laughed or folks repeating the TV spots. It was very Sleazy on Intel's Part.

If anyone remembers, the K7 and the P3 were neck and neck in performance per clock speed and AMD was getting pretty competitive, especially when they hit 1Ghz first. So, if you can't flat out win that way, come out with a CPU that will totaly own the competition in regards of it's clock speed. Most "common" people have no clue other then this Compaq with a 1.33Ghz AMD Athlon T-bird sure sounds a lot slower then this HP with a 1.5Ghz P4, when in fact that T-Bird was by far the better performer. It didn't matter, though, higher number means better right?

Even if P4 1.5 was slower but not in all apps, it had very stable motherboards. But didn't AMD add pipes first? Yet even the Willy didn't get as bad as Conroe spanked X2 or even bad as X2 spanked P4D and etc...

I guess it left me with a feeling that Intel was trying to scam. Some kind of indirect false advertising. It made me resist buying a P4 or letting anyone that asked me for advice buy a P4.

But hey, the Marketing department is not in charge over at Intel anymore. Intel has vindicated themselves, the Core 2 is awesome and awesome in a genuine way, at an awesome price. AMD milked us a little too much with crap like $390 X2 4200's and now they are gonna pay for it.

Yes it was a scam and very similar to what AMD is doing now LOL!

 

[Donnie27]

Not all about Speed? I remember a while back Intel kept promising us 4-5GHZ by xxxx date.

There were always Internal fights at Intel where it as called the Brainiac vs. the Speed Demon, the Brainiac side won! According an Intel Employee I've known for about 10 years, that fight started when P4 first tapped out.

 

[Met-AL]

Even if P4 1.5 was slower but not in all apps, it had very stable motherboards. But didn't AMD add pipes first?

Nope, the Athlon stayed the same as far as that is concerned all through the different versions. The big change was from the T-Bird to the Paly cores when AMD increased the performance of the branch predictor which is the official reasononing behind the XP rating which was a XP chip at a certain rating matched the performance of a T-Bird at that speed..for example a XP 1600+ (1.4Ghz) = a T-Bird 1.6Ghz.

Oh and yeah, there were some motherboards that were real turds. I never had one as I did some research before I did my purchases, but yeah, there was some real junk out there.

Yes it was a scam and very similar to what AMD is doing now LOL!

Scam? Can you explain what you are refering too? I explained my opinion in my scam statement.

 

[Donnie27]

Nope, the Athlon stayed the same as far as that is concerned all through the different versions. The big change was from the T-Bird to the Paly cores when AMD increased the performance of the branch predictor which is the official reasononing behind the XP rating which was a XP chip at a certain rating matched the performance of a T-Bird at that speed..for example a XP 1600+ (1.4Ghz) = a T-Bird 1.6Ghz.

Oh and yeah, there were some motherboards that were real turds. I never had one as I did some research before I did my purchases, but yeah, there was some real junk out there.



Scam? Can you explain what you are refering too? I explained my opinion in my scam statement.

All the BS about "Native Quad Core"? Or how about the all New Quadfather that's nothing more than a lightly modded Server Board. Why would someone be stupid enough to wait on K8L and two Quad core when Cloverton will be out almost a half a year sooner?

No, the AthlonXP didn't gain anymore pipes but the Athlon64 sure did.

Chip Architect

The point was to squeeze more MHz, the stages had to be increased. I much rather have seen Intel put more into Northwood C which didn't do poorly against AMD at all. Prescott was BS and that's why I never bothered buying one. The last models are better but they came too close to Conroe.

I'm not saying Competition didn't light a fire under Intel's ass and made them go to work or simply wake up. I'm saying the inspiration was already there and that inspiration wasn't AMD, sheesh!