Anandtech Previews 65nm

[Jonsey]

Ya know, the most underrated Intel products right now are their BTX stuff. Seriously, I put together a couple of systems, the latest with Evercase's mBTX offering...the final result was just sublime...it was so quiet, there was SOOO much functionality in such a little chassis, and damn if for once I was able to get the right riser cards to add a full size graphics solution. Oh and it kept everything pretty cool....

I have one of their latest BTX dell workstations at work. It is so quiet, even under load, I can't hear it on, even with my ear next to it.

 

[Donnie27]

Donnie27, you mention performance outside of gaming as if Intel is wining in that area...in my experience they are not...and I've experienced the latest P4's, PD's, A64's, and A64 X2's...

A64X2 did, not A64 until the FX55 launched and damned sure not AthlonXP. For almost 8 months, the only A64 were high priced to the point that same Sabrewolf or Tachy as he's known other forums couldn't even afford one. It takes more than a speedy processor to make a good computer and platforms have NOT been an AMD strong point as they depend on others to do it for them.

The article was about going from 90nm to 65nm. This same little 15 year old expert who Constantly flames anything that's negative about his beloved AMD or proves the slightest positive word having anything to do with the P4, is a joke at best. It's almost like its a sickness or a reflex action. Uh-Oh, there's something halfassed good about the P4, here's tachy to say otherwise. Again, AMD can only wish they sucked like that.

I didn't say the P4 was suprior performance wise. I also would compare it as say that it bests a P4 like a Farrari bests a Ford either, that's a laugh.

I'm calling you retard, there I put it in laymans terms. Btw, I have 3 P-M systems. It's just netburst that blows. So why are you posting? if you buy because intel is the larger company and has better advertising, but not better performance you shouldn't be on this forum

See, more of the same. That has absolutely NOTHING to do with Presler or Cedar Mill 65nm or anything. Then look at his lame assed post to me? So, is this cool now?

Donnie27

 

[Donnie27]

the "c" series was the only good p4 in netburst's history, it destroyed it's competition (barton)

Uh-Oh, the world is about to End! Something positive from the biggest Netburst basher on 6 forums, WOW!

Donnie27

 

[1c3d0g]

I have one of their latest BTX dell workstations at work. It is so quiet, even under load, I can't hear it on, even with my ear next to it.

Damn... looks like they improved their case designs then. Too bad they couldn't do that way earlier.

B.t.w., how's the room temperature?

 

[sabrewolf732]

A64X2 did, not A64 until the FX55 launched and damned sure not AthlonXP. For almost 8 months, the only A64 were high priced to the point that same Sabrewolf or Tachy as he's known other forums couldn't even afford one. It takes more than a speedy processor to make a good computer and platforms have NOT been an AMD strong point as they depend on others to do it for them.

?? how do you know how much money I have? You're making ignorant generalizations with no proof whatsoever. And what is the point of that? AMD has been the performance leader for quite a while. that is undebatable atm.

The article was about going from 90nm to 65nm. This same little 15 year old expert who Constantly flames anything that's negative about his beloved AMD or proves the slightest positive word having anything to do with the P4, is a joke at best. It's almost like its a sickness or a reflex action. Uh-Oh, there's something halfassed good about the P4, here's tachy to say otherwise. Again, AMD can only wish they sucked like that.

Again, you're making ignorant generalizations with no proof. Because I said prescott blows and a64 is the performance leader now makes me a huge amd fan boy huh? You're a moron.

I didn't say the P4 was suprior performance wise. I also would compare it as say that it bests a P4 like a Farrari bests a Ford either, that's a laugh.

When a 1.8GHz 3000+ ties a 3.6GHz pentium 4 in half life 2, I would say that is a HUGE performance deficit.

See, more of the same. That has absolutely NOTHING to do with Presler or Cedar Mill 65nm or anything. Then look at his lame assed post to me? So, is this cool now?

you're being a hypocrite. You told me I had no reason to post on this forum, what does that have to do with presler or cedar?

 

[sabrewolf732]

Uh-Oh, the world is about to End! Something positive from the biggest Netburst basher on 6 forums, WOW!

Donnie27

Again, you're just proving your idiocy. I say something positive about netburst and you reply sarcastically. I don't understand how saying the truth (amd 64 > netburst) makes me a !!!!!!. WTF? Ill say this right now, P-M > A64, does that make me a intel !!!!!! now? Or am I still a amd !!!!!!? You need to lay off the drugs donnie.

 

[Chris_Morley]

A64X2 did, not A64 until the FX55 launched and damned sure not AthlonXP. For almost 8 months, the only A64 were high priced to the point that same Sabrewolf or Tachy as he's known other forums couldn't even afford one. It takes more than a speedy processor to make a good computer and platforms have NOT been an AMD strong point as they depend on others to do it for them.

The article was about going from 90nm to 65nm. This same little 15 year old expert who Constantly flames anything that's negative about his beloved AMD or proves the slightest positive word having anything to do with the P4, is a joke at best. It's almost like its a sickness or a reflex action. Uh-Oh, there's something halfassed good about the P4, here's tachy to say otherwise. Again, AMD can only wish they sucked like that.

I didn't say the P4 was suprior performance wise. I also would compare it as say that it bests a P4 like a Farrari bests a Ford either, that's a laugh.



See, more of the same. That has absolutely NOTHING to do with Presler or Cedar Mill 65nm or anything. Then look at his lame assed post to me? So, is this cool now?

Donnie27

You make a lot of good points, Donnie, and you do so in very thorough manner. I agree that the flaming is uncalled for, unfortunately I am not a Moderator in this forum and we have a rule against being a 'mini-mod' in forums that we have no authority. I recommend you report offensive posts and allow the forum mods or an admin to take care of this.

 

[Jonsey]

Damn... looks like they improved their case designs then. Too bad they couldn't do that way earlier.

B.t.w., how's the room temperature?

Our A/C is always on the fritz, so temp varies from high 60's to low 80's. Under all circumstances, even heavy load, it is dead silent. It is only a 3GHz model though.

 

[Lethal]

Oh yeah, I can post wherever I damn well please ...

Not when you're banned, you can't. I'm giving you 3 days off for flaming in this thread after not heeding a previous warning.

 

[Donnie27]

You make a lot of good points, Donnie, and you do so in very thorough manner. I agree that the flaming is uncalled for, unfortunately I am not a Moderator in this forum and we have a rule against being a 'mini-mod' in forums that we have no authority. I recommend you report offensive posts and allow the forum mods or an admin to take care of this.

No biggie!

I had 9 AMD systems (over the years), 7 Intels as I added an old Pentium system for my little girl. Next system will make number 10. No one has to bash either system to like or choose the other. But that's too simple sometimes. AMD has a great product yet, there are still reasons folks buy Intel even if they don't make the fastest processor. I'm getting AMD next but I'm NOT about to call folks buying P4s ding bats or worse.

I'll reserve judgement and just stick to talking about temps on 65nm until the reviews pop up instead of the previews and teasers like this one. Intel has always taken something out, added something else or tweaked other parts. They've almost never just made any of the Netburst based Processors smaller only.

Donnie27

 

[Andross182]

No biggie!

I had 9 AMD systems (over the years), 7 Intels as I added an old Pentium system for my little girl. Next system will make number 10. No one has to bash either system to like or choose the other. But that's too simple sometimes. AMD has a great product yet, there are still reasons folks buy Intel even if they don't make the fastest processor. I'm getting AMD next but I'm NOT about to call folks buying P4s ding bats or worse.

I'll reserve judgement and just stick to talking about temps on 65nm until the reviews pop up instead of the previews and teasers like this one. Intel has always taken something out, added something else or tweaked other parts. They've almost never just made any of the Netburst based Processors smaller only.

Donnie27

alot of it has to do with marketing schemes....hell most people that come up to my tech counter say they own an intel computer just because intel pays so much to have their name out there

 

[The Donut]

From a NON-GAMING stand point i'm an Intel fan. Earlier this year, I switched from my P4 2.6C to an Athlon64 3500+, SLI, the works.. and in gaming - I loved it. But, soon as I started doing my work, (multiple spreadsheets, photoshop, illustrator etc open) the machine felt incredibly sluggish. To the point that I sold it 2 weeks later and returned to my P4 and am currently waiting for a _good_ reason to upgrade.

I'm not bashing either side, but many of the peoples arguments (first page, especially) are that the AMD crushes Intel in gaming - those people have to realize that gaming, while a growing audience, certainly isn't Intels target audience.

 

[Donnie27]

alot of it has to do with marketing schemes....hell most people that come up to my tech counter say they own an intel computer just because intel pays so much to have their name out there

Intel's Marketing is way over rated when you stop and think about it. Some Quarters AMD has sold up to 55% of the processors sold in the U.S., right where most of Intel's ads show. Now that's with almost NOTHING coming from AMD. So when folks say AMD is stupid for not marketing, I tend to disagree and think that money saved is good, hehehe! I think Intel wastes too much on TV ads.

90nm worked, Prescott was just bad=P Pentium-M proved 90nm was good for Intel. If Pentium N (Netburst) gained 600MHz without increased temps, that's a step in the right direction. If this same thing happens with Pentium-M then lookout! With 65nm 3.4 giving off as much as the 90nm 2.8 that's a start. As the old old photo showed, this thing started out at 2GHz. This puts pressure on AMD as the follow on Ceadar Mill will be dirt cheap and sold like Celerons. The biggest problems for Intel and AMD have been their own low cost models, NOT each other. People in less developed Countries can't afford bare X2 and P4EE processors that costs more money than they might make in one year. Yet, that's where the most Growth is. Many of the Mid ranged products aren't affordable to these folks either.

If I were some of the Hardcore AMD and Intel Followers, I'd be pissed that they'd call me smart on the hand and treat me like I'm dumbass on the other by overpricing X2, higher end Athlon64's, P4EE and the biggest joke of them all, the 840D-EE. It's almost like a bad JOKE or something.

Donnie27

 

[Donnie27]

From a NON-GAMING stand point i'm an Intel fan. Earlier this year, I switched from my P4 2.6C to an Athlon64 3500+, SLI, the works.. and in gaming - I loved it. But, soon as I started doing my work, (multiple spreadsheets, photoshop, illustrator etc open) the machine felt incredibly sluggish. To the point that I sold it 2 weeks later and returned to my P4 and am currently waiting for a _good_ reason to upgrade.

I'm not bashing either side, but many of the peoples arguments (first page, especially) are that the AMD crushes Intel in gaming - those people have to realize that gaming, while a growing audience, certainly isn't Intels target audience.

I still use a 2.6C right now. Eventhough I did see the steam warning, I played the Forgotten Coast Demo last night. I had it set to med and 1024 X 768 with only a 9700 Pro. It's something I'll get for sure. I want to do a system just like what you did and run both computers togethter with the Stardock app called Multiplicity. Next year, I'll upgraded that AMD system to an X2 when they move to the next socket, stay right off the upgrade curve.

The computer will look like;
GeForce 7800GT
DFI either the Infinity or the lowest end SLI Model
3500+ might be slightly overclocked
250GB 16MB cache WD HDD Getting it this week.

I already have the
Case, a very nice sized Black Mid-Tower SunFlower 461
Antec 550W TP 2.0 SLI PSU
LiteOn CD-R/DVD-Rom Combo
NEC 3540B 16X DVD-+Writer
Floppy
1GB of PQI PC-4000
X-Fi Xtreme Music, may upgrade to a Platinum
I run Dual CRTs, a 21" and one 19"
KB, Mouse and yada yada.

5.1 receiver + Home Speakers are used instead of Computer speakers. I'm using the X-Fi and RAM in my Current computer but the rest is just setting there waiting. I have 512MB of PC-3500 and an Audigy 2 ZS that will replace or fill once this stuff is moved.

Both of my computers will be connected to the Receiver. The other will be simply for Stereo Music and other stuff Intel is good at.

Donnie27

 

[1c3d0g]

Our A/C is always on the fritz, so temp varies from high 60's to low 80's. Under all circumstances, even heavy load, it is dead silent. It is only a 3GHz model though.

Well...considering that it's a Prescott, and one clocked @ 3GHz at that, I'm dissapointed they couldn't do the same with a Northwood 2.6GHz, 'cause mine is still screaming like a little girl (even after changing the main fan to a Panaflo)...

 

[sabrewolf732]

woot im unbanned. Now that I'm unbanned I would like to say sorry for the flaming, but donnie, you're still wrong We can agree to disagree. Thanks for unbanning lethal

 

[(cf)Eclipse]

Intel has always taken something out, added something else or tweaked other parts. They've almost never just made any of the Netburst based Processors smaller only.

williamette -> northwood was simply a shrink + cache, no?
that was definitly intel's prime imo. northwood was the shit for quite a while


edited cause i had a brain fart

 

[Donnie27]

williamette -> netburst was simply a shrink + cache, no?

No it wasn't. Several changes were made. Even with both at 2GHz/400MHz FSB, an apples to apples reveiw showed Northwood 4 to 8% faster even at the same speed and FSB. Remember, Intel doubled the sized of the Willy's 256K L2 on a 18nm processor to 13nm Northwood with 512K. Parts of the Layout also moved or was changed.

It can be said that AMD was the one who Changed Names and Charged for every improvement they made. Sorry that's sleaze at its best. As the Northwood grew. Intel Kept the prices very close, the same or even cheaper. Northwood performance increased with its 533 and 800MHz FSB. Intel still held the prices down and still called them Northwoods until Prescott launched.

Donnie27

 

[robberbaron]

williamette -> netburst was simply a shrink + cache, no?
that was definitly intel's prime imo. netburst was the shit for quite a while

It's Northwood isn't it?

And it allowed for a few tweaks like Hyperthreading and I'd assume better SSE and FPU.

"It can said that AMD was the one who Changed Names and Charged for every improvement they made. Sorry that's sleaze at its best. As the Northwood grew. Intel Kept the prices very close, the same or even cheaper. Northwood performance increased with its 533 and 800MHz FSB. Intel still held the prices down and still called them Northwoods until Prescott launched.

Except that you needed a brand new chipset for each iteration of the Pentium 4, whereas the AMD chipsets tend to last years.

Also, Northwood was the same basic silicon with each FSB increase. And I'm trying to understand how giving a NEW core a NEW core name is a bad thing? It allows for a more accurate description of the CPU. I really wish I could follow your logic as you grasp for straws in insulting AMD.

 

[(cf)Eclipse]

And I'm trying to understand how giving a NEW core a NEW core name is a bad thing?

agreed. saying venice vs san diego is a good way of differentiating between a 512kb or 1mb cache revE cpu
however, on the other side of the coin, intel doesn't mess with cache size and such nearly as often

 

[Donnie27]

It's Northwood isn't it?

And it allowed for a few tweaks like Hyperthreading and I'd assume better SSE and FPU.

Except that you needed a brand new chipset for each iteration of the Pentium 4, whereas the AMD chipsets tend to last years.

Also, Northwood was the same basic silicon with each FSB increase. And I'm trying to understand how giving a NEW core a NEW core name is a bad thing? It allows for a more accurate description of the CPU. I really wish I could follow your logic as you grasp for straws in insulting AMD.

No they didn't. My Asus P4T-E Sc-478 was compatible with both my first P4 1.7 Willy Sc-478 and My first Northwood 2GHz Sc-478. Sc-478 wasn't a secret and it was widely known it would replace Sc-423

Pre-Hammer AMD also went from 100, 133, 166 and 200 x 2 or 200, 266, 333 and 400MHz FSB repectively. Yes we're on nForce 4 now. Now go back and run a Barton on a nForce 1? Yes, and who in the hell wants to hold on to those old limted edition first Sc-939 boards? Ohhhhh, use Ad-in SATA-controller cards, stick with old AGP, sure, right? Most DIYers upgrade boards anyway, AMD or Intel users.

There is no phucking grasping for straws and if you're buying AMD Processors, you should feel insulted by AMD. First of all, it's NOT AMD chipsets since they rely on and are at the mercy of 3rd parties. In 2.5 years, we've seen the long forgotten nForce 3 150, 250, 250GB. Then there are at least 4 models and at least 3 revision of the nForce4 and that's just 939. The last model being SLI with two X16 PCI lanes instead of 16X that drops to 8 when in SLI Mode.

Again, AMD charged for each new stepping of the same Processor/s. Even when these new steppings were FIXES. Rev-e is a prime example. The only real difference between it and the privisious 90nm model is AMD slowed down the Memory controller to make it more compatible when all four memory slots are used. The Controller now didn't kick into a type of *Super Stability Mode and drop the RAM's speed to 333MHz instead of 400MHz. There's not a damned thing added to the Processor Core, package or etc., to justify AMD charging extra=P It's why I waited on Price cuts before buying.

For the record, Intel SUCKS IMHO for all the Newer chipsets after i875. Before that though, I'd disagree.

Donnie27

 

[Asian Dub Foundation]

Merom is what I want to see.

yep, i have hopes for the conroe, just cuz i want to see a good fight between intel and AMD

 

[Donnie27]

agreed. saying venice vs san diego is a good way of differentiating between a 512kb or 1mb cache revE cpu
however, on the other side of the coin, intel doesn't mess with cache size and such nearly as often

Originally Posted by robberbaron
And I'm trying to understand how giving a NEW core a NEW core name is a bad thing?

To you both!

You ask how?
A. When it's NOT a new Core, sheesh! Venice and San Diego are the same stepping with different amounts of Cache. So give me some differences between The Winchester D Core Stepping and Venice? You can't because there AREN'T any. The Name aren't what's sleazy about it, it's their price increase LOL!

Donnie27

 

[sabrewolf732]

No they didn't. My Asus P4T-E Sc-478 was compatible with both my first P4 1.7 Willy Sc-478 and My first Northwood 2GHz Sc-478. Sc-478 wasn't a secret and it was widely known it would replace Sc-423

Pre-Hammer AMD also went from 100, 133, 166 and 200 x 2 or 200, 266, 333 and 400MHz FSB repectively. Yes we're on nForce 4 now. Now go back and run a Barton on a nForce 1? Yes, and who in the hell wants to hold on to those old limted edition first Sc-939 boards? Ohhhhh, use Ad-in SATA-controller cards, stick with old AGP, sure, right? Most DIYers upgrade boards anyway, AMD or Intel users.

There is no phucking grasping for straws and if you're buying AMD Processors, you should feel insulted by AMD. First of all, it's NOT AMD chipsets since they rely on and are at the mercy of 3rd parties. In 2.5 years, we've seen the long forgotten nForce 3 150, 250, 250GB. Then there are at least 4 models and at least 3 revision of the nForce4 and that's just 939. The last model being SLI with two X16 PCI lanes instead of 16X that drops to 8 when in SLI Mode.

Again, AMD charged for each new stepping of the same Processor/s. Even when these new steppings were FIXES. Rev-e is a prime example. The only real difference between it and the privisious 90nm model is AMD slowed down the Memory controller to make it more compatible when all four memory slots are used. The Controller now didn't kick into a type of *Super Stability Mode and drop the RAM's speed to 333MHz instead of 400MHz. There's not a damned thing added to the Processor Core, package or etc., to justify AMD charging extra=P It's why I waited on Price cuts before buying.

For the record, Intel SUCKS IMHO for all the Newer chipsets after i875. Before that though, I'd disagree.

Donnie27

i875 was good. However if you notice venice's are cheaper than winnies. so idk what you're talking about. What premium?

 

[robberbaron]

To you both!

You ask how?
A. When it's NOT a new Core, sheesh! Venice and San Diego are the same stepping with different amounts of Cache. So give me some differences between The Winchester D Core Stepping and Venice? You can't because there AREN'T any. The Name aren't what's sleazy about it, it's their price increase LOL!

Donnie27

D0 3200 is $190
http://www.newegg.com/Product/Product.asp?Item=N82E16819103502

E3 3200 is $152
http://www.newegg.com/Product/Product.asp?Item=N82E16819103535

That sure is a huge markup of -$38.
Also, San Diego is an E4 stepping, venice is E3 and some are E6. At least get that part right.
Venice has SSE3 (the D0 does not, not that SSE3 matters), a much better memory controller, typically overclocks higher than winchester thanks to the improved manufacturing process through DSL, all the while keeping the same thermal envelope AND the same price (if not, a lower price.)

And yes, the A64 memory controller had limitations, but in spite of those limitations, it still performed better bandwidth and latency-wise than the Intel chipset's memory controllers! So the fix to the "problem" you invented was merely icing on the cake.

Each time they added a new revision to the AMD64 lineup, they usually had a price drop shortly after all across. Each PR shares the same price, and their prices have only gone down with time. You need to at least keep track of the numbers before you spout off about them.

 

[Donnie27]

D0 3200 is $190
http://www.newegg.com/Product/Product.asp?Item=N82E16819103502

E3 3200 is $152
http://www.newegg.com/Product/Product.asp?Item=N82E16819103535

That sure is a huge markup of -$38.
Also, San Diego is an E4 stepping, venice is E3 and some are E6. At least get that part right.
Venice has SSE3 (the D0 does not, not that SSE3 matters), a much better memory controller, typically overclocks higher than winchester thanks to the improved manufacturing process through DSL, all the while keeping the same thermal envelope AND the same price (if not, a lower price.)

And yes, the A64 memory controller had limitations, but in spite of those limitations, it still performed better bandwidth and latency-wise than the Intel chipset's memory controllers! So the fix to the "problem" you invented was merely icing on the cake.

Each time they added a new revision to the AMD64 lineup, they usually had a price drop shortly after all across. Each PR shares the same price, and their prices have only gone down with time. You need to at least keep track of the numbers before you spout off about them.

First of, you're talking about after price cuts and I'm talking about as a whole. You're talking at one Web site and I'm talking all of them. I like Newegg the best but they are NOT the only place I shop for hardware.

$330 - athlon 64 4000
$334 - athlon 64 4000 90nm rev
$228 - athlon 64 3800
$248 - athlon 64 3800 512k 90nm rev e
$165 - athlon 64 3700
$234 - athlon 64 3700 90nm rev
$185 - athlon 64 3500
$201 - athlon 64 3500 939pin
$201 - athlon 64 3500 90nm 939pin
$201 - athlon 64 3500 512k 90nm rev e
$153 - athlon 64 3400
$139 - athlon 64 3200
$150 - athlon 64 3200 939pin
$150 - athlon 64 3200 90nm 939pin
$150 - athlon 64 3200 512k 90nm rev e
$117 - athlon 64 3000
$142 - athlon 64 3000 939pin
$142 - athlon 64 3000 90nm 939pin
$139 - athlon 64 3000 512k 90nm rev e

Mixed bag but after Price cuts.

Current pricewatch prices as of 8:35AM CST. This is the way AMD should have priced them right from the start, NOT after sales slowed Now that might have caused our disagreement but that's why I said I waited for the Price cuts. I did the same thing with Intel BTW. The Other Point was Intel left the Northwood name and price nearly the same as they went from A, B, and C. They didn't Charge for A0, B, C, D, MO or etc..... Prescott has almost done the same. Surely you or anyone else here aren't going to tell me Prices have always looked like they do above, right?

For the record, I didn't say anything about different names for smaller core or going from 130 to 90nm for example. More cache on the processors should cost more as well. I do complain about the same Processor and renaming and charging for a NEW STEPPING OF THE SAME. Caps to show importance, not shout.

The Rev-E overclocks better because of relaxed timings. It runs slightly slower at Stock in most apps because of these relaxed timings.

Oh, too bad 11 of the 13 instructions that make up SSE3 relate to Hyperthreading, something AMD doesn't use. So that becomes almost Gimmick at best. It is still the same Core and Rev was a fix=P

Anyway, thanks for keeping this civil.

Donnie27

 

[Donnie27]

i875 was good. However if you notice venice's are cheaper than winnies. so idk what you're talking about. What premium?

Right.

Donnie27

 

[robberbaron]

I can understand from a flat out, basic consumers standpoint, that D0 - E3 was an insignificant change to the core, but flat out basic consumers don't even know what a core is. The new name allows OC'ers to get the right chip and get the best speeds - surely you've shopped for a chipf or the sake of OCing it, yes?

Anyways, it's been a great discussion and I'm glad it has been civil, unlike other convos in this forum.

 

[Donnie27]

I can understand from a flat out, basic consumers standpoint, that D0 - E3 was an insignificant change to the core, but flat out basic consumers don't even know what a core is. The new name allows OC'ers to get the right chip and get the best speeds - surely you've shopped for a chipf or the sake of OCing it, yes?

Anyways, it's been a great discussion and I'm glad it has been civil, unlike other convos in this forum.

If it were just names, that would be fine eventhough Overclockers don't need to be treated like dummies. We already look at shipping lots and know full well what a stepping is. Again, that's one of the reasons I said many should be insulted. There are no basic consumers building their own systems. It's meant for us, NOT them. There are no DIYers too dumb to go to overclocking sights and look for best overclockers per shipping lot or Core stepping. Yet, name isn't the only thing but prices they asked for these was a problem. Thank goodness for the cuts and that's seen as a good thing IMHO.
All we needed to know was Rev or Stepping E just as we have done since the days of the CeleronA.

You might think I'm just knocking AMD. I already posted that the Guy at Intel who came up with the whole P4EE schlock, should be fired. Then the next guy (or chick) should be the one at AMD who came up with the FX. I don't like the Idea that they think folks in this sector has money to burn or something. It's like BOTH of them are minning for suckers. No matter who's side anyone is one, this should cause OUTRAGE, not folks making excuses for them. Hell, they'd have cut prices sooner. AMD says their users are smarter, then treats them like Dummies, I don't get it.

Back to the subject.

Intel moved to 65nm and I'm sure they didn't just make a smaller Prescott. Transistor counts will tell the real story on Ceadar Mill vs. a Prescott with the same config as far as L2 goes. The temp and power usage improvements is a big step in the right direction. Overclockers everywhere laughed at Anand's crack about no getting more than 3.9GHz with a P4. Dothan, Yonah and etc.. is what most of my Buds are waiting for.

Donnie27

 

[robberbaron]

Anand only got the P4 to 3.9? That's pretty below average. Many of those goofy 2.66GHz chips hit 4.0GHz, and many even get to 4.2GHz on air.

As for the core names, I hear you. But for me, personally, it's a lot easier to have different names for different chips. Like, when you say 3200+ CG, you could be either talking about the 1mb Clawhammer or 512kb cache Newcastle. So the core name does help get an idea of what chip someone I'm helping is using/planning to buy/whatever. The only downside is having to answer questions of "What is BI? What does Venice mean? How do I overclock my Jackhammer RevF 3257+?"

The overall power reduction looked to be about 25-33% from 90 to 65nm. I'd say that's a success, given how shitty netburst is for efficiency.

 

[Donnie27]

Anand only got the P4 to 3.9? That's pretty below average. Many of those goofy 2.66GHz chips hit 4.0GHz, and many even get to 4.2GHz on air.

As for the core names, I hear you. But for me, personally, it's a lot easier to have different names for different chips. Like, when you say 3200+ CG, you could be either talking about the 1mb Clawhammer or 512kb cache Newcastle. So the core name does help get an idea of what chip someone I'm helping is using/planning to buy/whatever. The only downside is having to answer questions of "What is BI? What does Venice mean? How do I overclock my Jackhammer RevF 3257+?"

The overall power reduction looked to be about 25-33% from 90 to 65nm. I'd say that's a success, given how shitty netburst is for efficiency.

My bad! He said the same thing about Pentium-D @3.9. He admits that 4.2 was more likely. So I have to clear that up right now.

I also think that eventhough I don't see myself Buying one, I think Ceadar Mill just might take up where the 2.4 and 2.6C or the left off. If anything impressed me, it was about Anand's little preview, it was this;

First up was Cedar Mill, which we had running at 3.60GHz at 1.300V by default. Our first test was to see how far we could get it without increasing the core voltage, and the best that we could do there was 4.25GHz (250MHz x 17.0).

Donnie27

 

[sabrewolf732]

Netburst will be alive and kicking in Conroe, right along with Banias. Funny that everyone (amd lovers) keep thinking that Netburst is not a great Architecture, yet it's fast enough to be faster than A64s in office apps and multimedia. X2s? I'm not buying into the Dual Core stuff, just yet and if I do, it will certainly be Presler........never amd. My current CPU is more than enough. BTW, P4s and Ds are meant to be run overclocked, heck even Intel's own MoBos allow you to overclock at least 10%, certainly a good start, perhaps one day they'll go the ASUS route?

Pentium D's oc'ed on what cooling? And netburst is not a great architecture, which is why Intel is switching to a more pentium 3 esque architecture.

I'm not buying into the Dual Core stuff, just yet and if I do, it will certainly be Presler........never amd

A bit biased?

And if you think (amd) they're fast at gaming? think again! BENCHMARKS SUCK. I run my CPU at 4,020mhz (<---notice, no CPU-Z ss there, unlike you know whom) which is plently enough for "any" of todays apps. Zero lag, highly reliable, stable and overclockable. 2.2 with a 2.8 max?...nah I don't think so, old school is better......4.0+ghz.

If you deny the fact that amd's are the best gaming processors obviously something is wrong, this has been stated at every single hardware site. Also, % overclock is what matters, not overall MHz. This is pretty obvious.

Integrated Memory Controller such a great idea? for what? games?......hhmmm.....Pentium M with a 490mhz (more) overclock from stock 2.26 is faster than the almighty X2 4800+ with the latest and greates amd has to offer.......see here please. and obviously is not that great for office tasks nor multimedia nor price either.

the x2 is not intended for games. If you want a amd processor intended for pure gaming right now your best bet is single core. But the Pentium M is faster in gaming, that is undesputable.and note, the pentium M does not rely on the crappy netburst arc.

oh! Crap, back on topic.......Osirus35, here's a link on how Conroe will be based on Banias and Netburst: Intel Conroe/Woodcrest starts with over 3 GHz. <--noticed how that thread has no replies, because it scares "them".

WTF? How does that make them based on netburst? Because they're being released at over 3GHz??? And Banias isn't a architecture lol

I can't wait to get a 65nm Presler and overclock the heck out of it......which brings up a question? Does amd have any plans for 65nm CPUs? My palm reader, the most reliable source I have for news on CPUs and technology.......because Anand which I'm pretty sure he doesn't have a clue what ViiV is yet (and THG are) is like Donnie says: AMD cheerleaders, tell me otherwise: AMD and 65nm CPUs, another delay.

You answered your own question. "Does amd have any plans for 65nm cpu's?" ROFL.

I wish this side of the forum would be lock from amd zealots, every thread I see pretty much end up being the say way, I've never posted in any amd thread nor I ever will because I could careless anything related to amd unless it's bad news

Actually arguing with someone is a good thing. When you dont argue is when zealots are created .Perhaps that's why there are zealots like me =P

So far the "most" civilized forum I've been is OCForums.com. I sure contribute in this amd vs Intel yet again another thread but it's quite ridiculous to think Netburst is not good, it's the greatest! Can you said Pentium D @ 5.2ghz? .

and to think what temps it has etc


rofl you're funny. and it would still get beat in gaming by a venice on air I'm not anti intel, just anti netburst, though it's had it's days (the c series in particular was a nice series)

 

[Magik Smoke]

Bleh! Fan boys in general suck ass.

I will buy whichever processor offers the most bang for the buck. Which is why I OC...

I’m starting to despise both AMD and INTEL because they BOTH have gone to great lengths to confuse the VALUE issue.

 

[sabrewolf732]

Cedar Mill, Prescott VT and Presler are still Netburst.

So? They are the last of netburst, and if you can read the article intel won't be pushing them much.

Nah!, I just don't like amd based systems due to their highy unreliability and unstability rate. I don't drink Pepsi, it's too sweat, I don't like Pizza Hut, it's too greasy.....biased? No! it's called "choice" and because I don't and will never get an amd based system doesn't mean "bias".

Yes it is a bias. You are uninformed. There are no stability issues. Acutally there are stability issues with EVERY platform.

Best? being best doesn't mean you can't do it with an Intel. Every Intel based machine I've built is capable enough to run any current game on the market...........that's why you have video cards such as the 7800 GTX and GT.
You must be one of those persons whom think that an FX-57 with a 6600 GT is fater than a Pentium 4 with a 6800 Ultra......don't you? So much for CPU power for gaming, isn't it?

Now you're just trying to avoid the facts. IT IS THE BEST AT GAMING. That is undesputable.

Right, noticed also that Pentium M does not rely on the crappy "on-die memory controller".

Not really crappy as due to this AMD has very low latency and is much more unified. What makes it crappy? you're just being a !!!!!!, "if it's from amd, it's crap". That's how you're thinking.

Well, you have to click on the links and you'll see this: Netburst + Banias. Banias not an architecture you said?..........

Banias is not an architecture, it's a pentium M core. Preceeding the Dothan.

I just like to deliver the bad news, I could careless if they ever do but I'm sure it does for those whom think the little engine that can.....can't!

What bad news did you deliver? that amd has the fastest desktop processsors right now?

Arguing = wasting your time. like you do for the most part and adds nothing constructive to the point.

WTF? How does it not add anything constructive. When you argue is how you make progress.

I may be funny but it's better than to be an ignorant.

You're trying to call me ignorant? I advise you go to dictionary.com and look up the word hypocrite.

 

[Donnie27]

If you think Anand did a biased job.

http://www.tomshardware.com/cpu/20051007/index.html

Today, AMD has the lead in both the single and the dual core desktop segments thanks to processors that not only perform better in most typical benchmarks, but they also require a whole lot less of power to do the job. This is obviously why Intel now focuses on platform solutions rather than on components. However, we have to emphasize that this is the clear trend for the whole hardware industry, since hardware performance is something that is sufficiently available to everybody today. Even a low-cost processor will be fast enough to power the vast majority of today's applications. Convincing people to buy something new requires the vendor to show what can be done with the new device.

They must have drank a few too many Andechs back in 1994 there used to be an old saying when it comes to computer Hardware. "You don't win a race by leading a few Laps and the fastest Car doesn't always win." Say again, AMD sales didn't pick up until Semron production picked up. Hint! So common users aren't Clueless, they buy what they need to surf the web, make and send e-greetings cards, use word and most business users do the same. That's the real problem both AMD and Intel face is 95% of users don't need what most Geeks think these users need. Intel's advantage is in Volume. They, unlike AMD, don't need to have a higher mark up on EACH processor they sale. The whole Performance Crown thing is more of a Joke than anything. Both are in Business to make money.

What Tom's does show is;
It Is Not Just A Die Shrink! Just as I said. L1 went from 12 to 16KB for starters.
Gigabyte GA-8I955X Royal no new motherboard is needed beyound that. I have sneaky feeling Ceadar Mill will work fine on 915 and 925. Presler is another story though.
Intel will make a killing on them. 300mm wafers with how many 65nm processors on each?
Intel made a smooth transistion to 90nm and now they'll do the same with 65 and no reason to believe 45nm will be a problem either.

None of this needs to degrade to just another AMD vs Intel shouting match.

Donnie27

 

[(cf)Eclipse]

I'm not anti intel, just anti netburst

same. i'm actually considering an intel rig to play with some ddr2.. if i had the money for it.

The Rev-E overclocks better because of relaxed timings. It runs slightly slower at Stock in most apps because of these relaxed timings.

incorrect:

What did the engineers focus on this time? Of course, they again made a few optimizations in order to improve the memory controller performance. The tests of Venice based processors showed that they were faster than their Winchester based counterparts.

http://www.xbitlabs.com/articles/cpu/display/athlon64-e3-mem.html

http://www.xbitlabs.com/articles/cpu/display/athlon64-venice_8.html
http://techreport.com/reviews/2005q2/athlon64-venice/index.x?pg=4

it actually runs a weeee bit faster than it's older siblings at a given clock

also, your logic for saying that a san deigo is the same as a venice with more cache isn't quite true. if it was, the bandwidth numbers between the two would be the same. they're not. there are slight tweaks that give venice the edge for some reason.

but i digress, this is an intel thread. i'll stop talking about amd now

you said that they are in business to make money. it's common practice to show off a really fast rig and say something along the lines of "this computer has a processor that is directly related to this awesome performance award winnner blah blah blah". it's a surprisingly powerful tactic.. and one of the reasons both intel and amd have stuck with the pentium and athlon names for so long. same goes for ati and nvidia too

 

[sabrewolf732]

same. i'm actually considering an intel rig to play with some ddr2.. if i had the money for it.

same here, DDR2 is so god damn cheap! If I could get a decent matx ddr2 mobo and a p4 for a decent price I would prob pick a p4 up

 

[Donnie27]

All Intel has to do is cool down Prescott and with Cedar Mill, they seemed to have done just that. No matter how many AMD leaning folks say otherwise, it will just be another hit. Overclockers will try, like and buy, it will hold Intel over to the real* processors show up and they will make even more money.

It sounds to me that there are bitter folks that Intel beats AMD to yet another milestone, 65nm. They'll be close to shipping 45nm when AMD finally ships 65nm. If AMD takes two years to RAMP it up like they did with 90nm, things will look a lot differently.

incorrect:

http://www.xbitlabs.com/articles/cpu/display/athlon64-e3-mem.html

http://www.xbitlabs.com/articles/cpu/display/athlon64-venice_8.html
http://techreport.com/reviews/2005q2/athlon64-venice/index.x?pg=4

it actually runs a weeee bit faster than it's older siblings at a given clock
but i digress, this is an intel thread. i'll stop talking about amd now

you said that they are in business to make money. it's common practice to show off a really fast rig and say something along the lines of "this computer has a processor that is directly related to this awesome performance award winnner blah blah blah". it's a surprisingly powerful tactic.. and one of the reasons both intel and amd have stuck with the pentium and athlon names for so long. same goes for ati and nvidia too

You might want to see real world tests of the two, I saw results at overclockers, Xtremesystem and others saying some different than XBits.

They're in business to make money, no spin needed=P

http://www.devhardware.com/c/a/Computer-Processors/AMD-Venice/3/

Oops! Look at the frequency. That memory controller doesn’t look like it is quite fixed yet. This was the auto setting in the BIOS for the memory speed. I next tried to set it to 200 MHz, and Windows couldn’t adjust. Next I put it back onto auto, and then got into windows, and with A64tweaker, I set the speed from 166MHz, to 200MHz, fined up CPU-Z. Same results, like I never even changed it. So much for that improved memory controller; it looks to me like they broke it.

http://www.anandtech.com/mb/showdoc.aspx?i=2264&p=3

As we discovered most recently, AMD's 90nm CPUs are of a brand new revision that featured a number of bug fixes and performance improvements. Internally AMD refers to this CPU revision as Revision D. However we noticed something very interesting when looking at the latest AMD roadmaps while in Taiwan: the 90nm chips listed as San Diego and Venice (Athlon 64 FX and Athlon 64 respectively) claimed SSE3 support as a feature, but the current 90nm chips do not have SSE3 support.

Internally, AMD's San Diego and Venice CPUs are referred to as nothing more than Athlon 64 Revision E chips. Revision E includes even more bug fixes and performance improvements over those we found in Revision D, including support for the 13 new instructions that were added with Prescott, more commonly known as SSE3.

In the 500MHz DDR story they wrote "we took a look at the performance improvements offered by SSE3 support and came up empty handed ."

Now that's improvements and fixes to an already perfect processor uh?

http://www.vr-zone.com.sg/?i=1123

Rev E has better mismatched DIMM support and has 2-beat DRAM timing for 939-pin processors. Rev E also includes better power management features and an enhanced DRAM controller. Indeed, Rev E will come with SSE-3 support with 11 of out 13 instructions present in Prescott excluding MONITOR and MWAIT. Other improvements include improved HyperTransport bandwidth for 940-pin at 2GT/s, Adaptive Prefetch to optimize cache hit rate, two additional write-combining buffers (total 4), XOR DRAM bank address to prevents page-thrashing on cache write-backs and an expanded clock ramp hysteresis counter.

Similar story at 2CPU and Geek.com..

That's (page-thrashing) similar to the Linux cache flush bug. So jump on someone else, I didn't make this crap up. The Guys at Xtremesystems are 10 to 1 AMD followers. They and some at Extremetech talked about the seemingly relaxed Memory Controller on Revision E. The all 4 the slots full bug was found a long time ago and some folks are still having problems with at least some memory on the nForce 4 with Rev-E. Others complaining about this was Tech-Report, GamePC, Lostcircuits, Ace's and etc...

This is what happens when sites look to cater to Ph@ns instead instead giving them info. Sure they can look at the forum index to see 83 to 13 and cater to the 83, but at least IMHO, [H] has the balls to NOT pull a XBits, Anantech, or Tom's to steal readers from the AMDZone or real Fansites (in this case Fansites shouldn't be edited).

You search and see how many hits you get?

Donnie27

 

[sabrewolf732]

I have to agree, almost everywhere I read venice performed a tiny tiny bit worse than winnie. But that's not why venice's oc so good. I will actually prob buy a 65nm p4 since it seems like they got heat under control. If it can hit 4.5GHz on decent air and keep the temps down I'll be happy Maybe the 65nm stuff will be the next "c" series.

 

[Duke3d87]

Oh!....Enlighten us with your wisdom on how Netburst is so different from the Northwoods to the Prescotts?

It's not really that different. I think what he was saying is the application of the NetBurst architecture. The Pentium 4C is known to be the best Pentium 4 core ever made. It gave AMD a serious run for its money. The 20 stage pipeline's penalties were slowly being masked and the muscle of the NetBurst architecture in multimedia began to filter through. Then, they released Prescott. It ran hotter, had a pipeline that was 11 stages longer and had more cache. The L1 cache was said to be slower, and desipte the larger cache, failed to outshine Northwood. The Prescott core is pretty heavily tweaked, but much of the benefits are masked by the 31 stage pipeline.

if your only respond is going to be "heat", I dobut it, my 650 idles at 34c and highs in the 51c to 53c overclocked at 4.0ghz and at stocks it does 31c. 44c at 4.2ghz max on air. Sure not using Intel's "crappy" Heat-Sink/Fan but a Zalman CNPS series......Power consumption? again, nothing a good quality PSU can't handle. 300x14 common on water, yes I can get my 650 @ x14 multi with my P5WD2-Premium with 0519 BIOS. 250X14 = 3.5ghz and look at the voltage? Stock baby, stock! . I've had it to 260x14 and temps still way under control. Next it'll be CoolerMaster Mini120 water kit and off I go to 300x14.

And when he says heat, look at the TDP. Intels have a record TDP. They level out at 130 watts and that's not the actual power consumption. It is the amount of heat that any given HSF should be able to dissipate. And, i've heard different stories about the heat. Some say it's not that bad while others day that it's horrible. All this will soon be in the past when Conroe awakens.