According to Anandtech, Conroe beats FX-60

[TooNice]

It seems to me that when the Athlon came out initially as a 650mHz and then 700mHz, that the Pentium !!! was still at 550mHz.

I am pretty sure it started at 500mhz actually, but in very small quantity.
600Mhz P3 was available at the time of the K7 700Mhz, but by then AMD has already established itself as the new performance leader. Clock per clock the AMD was better (though not hugely in contrast to the P4), but it also scaled better than the P3 architecture Intel had at the time.

It was exciting time indeed to see the underdog win a significant victory.

Just wondering though, do we know if the Conroe will require a new mounting system? i.e. Are the current HS compatible with it (without modding)? The impression I have is that things won't be getting much hotter than the Prescott, so a high end HS today should be able to cope with the Conroe. But will they be compatible?

[JazzJackRabbit]

Clock per clock the AMD was better (though not hugely in contrast to the P4), but it also scaled better than the P3 architecture Intel had at the time.

Clock per clock AMD was not better until they released thunderbird. First athlon generations were actually slower to MHz equivalent Intel CPUs. And even after thunderbird release I'd say they were tied performance wise.

[NeilBlanchard]

Hello,

You must be refering to the Coppermine P!!!? Because the early P4 required at least 400-500kHz to even equal the performance of the Coppermine or the Athlon.

As I remember it, I bought an IBM Athlon 550 machine in just about the first 2 weeks they became available, and that was the slowest model of three that they sold at the time. The 700mHz came out relatively soon afterwards. The fastest P!!! at the time was the 500mHz, or maybe they came out with the 550mHz, but it cost a lot more than the Athlons. Intel totally underestimated the Athlon's speeds. They might have realized that since the DEC Alpha was so darn fast...

[kenji]

"i will never buy intel"

just wanted to start with that right off the bat. I think AMD is a better company for more reasons than just superior products. Or maybe that's just because Intel is a worse company for more reasons than just inferior products..

Let's see, I'm only twenty one, but if I remember correctly..

Commodore64
Intel/AMD/Intel/AMD
blah blah i don't remember I was like 4.
Intel Pentium, Then a Pentium PRO, a few amd chips (geez they are probably still all in a box somewhere)
Even maybe a "CYRIX 686!!!" .. laff
Fast forward.. I started building systems at like 10?..
AMD K5
AMD K6.. As I recall a lot of these were running off the same S7 or SS7 mobos as the pentiums I took out to replace with.
AMD K62.. was it a 400? I remember oc'ing it to 420.. kids..
AMD K62 500. The k62's all ran Quake 1 Team fortress (including GLquake just fine, with my graphics blaster or whatever it was and eventually my wonderful Voodoo3 3000 agp <AGP WOW
AMD athlon 1 ghz
AMD athlon 1.2 ghz
AMD athlon TBIRD 1.4 ghz
AMD duron lost in there somewhere
Dad hand me down DELL inspiron laptop, 2.4 ghz I'm typing on
AMD Athlon XP barton 35w 2400+.. HTPC

and the next chip I purchase will certainly be an AMD, regardless of the performance leader at the time. The motivation for my loyalty to the company has changed, but the loyalty itself has only gotten stronger. I won't clarify exactly because I'd just start another OT flamewar "IT HASNT BEEN PROVEN BY A COURT OF LAWWW."


On the Energy star issue:
Why shouldn't we be getting more performance at lower wattages? Look at the advance in speed/power over just MY lifetime, why not lower energy? Some would say this is because the chipmakers are focused on what consumers want, but what about large scale servers etc that require less power for a smaller bill? I think the government stepping in will be the perfect catalyst for these chipmakers to really start firing up the grill when it comes to efficiency.

-Kenji

[kenji]

Maybe "firing up the grill" was a poor choice of words.. haha. How about: firing up the high efficiency grill which uses new energy efficient gas with a high combustion to waste ratio? now that just sounds funny, Let's just say chipmakers should start putting thier energy efficient thinking caps on

[TomZ]

Processors are not religion. There is no need to devote yourself to AMD or Intel. There is no need to preach to others to do the same. They are just processors - not a life decision - not something that defines who you are - not something that really affects your life. I think it's fine to have a preference - but blind faith in a company is misguided.

Both AMD and Intel, and 99.9% of other companies, are out to make a profit. They don't really care about you, or about me, except so far as to get you or I to purchase their products. AMD is not "good" and Intel is not "evil." They both have the exact same objective: profit. Both companies have lots of smart, motivated people working hard and doing their best. Both companies make strategy and execution mistakes, just like all real companies out there.

It's fun to debate about which is better, but you need to keep it in perspective.

[stupid]

Processors are not religion. There is no need to devote yourself to AMD or Intel. There is no need to preach to others to do the same. They are just processors...

It's all about efficiency and performance, and it is best to choose which processor most suits your own needs. For now I would say AMD has Intel over a barrel, general speaking. But Intel's Conroe/Merom looks very promising and will most like turn the tables on AMD. This is a good thing for everyone, including AMD cultists because it is part of competition and will only further better products down the road.

Now if you will excuse me, I am late for my hourly ritual of praying before my nVidia altar.

[jaganath]

AMD is not "good" and Intel is not "evil."

The existence of AMD is good (insofar as they provide competition for Intel and prevent them establishing a monopoly on the chipmaking industry). The possibility that Intel will become a monopolist company is an evil, or at least disturbing, prospect.

It's fun to debate about which is better, but you need to keep it in perspective.

It's easy to make glib rationalisations such as "whatever decision I make personally, it doesn't matter", this is in large part the mentality that has caused much of the pollution and resource wastage that goes on in the world. Whether we like it or not, the choices we make as individuals have consequences, and it is up to the individual to decide whether their conscience allows them to live with those consequences.

[TomZ]

The existence of AMD is good (insofar as they provide competition for Intel and prevent them establishing a monopoly on the chipmaking industry). The possibility that Intel will become a monopolist company is an evil, or at least disturbing, prospect.

I totally agree - and if there were three companies, even better!

It's easy to make glib rationalisations such as "whatever decision I make personally, it doesn't matter", this is in large part the mentality that has caused much of the pollution and resource wastage that goes on in the world. Whether we like it or not, the choices we make as individuals have consequences, and it is up to the individual to decide whether their conscience allows them to live with those consequences.

I totally agree. But what kenji said was, "and the next chip I purchase will certainly be an AMD, regardless of the performance leader at the time." This runs counter to making an educated choice that would benefit consumers, society, etc.

Anyone who feels that AMD has always used power efficiently (i.e., an ethical argument about usage of resources), please come over and feel my old space heater, which was actually my dual Athlon MP system. That computer literally heated my office in the winter (I closed the heating ducts in the room) and drove me crazy in the summer. Again, not making any judgements about good or bad, just saying that AMD also made hot processors.

[TooNice]

Clock per clock AMD was not better until they released thunderbird. First athlon generations were actually slower to MHz equivalent Intel CPUs. And even after thunderbird release I'd say they were tied performance wise.

Battle at 600Mhz.
I didn't count, but I am sure the Athlon beat the P3 more often than not... and was on par about half the time with the overclocked P3 thrown in the test.

Now for SSE optimised application, it might be true but I am still inclined to say that the early Athlon already had a small clock per clock advantage over the P3.

[I am also not a hugely loyal consumer. Whoever get the best product for my budget gets my money. But I admit that it does put the underdog under pressure: the general public probably go for Intel regardless, whereas the non loyal enthusiast will only consider the underdog if the later produce something better at same/lower cost]

[Beyonder]

For me, it's all about the cheapest, most efficient processor available. I don't care who makes it. I just want to reap the benefits.


That being said, I neither trust nor care about the Conroe benchmarks. I'll believe it when I see the real thing, and until then, they might as well be claiming the Easter Bunny had sunday brunch with Elvis while pigs catapulted through the sky. The proof is in the pudding, and I don't see any pudding.

[Erssa]

HKEPC has done some AM2 vs s939 comparison benchmarks. Judging from them, it looks like the Intels prediction of "fx-62" performance will be pretty close to truth.

[jvrobert]

HKEPC has done some AM2 vs s939 comparison benchmarks. Judging from them, it looks like the Intels prediction of "fx-62" performance will be pretty close to truth.

Wow, don't know if this has been discussed yet but this is extremely sweet. Guy has Conroe running _passive_ cooling, and it's posting exceptionally high performance numbers.

Here's post where he says he's running passive, but see whole thread for performance numbers. We're talking fx-60+ numbers from a middle end CPU, with passive cooling.

http://www.xtremesystems.org/forums/sho ... tcount=580

[MooseX]

Wow, don't know if this has been discussed yet but this is extremely sweet. Guy has Conroe running _passive_ cooling, and it's posting exceptionally high performance numbers.

Here's post where he says he's running passive, but see whole thread for performance numbers. We're talking fx-60+ numbers from a middle end CPU, with passive cooling.

http://www.xtremesystems.org/forums/sho ... tcount=580

Ever heard of this thing called 'viral marketing'?

http://en.wikipedia.org/wiki/Viral_marketing

[TomZ]

Ever heard of this thing called 'viral marketing'? ;

I was totally expecting someone to suggest something like this. After all, AMD has a monopoly on making good products, right? We just can't accept at all the possibility that Intel may have a winner on their hands, can we?

[jvrobert]

Wow, don't know if this has been discussed yet but this is extremely sweet. Guy has Conroe running _passive_ cooling, and it's posting exceptionally high performance numbers.

Here's post where he says he's running passive, but see whole thread for performance numbers. We're talking fx-60+ numbers from a middle end CPU, with passive cooling.

http://www.xtremesystems.org/forums/sho ... tcount=580

Ever heard of this thing called 'viral marketing'?

http://en.wikipedia.org/wiki/Viral_marketing

Haha, yeah, I love when new buzzwords are born. Everyone tries to apply them in all kinds of wacky situations. "Viral Marketing" and "Meme" are the current buzzwords du jour. People just like to say the words, I think, makes them feel like they're in a William Gibson cyberpunk noverl or something.

Anyway, the relevence is clear (if he's not lying or misstating) to a silent PC forum, especially since you can buy nearly this performance/low wattage in a Core Duo _right now_.

[stupid]

Conroe's preliminary benchmarks were good enough for me to postpone my planned upgrade. How will the retail version of Conroe perform? No one knows exactly. If for some reason Conroe actually falls flat on it's face, then I'll upgrade to a Socekt AM2 Athlon X2, and Intel will have hell to answer for as their reputation has a lot riding on their Pentium 4 replacement.

I'll just wait on the sideline for a few months for official reviews to come out with some hard data. Until then my good 'ol trusty Athlon XP-M 2600+ is still my PC workhorse.

[frostedflakes]

I don't think a passive Conroe is that hard to believe, TDP for this line is 65w (although I think XE TDP shoots up to like 95w -- maybe they plan to run a lot of extra voltage through them to hit such high clocks?).

Although I think at 65nm AMD will be able to put up very impressive power numbers, probably better than Conroe's, thanks to technologies such as SOI, DSL, etc. But in the end (for us, at least) it will come down to performance per watt, which I'd think they'd be pretty similar in or if anything Intel pulling slightly ahead.

Just my humble $0.02, though.

[Shining Arcanine]

Meanwhile, AMD's "wide" approach, with it's nine-issue design seems to be the better one. They invented the x86 64bit instruction set, and a well set up for dual/quad cores with Hypertransport, onboard memory controller, etc.

AMD has a three-issue design. If it had a 9-issue design, you would see the Athlon outperforming the Itanium.

By the way, Intel had a 64bit x86 processor architecture with an integrated memory controller in development long before AMD even thought of it. It was called the P7 architecture. Intel canned it when they started developing the Itanium. If anyone invented 64bit x86 instructions, it was Intel; they were just too stupid to go ahead with them.

I still think that unless AMD screws their die shrink up pretty seriously - AMD will at worst give Conroe a run for it's money - Mind you my old uATX Gigabyte system which is only a 2.8 northy with a 9800xt radeon runs just
about everything out there including the latest games at both reasonable frame rates and reasonable speed all be it without the bells an whistles. My newest build an AMD X2 with all the new B.S - gets a few more FPS - can run with all the bells an whistles - is loud - and has been problematic - although all the issues have been solved They should not have been there to start with.
Are all these "improvements" benifiting the consumer ? I think they need to get their stuff tested BEFORE they sell it ! Hope they get it right this time .

How? AMD needs to scale their architecture to 3.5+ GHz just to compete with a mid-range Conroe. How on earth will they give Conroe a run for its money?

[MikeC]

How? AMD needs to scale their architecture to 3.5+ GHz just to compete with a mid-range Conroe. How on earth will they give Conroe a run for its money?

It's far too early to talk about what future CPUs from either Intel or AMD will do. Until production samples of new products from both companies have been tested by a reasonable number of quality hardware review sites, this kind of comment is unfruitful. Please refrain from starting brand wars.

[TomZ]

this kind of comment is unfruitful

With all due respect, I find the discussion interesting, and I think the comments that were made are justified and are not inflammatory. The recent Conroe benchmarks seem to show that AMD is not going to keep ahead of Intel with just small incremental improvements. I'm not really sure what drew your ire.

[Beyonder]

What benchmarks? The benchmarks for the processor that you won't be able to purchase for months ("sources" say 2H 06, and more recent "sources" say Q4) which were conducted by the company selling the product? There's another buzz word commonly associated with that type of product.

MikeC's comment is extremely valid: there is little merit in counting your chickens before they hatch.

[stupid]

There's another buzz word commonly associated with that type of product.

I wouldn't actually call Conroe vaporware since it hasn't even missed its launch date yet. As far as I know, Conroe on course to be delivered to OEMs in Q2 '06, and available for retail sale in Q3 '06.

What Intel did is more of a teaser. I wouldn't consider anything to be vaporware unless it is seriously delayed. Duke Nuke 'Em Manhattan Project comes into mind. Even if Conroe slips one quarter it is more of a delay and shouldn't be labeled as vaporware.


While the initial Conroe benchmarks certainly do look enticing, and I've read that an independent 3rd party has confirmed the overclocked FX-60 is a realistic estimate of future Athlon 64s based on K-8 (can't remember the source), it is still preliminary. All the tests result were controlled by Intel, that's not say that they manipulated the numbers; just that they picked the benchmarks that most likely showed the biggest performance gap.

Unlike the last couple of years 2006 represents a fundamental shift in the CPU arena. Intel is introducing a new architecture that can execute 4 instructions per cycle and has basically abandoned the Pentium 4 design in favor of something more akin to the Pentium M. While AMD is not releasing a new core this year, it is laying the ground works for their K-8 successor with Socket AM2 and the adoption of DDR2; actually more like finally getting on the bandwagon since memory manufacturers are shifting away from DDR. AMD is also releasing a 35w TDP version of the X2 3800+ on socket AM2, and the X2 4200+ thru X2 4800+ will have 68w (or somewhere near that) TDP this year.

I for one have decided to postpone my upgrade (Athlon XP) until the retail release of Conroe. Will Conroe live up to it's teaser benchmarks? Will it run far cooler than the Pentium 4 space heater? Who knows? If it flops, then I'll simply upgrade to Socket AM2.

[quikkie]

By the way, Intel had a 64bit x86 processor architecture with an integrated memory controller in development long before AMD even thought of it. It was called the P7 architecture. Intel canned it when they started developing the Itanium. If anyone invented 64bit x86 instructions, it was Intel; they were just too stupid to go ahead with them.

I did some googling and found 3 mentions of intel and P7 in the same sentence, a chip that never left R&D design docs back in the early 90s (unknown if it had on-die mmc) which was canned long before itanium was dreamt up, a chip marketed as having netburst and the merced core (itanium). which is it? and do you have any links for me to read?

I read that MS pressured intel to adopt AMDs x64 extensions so that there weren't two different and probably incompatible extensions to support Otherwise the original P7 instruction set may have seen the light of day after all.

[stupid]

I read that MS pressured intel to adopt AMDs x64 extensions so that there weren't two different and probably incompatible extensions to support Otherwise the original P7 instruction set may have seen the light of day after all.

Not sure if "pressured" is the right term. There were two different x64 extensions, MS simply decided that AMD's x64 was better.

[Beyonder]

There's another buzz word commonly associated with that type of product.

I wouldn't actually call Conroe vaporware since it hasn't even missed its launch date yet.

I actually somewhat agree--it doesn't fit the exact definition of vaporware. That being said, Conroe is "vaporware" in the literal sense, because it is not a product anyone can actually purchase, depite already being heavily marketed via Anandtech. Posting illegitimate benchmarks of a product that isn't yet available is a means of preventing people from purchasing more enticing options now, and there has been wild speculation about the release date.

There has been no official launch date from Intel. Dates range from the wildly optimistic Q2 (in a month and a half, and no official launch date yet? Hmm...), while others have speculated as late as Q4 (in other words, October).

Here is an article specifying server chips in Q3 of 06, which implies that you're not going to see these in a desktop in the very near future--definitely not in Q2. Then again, I shouldn't fan the flames: all of this is just blatant speculation.

While the initial Conroe benchmarks certainly do look enticing, and I've read that an independent 3rd party has confirmed the overclocked FX-60 is a realistic estimate of future Athlon 64s based on K-8 (can't remember the source), it is still preliminary. All the tests result were controlled by Intel, that's not say that they manipulated the numbers; just that they picked the benchmarks that most likely showed the biggest performance gap.

Consider this: we may be having this conversation all over again when AMD demonstrates some future architecture that stomps all over Conroe in September. Wild speculation will ensue. Consumers will delay purchasing descisions based upon said wild speculation.

I for one have decided to postpone my upgrade (Athlon XP) until the retail release of Conroe. Will Conroe live up to it's teaser benchmarks? Will it run far cooler than the Pentium 4 space heater? Who knows? If it flops, then I'll simply upgrade to Socket AM2.

...you just made some Intel marketing executive smile.

[Shining Arcanine]

What benchmarks? The benchmarks for the processor that you won't be able to purchase for months ("sources" say 2H 06, and more recent "sources" say Q4) which were conducted by the company selling the product? There's another buzz word commonly associated with that type of product.

MikeC's comment is extremely valid: there is little merit in counting your chickens before they hatch.

http://www.anandtech.com/cpuchipsets/sh ... spx?i=2738

Socket AM-2 does not significantly improve performance and there are Conroe benchmarks of engineering samples at xtremesystems.org:

http://www.xtremesystems.org/forums/sho ... hp?t=95021

By the way, Intel had a 64bit x86 processor architecture with an integrated memory controller in development long before AMD even thought of it. It was called the P7 architecture. Intel canned it when they started developing the Itanium. If anyone invented 64bit x86 instructions, it was Intel; they were just too stupid to go ahead with them.

I did some googling and found 3 mentions of intel and P7 in the same sentence, a chip that never left R&D design docs back in the early 90s (unknown if it had on-die mmc) which was canned long before itanium was dreamt up, a chip marketed as having netburst and the merced core (itanium). which is it? and do you have any links for me to read?

I read that MS pressured intel to adopt AMDs x64 extensions so that there weren't two different and probably incompatible extensions to support Otherwise the original P7 instruction set may have seen the light of day after all.

I do not quite understand what you mean by "which is it?" I only said one thing. The P7 architecture was in development long before AMD thought of 64bit processors with integrated memory controllers but was canned the moment Intel began developing Itanium. I learned that from a tibbit I read at AcesHardware.com. It seems to have since been taken offline/made inaccessible to Google through an archiving mechanism so I cannot locate it, but that is what I read.

[Erssa]

While the initial Conroe benchmarks certainly do look enticing, and I've read that an independent 3rd party has confirmed the overclocked FX-60 is a realistic estimate of future Athlon 64s based on K-8 (can't remember the source), it is still preliminary. All the tests result were controlled by Intel, that's not say that they manipulated the numbers; just that they picked the benchmarks that most likely showed the biggest performance gap.

Consider this: we may be having this conversation all over again when AMD demonstrates some future architecture that stomps all over Conroe in September. Wild speculation will ensue. Consumers will delay purchasing descisions based upon said wild speculation.

I doubt we won't have this kind of conversation. We know from the roadmaps that AMD won't be releasing anything this year, that could stomp Conroe. Sure the current architechture is great and it will prove itself great again in the future, but not during this year, probably not even next year with 65nm, it will probably take quadcore for AMD to gain the performance crown back. Check the latest Anandtech AM2 preview. It's like they say:

"The disheartening news for AMD and its fans alike is that if AM2 can't offer significant performance increases over what we have now, then all Intel has to do is execute Conroe on schedule, delivering the performance we've been promised and 2006 will be painted blue. AMD has been telling us that 2007 is the year we'll see major architectural changes to their processors, so AM2 may very well be as good as it gets for now. That's still very good, of course - the fastest X2 chips still outperform the fastest Pentium D chips - but it looks like after three years K8 may finally get some competition for the performance crown."

I for one have decided to postpone my upgrade (Athlon XP) until the retail release of Conroe. Will Conroe live up to it's teaser benchmarks? Will it run far cooler than the Pentium 4 space heater? Who knows? If it flops, then I'll simply upgrade to Socket AM2.

...you just made some Intel marketing executive smile.

Yeah he probably did. I think it was the whole purpose of the Conroe benchmarks.

I have to agree with many people here. Those Conroe benchmarks will hardly reflect the situation when Conroe is released. The performance lead for Intel will likely be even greater since Conroe has more time to develop before the release then AMDs AM2 have. AM2 should be basically done now as it's release is less then 2 months away... I am actually glad that AM2 doesn't offer performance benefits. It prevents my x2 3800+ from dropping in price .

[stupid]

I actually somewhat agree--it doesn't fit the exact definition of vaporware. That being said, Conroe is "vaporware" in the literal sense, because it is not a product anyone can actually purchase, depite already being heavily marketed via Anandtech.

Can't blame Intel if someone else is cheerleading.

There has been no official launch date from Intel. Dates range from the wildly optimistic Q2 (in a month and a half, and no official launch date yet? Hmm...), while others have speculated as late as Q4 (in other words, October).

Here is an article specifying server chips in Q3 of 06, which implies that you're not going to see these in a desktop in the very near future--definitely not in Q2.

Ah, rumors, rumors, rumors...

Consider this: we may be having this conversation all over again when AMD demonstrates some future architecture that stomps all over Conroe in September. Wild speculation will ensue. Consumers will delay purchasing descisions based upon said wild speculation.

That will always be the case. However, considering my Athlon XP will be nearly 5 years old by the time I upgrade, it be either Conroe or AM2; whichever CPU offers the best performance for around $600.

[Beyonder]

I doubt we won't have this kind of conversation. We know from the roadmaps that AMD won't be releasing anything this year, that could stomp Conroe.

This is exactly my point. AMD and Intel demo stuff that won't be available for months (in the case of Conroe, I bet it will be six months between benchmark preview and RTM). You're presuming I'm talking about AM2--I am not. September would be idea to demo something you're going to release in six months (i.e. 2007).

My comments have nothing to do with AM2, and everything to do with the fact that AMD and Intel love showcasing stuff that mere mortals cannot purchase, which is why these conversations are stupid, and also why we will have this conversation all over again.

And, with that, I sign out. Been fun.