Conroe: The truth?

[AgentVX]

and just what is making you think scores will scale 100%

Of course it won't, but in his comparison you should also take into consideration the gain from overclocking a Conroe as opposed to running it at stock 2.4GHz. That would obviously give it another boost in addition to the normal xx % you get from SLI/Crossfire. It still may not end up being 100%, but my point is there's the CPU overclocking factor too.

[RPGWiZaRD]

I think ppl in here forgot one very basic thing. Let's forget about the numbers for now and just think some very basic thoughts and forget who will own who.

This is competition, and in competition it's only natural for the competitors to swap the performance crown every now and then. A company not ever being able to take the performance crown is a failing company imo. Let's say Conroe might win specific 20 random software benchmark battles in this 65nm vs 90nm battle which would only be natural. It's about time we see such changes. Now in 2007 when AMD launches their own 65nm chips perhaps we see the slight performance favor for the AMD chips and who knows what Intel will counter with then. I'm only saying one company can't always have the lead or there will be no competitors. Intel might have the lead for 6months, AMD might have the lead the 6months after that etc, it always keeps changing but fanboys usually don't tolerate this very well.

Other than that, a tough competition is always good, forcing the developers to make even more radical changes and improvements in their chips to take back the lead. Thus, we'll see even bigger performance improvements.

There's one thing I don't really like about Intel at this point tho and it's their marketing strategy, other than that I have no problems with neither of them.

This fanboyism really sux tho.

[DilTech]

I think ppl in here forgot one very basic thing. Let's forget about the numbers for now and just think some very basic thoughts and forget who will own who.

This is competition, and in competition it's only natural for the competitors to swap the performance crown every now and then. A company not ever being able to take the performance crown is a failing company imo. Let's say Conroe might win specific 20 random software benchmark battles in this 65nm vs 90nm battle which would only be natural. It's about time we see such changes. Now in 2007 when AMD launches their own 65nm chips perhaps we see the slight performance favor for the AMD chips and who knows what Intel will counter with then. I'm only saying one company can't always have the lead or there will be no competitors. Intel might have the lead for 6months, AMD might have the lead the 6months after that etc, it always keeps changing but fanboys usually don't tolerate this very well.

Other than that, a tough competition is always good, forcing the developers to make even more radical changes and improvements in their chips to take back the lead. Thus, we'll see even bigger performance improvements.

There's one thing I don't really like about Intel at this point tho and it's their marketing strategy, other than that I have no problems with neither of them.

This fanboyism really sux tho.

QFT... Without competition, you just get speedbumps every year without any progress.

[ORCBEAST]

and just what is making you think scores will scale 100%

Get Glasses

DiLTecH:
Exactly, CF/SLi doesn't scale 100% on ANY platform...
In 3dmark05, it only 25-35%.... That's it. Single card of 11k means sli/cf score of about 13-14k.

You havent taken into account overclocking. I was not talking about stock systems. If with adding a extra card will add only 35% in 3Dmark05 then you wont be able to hit 20667 in 3DMARK 05.

this guy is pathetic, plagiator. He took Victor Wang result and post it as his result.

Can you all see- that all is VictorWang number, minus two benchmark with conroe have advantage.

I said, it's a good hoax.

Well put.
No one seems to care that the guy didnt add the extra two benchmarks..But they seem to care when I am suggesting something with it MIGHT score higher or lower..


ADDITIONAL COMMENT:

that doesnt change the fact that he is right... the prices set for conroe are lower then an FX-60 and the cpu is much faster... obviously it isnt available yet, but thats why this thread is in a news forum.

If you don't think it's faster then im not sure what to say... a 2.4 ghz conroe has been shown outperforming an FX-60 in pretty much every bench we can think of. from the looks of it it takes about a 2.7-2.8ghz X2 to be even with the 2.4ghz conroe.

Thx you seem to understand.

CARFAX got in contact with one of the sciencemark guyz. He had this to say.

In any case, I see this as a *strong* showing for Conroe, not a weak one as the blogger claims. In the benchmarks that matter (BLAS, MolDyn, Primordia) Conroe is at least equal to if not exceeding an Athlon at the same clock speed. The fact that in 32-bit mode, an Athlon64 clocked 400 mhz higher cannot exceed Conroe's performance gap is telling

http://www.xtremesystems.org/forums/...&postcount=901

[fhpchris]

Everyone is saying that Conroe will be great, and we get benches like PI to show that, but no one is thinking that it Might be worse in alot of others -- Including 3d03/05/06, once that 4mb cache is trashed, it is on many occasions slower than a X2, as I have shown.

No one is talking saying that that 2.4ghz X2 with a identical video card can pull similar numbers, or damn near identical numbers -- with one of the slowest S939 boards and a 2gb kit, that is prolly running @ 2T. (A8N sli 1T issues)

I am here to say that these stupid FX-60 comparisons are not valid -- because in some benchmarks a simple dual core A64 will be just as fast at the same clockspeed. Many people on these forums will also be saying that "this 299$ 3800+ beats stock Conroe this or that" at some 3d applications.

Do you not understand that nearly every single 03/05/06 score in the top 10 are all AMD for a reason? PI =/= 3D. I dont think it is impressive that a 65nm DDR2 part can just Tie a DDR1/90nm part.

Yes, in benchmarks like Pi, that do not trash the cache, it is much faster than a A64, but in 3D it just isnt.

[Daveb2012]

But honestly, come on. I seen the numbers.. conroe isn't THAT much faster than an FX-60 for example.

well it cost about 75% less.

[Celeron Gamer]

Before you guys go bantering on about WR, and stuff. You realize, that Conroe is not even in public hands, and we do not even have proper support, or knowledge of steppings yet. While you boast about AMD's WR that the S939 platform has been in retail market 1-2 years, and 6+months for the FX-60.

Until Conroe is released, and for atleast 2-6 months for enthusiasts to gain any knowledge, you won't see any WR, that's for sure.

Anyways, that guy has banters the best BS I've ever seen. Biased BS are certainly entertaining!

[zakelwe]

Yes, in benchmarks like Pi, that do not trash the cache, it is much faster than a A64, but in 3D it just isnt.

Actually I tend to disagree with you on that. When looking at cpu performance in 3dmark the best tests are 2001 car and lobby and 03 GT1, comparing Kyosen's Yonah to Mickey Mouse aka Jasons FX-60 we have

(Jason scores first)

car low 673v587
car high 205v174

lobby low 558v505
lobby high 250v225

03 GT1 663v641

Apart from car low that is a maxium of 10% or so in the FX60's favour and considering the Yonah is a mobile chip, is only watercooled and on a non enthusiast motherboard that is not too shabby result. And of course memrom and Conroe are faster. I believe FUGGER has crossfire setup as well as conroe so when all these teething problems with boards and BIOS's are sorted out maybe we will be able to see more clearly. It's looking as if Yonah at least will clock as high as Fx-60 so you have to assume Conroe and memrom will clock as least as high and maybe more so with their netburst/PIII melded new architecture.

I get the feeling this 4mb cache thrashing thing is more a symptom of clutching at straws by people wanting something to be so ( ie FUD ). We shall see though as is said by most rational folk !

Regards

Andy

Regards
Andy

[Xenogias]

conroe cpus will be faster than AMD cpus for most things when it is released. I think you can synthesize all the data available now and pretty easily come to that conclusion.

What concerns me is the platforms they will release it on. If you're concerned about 3d and gaming then you better hope there are some motherboards much better than Intel's P4 lineup. Intel and nvidia and ATI all dropped the ball on making dual graphics boards that scaled as nicely as AMD platforms. I dont know why but they did... you just cant beat s939 for SLI and CF, thats why all the top scores in any modern 3d benchmark (2001 is obviously excluded here) have been AMD for so long not because of the cpu silicon. Think about it this way: You could have the best clocking FX-60 and get it up to 4ghz and set some records but once a new graphics board comes out, you will be beaten by systems running cpus with much much much lower clocks. For this reason when building a new system i always prioritize funds for video hardware and usually get a lower clocked cpu, i think many of you feel this way too--I dont see any benefit in windows from newer cpus unless you're really concerned about that extra 5 minutes it takes you to encode a DiVX movie once a year. Why more people aren't yelling and screaming about the conroe platform I dont know.

[Mikesta]

There is no doubt that Conroe will offer a substantial performance increase over the the 1st generation AM2 A64's. Probably in the order of 15-20% clock for clock. This is especially true for single threaded applications that are cache hungry or light multi-tasking.

However it does seem to me that the more heavily one multitasks the less of an advantage the Conroe will have over the A64.

Imagine a scenerio where 1 plays Oblivion on high settings, and has some heavy distributed projects running at the same time. Will Conroe hold as big an advantage? To my (admittedly little) knowledge probably not. As that (Biased) article pointed out, the Conroe is less efficient under very heavy (cache dependent) workloads. I'm not sure what the author did, nor am I sure how Sciencemark works but intuitively I suspect that if you ran two instances of many of those tests, the A64 will come out looking better.

For the Purposes of members here at XS Conroe will be THE cpu to have, no doubt about it it will hammer the AM2 A64.

However as 'bigger and better' applications get developed, and a greater amount of multi-threading takes place the advantange Conroe has will at the very least be substantially diminished. People are starting to use their PC's differently (alot more multitasking of CPU intenisve programs) and most benchmarks we use here don't take that into acoount.

What we can take away from this is that the A64 is still a better design for the future then even 'Core'. Once Quad-Core's come out I suspect that even an A64 based on todays tech. will do very well against its Intel equivalent. I'm not taking into account K8L. For the AM2 doubters out there Quad-Cores is where AM2 platform will truly make a sense as a platform over 939.

For now Intel looks like it will dominate quite substantially the typical workloads of the current market.

[MaxxxRacer]

What we can take away from this is that the A64 is still a better design for the future then even 'Core'. Once Quad-Core's come out I suspect that even an A64 based on todays tech. will do very well against its Intel equivalent. I'm not taking into account K8L. For the AM2 doubters out there Quad-Cores is where AM2 platform will truly make a sense as a platform over 939.

For now Intel looks like it will dominate quite substantially the typical workloads of the current market.

Well said. I believe this will hold true until Intel rids itself of the FSB and moves to an on die memory controller.

And when quad core time comes around, I think Intel will have two choices.

1. Completely revamp the way cores talk to eachother (not using the FSB as they do now)
2. On die memory controller and crossbar switch.

[Ailleur]

Well said. I believe this will hold true until Intel rids itself of the FSB and moves to an on die memory controller.

And when quad core time comes around, I think Intel will have two choices.

1. Completely revamp the way cores talk to eachother (not using the FSB as they do now)
2. On die memory controller and crossbar switch.

Thats what CSI will fix, though it is already late and stated for 2007
Arstechnica article here

[nn_step]

Thats what CSI will fix, though it is already late and stated for 2007
Arstechnica article here

They cancelled that dude they needed the extra numbers on Conroe...

[Mikesta]

They cancelled that dude they needed the extra numbers on Conroe...

I thought that CSI was just delayed?

[nn_step]

I thought that CSI was just delayed?

I am just going by what I hear
http://www.theinquirer.net/?article=30710

[shadowing]

Everyone is all going on about how FX60 is much more expensive than Conroe. For all you know... the price can still decrease...

[viper650]

Everyone is all going on about how FX60 is much more expensive than Conroe. For all you know... the price can still decrease...

drop by 600 dollars? conroe is going to be better then most of amd for now. it will switch back and forth, but for now, amd fanboys need to realize they arent on top anymore. 07 will be a different story, and quite possibly different results. LET IT GO. i swear, reading posts by that fxx chris guy makes me want to stab myself in the eye.

[berk]

The truth is Conroe will spank until AM2 is revised in 2007.

[Mats]

Any price comparisation between todays A64 and tomorrows CD is just stupid. And yeah, we've seen the prices for both upcoming AM2 and CD, but remember that AM2 will show up first which makes them not comparable, besides any of the manufacturers can change the prices whenever they want. It's very possible that AMD will/must slash the prices in Q3, but telling us that now would be just plain suicide.