AT Benches Core 2 X6800 (2.93 Ghz Conroe) vs FX-62

[Polish3d]

Originally posted by: Absolute0
3.5-4 Ghz on air is what it seems, so long as you have a high enough multi so that the FSB can get high enough. The cheapest conroes will top out around 3-3.2 Ghz due to FSB constraints.

Intels don't seem to correlate so strongly based on week/stepping, seems the higher ones just clock better. E6600 looks to be the mainstream choice for overclockers, looking into a 6700 for myself because the 10x multiplier makes 4+ Ghz possible.

I'm thinking 6600/6700 myself. A 3.5-4ghz conroe, if you assume average of 25% faster than AMD at same speed is like having a damned 4.5ghz Athlon X2, which would be smoking fast

 

[meksta]

i think the difference between am2 and conroe will be even less when AMD goes to 65nm. The frequency wall will probably be raised from 3ghz to about 4ghz...only question is will AMD put in more cache during that time.

AMD will definitely cut prices in the next month or two...just a matter of how much.

 

[dexvx]

Originally posted by: meksta
i think the difference between am2 and conroe will be even less when AMD goes to 65nm. The frequency wall will probably be raised from 3ghz to about 4ghz...only question is will AMD put in more cache during that time.

AMD will definitely cut prices in the next month or two...just a matter of how much.

I love how you armchair engineers think.

Its easy! Just shrink to 65nm, which will automatically give a 33% speed bump. While we're at it, lets double teh cache. Oh maybe, add a FP unit and maybe shrink the die to 50 mm^2 for more profits.

 

[meksta]

let me rephrase then:

1) AMD is in the process of going to 65nm on 300mm wafers (30% die shrink + 50% area increase)

2) Every process shrink has produced higher clock speeds. I said the "WALL" will probably be raised, and didn't say they will initially come out with 4ghz parts. In fact this might take a while.

3) Ok the cache part I was speculating, but never said it was easy. It's an option and it would not be out of the realm for them to do so. because they have cores with 2 different cache sizes already.


I believe prices will be competitive and that is great for the consumer. Not a fanboy of either...

 

[dexvx]

Originally posted by: meksta
1) AMD is in the process of going to 65nm on 300mm wafers (30% die shrink + 50% area increase)

2) Every process shrink has produced higher clock speeds. I said the "WALL" will probably be raised, and didn't say they will initially come out with 4ghz parts. In fact this might take a while.

Its inaccurate. 130nm wall was at 2.4-2.6Ghz (FX-57). The 90nm wall appears to be around 3Ghz. Just because you shrink doesnt mean you increase speed. Northwood -> Prescott yielded a whopping 400Mhz increase.

Originally posted by: meksta
3) Ok the cache part I was speculating, but never said it was easy. It's an option and it would not be out of the realm for them to do so. because they have cores with 2 different cache sizes already.

AMD's cache technology has always been behind Intel's. I highly doubt they will increase it. Their Z-Ram technology will probably enabled them to fit more L3 into it (as Quad Core seems to indicate).

 

[F1shF4t]

Originally posted by: n19htmare

Originally posted by: Dark CupcakeSo much money for so little gain

Then I guess you havn't upgraded to "AM2".... now thats so much money for so little gain.

going to E6600, you'd gain a lot more than if you were going to AM2. Plus i can guarantee the E6600 would over clock a hell of a lot better than the X2-3800 (both in $300 price range).

See the point is if i was buying a brand new system, sure it made sence to get the E6600, but i already have mine, and its now at same speed as conroe (sure u can overclock the conroe and it will be faster but thats not the point) So it would be more of a side grade.
Plus my 3800+ can hit 2.9ghz, i dont call that a bad overclock (run it at lower, dont wanna get crazy with the voltage)

At this point for anyone with a X2 system it makes no sence to upgrade to conroe and even more less sence to change to AM2.

I would rather buy myself a new screen or a new gfx card.

 

[Diasper]

I also think there needs to be some more real world testing at high graphical settings - as others have said the firingsquad results are interesting.

However, (as I understand it) if that's just due to being GPU limited as opposed to anything relating to FSB, then Conroe is an excellent and superior processor to the AMD64. I'm basing this on Anandtech's minimum FEAR FPS where Conroe scored ~50% more. Minimum FPS is king and I'm surprised more people haven't picked up on this

Also as Anand stated we need to see more tests especially with Oblivion.

The only other point is more information on the DRM implementation should be given - at least I have yet to come across any information clearly qualifying it.

 

[DerwenArtos12]

Originally posted by: Diasper
I think there needs to be some more real world testing at high graphical settings - as others have said the firingsquad results are interesting.

However, (as I understand it) if that's just due to being GPU limited as opposed to anything relating to FSB, then Conroe is an excellent and superior processor to the AMD64. I'm basing this on Anandtech's minimum FEAR FPS where Conroe scored ~50% more. Minimum FPS is king and I'm surprised more people haven't picked up on this

Also as Anand stated we need to see more tests especially with Oblivion.

The only other point is more information on the DRM implementation should be given - at least I have yet to come across any information clearly qualifying it.

The problem with such testing is that at high graphical settings the system is very GPU limited, not CPU limited, to get the best results for testing a CPU, use a benchmark that really stresses the CPU primarily of not soley. For the sake of gaming, that would be minimum/maximum framerates at ~600x800 with absolutely no AA or AF. Multitasking situations are really going to be a measure of this processor as well. AMD has the multi-core crown till Conroe is widely available, even in area's where intel has the crown, to this day, like media encoding.

 

[Polish3d]

Here's a great idea:


Run it at 800x600

and then run it at 1600x1200

 

[JEDIYoda]

I applaude all you AMD peeps who even with over whelming prood that Conroe is the next big Chip are stayiing true to AMD hoping for some sort of miracle....

 

[F1shF4t]

Originally posted by: JEDIYoda
I applaude all you AMD peeps who even with over whelming prood that Conroe is the next big Chip are stayiing true to AMD hoping for some sort of miracle....

Huh, whats the point of ur post, flame bait?, might i recomend u to visit the thread in my sig.

 

[Keysplayr]

Originally posted by: Dark Cupcake

Originally posted by: n19htmare

Originally posted by: Dark CupcakeSo much money for so little gain

Then I guess you havn't upgraded to "AM2".... now thats so much money for so little gain.

going to E6600, you'd gain a lot more than if you were going to AM2. Plus i can guarantee the E6600 would over clock a hell of a lot better than the X2-3800 (both in $300 price range).

See the point is if i was buying a brand new system, sure it made sence to get the E6600, but i already have mine, and its now at same speed as conroe (sure u can overclock the conroe and it will be faster but thats not the point) So it would be more of a side grade.
Plus my 3800+ can hit 2.9ghz, i dont call that a bad overclock (run it at lower, dont wanna get crazy with the voltage)

At this point for anyone with a X2 system it makes no sence to upgrade to conroe and even more less sence to change to AM2.

I would rather buy myself a new screen or a new gfx card.

@ bolded statement above: It is huh? So, your 3800X2 @2.9 is the same "megahurtz" as the tested 2.93 Conroe, yes. Just making this clear. You are at the same clock speed, but no where near the performance. You would have to run your 3800X2 somewhere around 3.5GHz or better to compete with the 2.93 GHz Conroe. And you and I both know, that just isn't going to happen. And, like you said, the Conroe can always be o/c'd to add more hurtin. We need to wait for 65nm X2's/FX's to see those clocks (maybe).

Cupcake, just call it like it is man.

 

[PetNorth]

Originally posted by: kknd1967
Average of all application performance benchmark. Only exclude the Sysmark communication benchmark which is not very related to the CPU performance as the author mentioned.

Conroe looks pretty good . Basically match the IDF benchmark where 2.67G Conroe is 20% faster. 2.93G Conroe is 26.25% faster (takes about 5% hit due to sublinear scaling of performance in freq).


Conroe / FX62

Business Winstone 1.175
Sysmark office 1.26
Sysmark Doc 1.307142857
Sysmark Data 1.513513514
Multimedia Creation 1.057
Sysmark 1.273458445
Sysmark_3D 1.274052478
Sysmark_2D 1.287946429
Sysmark_web 1.256637168
Q4 1.238029146
FEAR_avg 1.178294574
FEAR_min 1.509433962
FEAR_max 1.141975309
BF2 1.202803738
------------------------------------------
avg speed over A64 1.262520544
IPC over A64 1.206504274

Only to clarify.

Why do you repeat results? I mean, for example, Sysmark_3D, Sysmark_2D, Sysmark_web results ARE INCLUDED in Internet Content Creation Overall. The same thing with sysmark doc, data etc are included in Office Overall. So the correct resume is:

Business Winstone 17,56% faster
MCC Winstone 5,76% faster
Sysmark Office Overall 26,67% faster
Sysmark ICC Overall 27,35% faster

So, overall in these tests X6800 2.93 is 19,33% faster than FX-62 2.8

The same thing with FEAR, if you have the average FEAR fps, why do you include max and min, if they are included in average fps?

So, in these games benches would be:

Quake 4 23,80% faster
FEAR (average fps) 17,83% faster
Battlefield 20,28% faster

Overall with these three games X6800 2.93 is 20,64% faster than FX-62 2.8

Of course, talking about games, benches without AA/AF, are useless for real life.
In real life, with AA/AF etc activated, differences are basically 0.

 

[Keysplayr]

Originally posted by: PetNorth

Of course, talking about games, benches without AA/AF, are useless for real life.
In real life, with AA/AF etc activated, differences are bassically 0.

Where does it show the difference being 0 when AA/AF are enabled?
I'm not saying you're wrong, but would like to see where you got your data from.

 

[PetNorth]

firingsquad have some result with Quake; I guess Toms with Far Cry

 

[F1shF4t]

Originally posted by: keysplayr2003

Originally posted by: Dark Cupcake

Originally posted by: n19htmare

Originally posted by: Dark CupcakeSo much money for so little gain

Then I guess you havn't upgraded to "AM2".... now thats so much money for so little gain.

going to E6600, you'd gain a lot more than if you were going to AM2. Plus i can guarantee the E6600 would over clock a hell of a lot better than the X2-3800 (both in $300 price range).

See the point is if i was buying a brand new system, sure it made sence to get the E6600, but i already have mine, and its now at same speed as conroe (sure u can overclock the conroe and it will be faster but thats not the point) So it would be more of a side grade.
Plus my 3800+ can hit 2.9ghz, i dont call that a bad overclock (run it at lower, dont wanna get crazy with the voltage)

At this point for anyone with a X2 system it makes no sence to upgrade to conroe and even more less sence to change to AM2.

I would rather buy myself a new screen or a new gfx card.

@ bolded statement above: It is huh? So, your 3800X2 @2.9 is the same "megahurtz" as the tested 2.93 Conroe, yes. Just making this clear. You are at the same clock speed, but no where near the performance. You would have to run your 3800X2 somewhere around 3.5GHz or better to compete with the 2.93 GHz Conroe. And you and I both know, that just isn't going to happen. And, like you said, the Conroe can always be o/c'd to add more hurtin. We need to wait for 65nm X2's/FX's to see those clocks (maybe).

Cupcake, just call it like it is man.

Maybe u should read my post properly, i will never buy any fx or extreme expensive edition.

I was talking about the e6600 which is a 2.4ghz conroe, and i said that my 2.7ghz X2 will perform roughly the same. Now u can overclock the conroe, duh i even mentioned that, but i desided to not take it into account.

Why? cause milage can warry, and i can either get a dud with same performance ie 2.4ghz, or it can overclock greatly with anywhere between 0% and 30% advantage.
If u can justify me spending 600 bucks now for that gamble, and a gamble for no purpose as i'm quite happy with my current rig, in fact i'm typing this on my 3200+ rig.

I am calling like it is. 2.7ghz x2 vs 2.4ghz conroe, not hard to belive performing the same if u look at AT's benchmarks.

 

[n7]

Very impressive if you play games at 1024x768

If you play games @ 1600x1200, there's no reason for me to upgrade to Core 2 Duo, since at that resolution, i'm so GPU limited there's barely any benefit with C2D.

I will likely upgrade later this fall anyway though, since encoding times etc. will be loads better with C2D.

 

[Keysplayr]

Originally posted by: Dark Cupcake

Originally posted by: keysplayr2003

Originally posted by: Dark Cupcake

Originally posted by: n19htmare

Originally posted by: Dark CupcakeSo much money for so little gain

Then I guess you havn't upgraded to "AM2".... now thats so much money for so little gain.

going to E6600, you'd gain a lot more than if you were going to AM2. Plus i can guarantee the E6600 would over clock a hell of a lot better than the X2-3800 (both in $300 price range).

See the point is if i was buying a brand new system, sure it made sence to get the E6600, but i already have mine, and its now at same speed as conroe (sure u can overclock the conroe and it will be faster but thats not the point) So it would be more of a side grade.
Plus my 3800+ can hit 2.9ghz, i dont call that a bad overclock (run it at lower, dont wanna get crazy with the voltage)

At this point for anyone with a X2 system it makes no sence to upgrade to conroe and even more less sence to change to AM2.

I would rather buy myself a new screen or a new gfx card.

@ bolded statement above: It is huh? So, your 3800X2 @2.9 is the same "megahurtz" as the tested 2.93 Conroe, yes. Just making this clear. You are at the same clock speed, but no where near the performance. You would have to run your 3800X2 somewhere around 3.5GHz or better to compete with the 2.93 GHz Conroe. And you and I both know, that just isn't going to happen. And, like you said, the Conroe can always be o/c'd to add more hurtin. We need to wait for 65nm X2's/FX's to see those clocks (maybe).

Cupcake, just call it like it is man.

Maybe u should read my post properly, i will never buy any fx or extreme expensive edition.

I was talking about the e6600 which is a 2.4ghz conroe, and i said that my 2.7ghz X2 will perform roughly the same. Now u can overclock the conroe, duh i even mentioned that, but i desided to not take it into account.

Why? cause milage can warry, and i can either get a dud with same performance ie 2.4ghz, or it can overclock greatly with anywhere between 0% and 30% advantage.
If u can justify me spending 600 bucks now for that gamble, and a gamble for no purpose as i'm quite happy with my current rig, in fact i'm typing this on my 3200+ rig.

I am calling like it is. 2.7ghz x2 vs 2.4ghz conroe, not hard to belive performing the same if u look at AT's benchmarks.

I believe I read your post just fine, thanks. But I did assume that you were talking about the 2.93 Conroe. My bad.

 

[Munky]

Originally posted by: keysplayr2003

Originally posted by: PetNorth

Of course, talking about games, benches without AA/AF, are useless for real life.
In real life, with AA/AF etc activated, differences are bassically 0.

Where does it show the difference being 0 when AA/AF are enabled?
I'm not saying you're wrong, but would like to see where you got your data from.

It may not be 0, and at 1024 res it will most likely not be 0, but it will still definitely be a smaller difference. AA/AF are completely gpu-bound functions.

 

[Keysplayr]

No offense Munky, but I've really grown tired of hearing the words "may" or "will" or "should". I am more looking forward to the "is" word. I know there is not that much available info yet. So I guess I will wait to see just like everyone else.

:beer:

 

[z3R0C00L]

Originally posted by: mamisano
These are all Intel supplied systems that are being tested here, with OSes, etc installed by Intel.

I will wait until I see an IN-HOUSE comparrison from a reputible site before I pass judgement.

Intel marketing at work once again...

Uh?!

You cannot do anything to an OS to give a CPU such a large boost. Conroe is for real. I think it's about time AMD fanboys swallow there pride and admit to it. If it were an AMD CPU showing the same early results you'd all be jumping on that like Jessica Simpson at a college frat party.

I currently run an AMD system, but I know (from looking at the specs etc) that Conroe will end up EVEN FASTER within the next year. You all forget we don't have SSE4 apps and games yet... much less heavy SSE3 useage.. all of which Conroe will compute with extreme relative ease.

Conroe is real. AMD better be and obviously are afraid. They'll strike back.. but don't expect a miracle.

 

[XNice]

no one is excepting any miracles, anyone in the amd camp is merely expecting their side to show their stuffs to. People are so quick nowadays to call someone a fanboy. I own an amd based system and like many other smarter people here, see no need to upgrade anytime soon and if it did, a gpu would be a better purchase than a cpu/mobo combo.

I for one do not like intel cpu due to the available motherboard chipsets. If they don't improve to the standard that NF4 series has set as far as feature set, reliability, and cheaper OC'ing, I'm not switching to any intel based systems yet, especially with availablity estimated to be scarce at best(~%25 by Q3????).

Everyone talks about the cpu's and all, but a big part of the picture is the new chipsets running these cpu's. So really we only have one third of the picture, the other 2 thirds(amd, mobo chipsets) are yet to be officially released paper or otherwise.

 

[XNice]

Originally posted by: z3R0C00L
You cannot do anything to an OS to give a CPU such a large boost. Conroe is for real. I think it's about time AMD fanboys swallow there pride and admit to it. If it were an AMD CPU showing the same early results you'd all be jumping on that like Jessica Simpson at a college frat party... don't expect a miracle.

Actually OS configuration plays a big part in performance, just look at the display adapter settings.Sorry but, this argument actually holds merit.

 

[kknd1967]

There are many ways to do average.
for example you say avg fps is important. I say both avg, min and max are important. others may feel only min fps is important.

My starting point is that if you treat each sample as an outcome, the more you take average the better, even some are correlated but they are different scenarios. Some benchmark's overall score is not really related to CPU as they factor in dirve i/o etc. For example the total 3DMark05 overall score means less about pure CPU performance as compared to 3DMark05 cpu score.

Originally posted by: PetNorth

Originally posted by: kknd1967
Average of all application performance benchmark. Only exclude the Sysmark communication benchmark which is not very related to the CPU performance as the author mentioned.

Conroe looks pretty good . Basically match the IDF benchmark where 2.67G Conroe is 20% faster. 2.93G Conroe is 26.25% faster (takes about 5% hit due to sublinear scaling of performance in freq).


Conroe / FX62

Business Winstone 1.175
Sysmark office 1.26
Sysmark Doc 1.307142857
Sysmark Data 1.513513514
Multimedia Creation 1.057
Sysmark 1.273458445
Sysmark_3D 1.274052478
Sysmark_2D 1.287946429
Sysmark_web 1.256637168
Q4 1.238029146
FEAR_avg 1.178294574
FEAR_min 1.509433962
FEAR_max 1.141975309
BF2 1.202803738
------------------------------------------
avg speed over A64 1.262520544
IPC over A64 1.206504274

Only to clarify.

Why do you repeat results? I mean, for example, Sysmark_3D, Sysmark_2D, Sysmark_web results ARE INCLUDED in Internet Content Creation Overall. The same thing with sysmark doc, data etc are included in Office Overall. So the correct resume is:

Business Winstone 17,56% faster
MCC Winstone 5,76% faster
Sysmark Office Overall 26,67% faster
Sysmark ICC Overall 27,35% faster

So, overall in these tests X6800 2.93 is 19,33% faster than FX-62 2.8

The same thing with FEAR, if you have the average FEAR fps, why do you include max and min, if they are included in average fps?

So, in these games benches would be:

Quake 4 23,80% faster
FEAR (average fps) 17,83% faster
Battlefield 20,28% faster

Overall with these three games X6800 2.93 is 20,64% faster than FX-62 2.8

Of course, talking about games, benches without AA/AF, are useless for real life.
In real life, with AA/AF etc activated, differences are basically 0.

 

[lopri]

You guys don't think faster GPUs will come during the early lifespan of Conroe? What about SLI? And since when did we benchmark CPUs @1600x1200 w/ 4xAA/8xAF? Hell, at that configuration P4 and A64 will probably produce similar numbers. CPU benches are done on low res for reason. And the faster the GPU, the wider the gap will be.