Conroe 2.4GHz benchmarked

[TrevorRC]

Now counting Windows. Which is several hundred megs on its own.

 

[Avalon]

23 minutes is a very good Spi32M run. The best I could do was ~26 minutes on my Opty 148 @ 3ghz, so this is good news. Still can't wait for the actual release and this chip being put to the test.

 

[lopri]

The data, while in its crude form, look to be in line with Anand's preview. Again very roughly Conroe seems 20% faster than A64, clock for clock with some quirks here and there (PI 1M, Sandra Multimedia, etc.). The situation is reversed now that AMD will have to crank up clocks to fight Intel's efficient architecture, which, by the way, I believe to be quite a possible scenario considering the recent clock speed jump of K8 CPUs.

 

[Alaa]

still no REAL benches..still not impressed! wuts the good thing in watching numbers?

 

[BrownTown]

im not surewhat you mean by 'real' benchmarks, most of these programs are designed to emulate a balanced work load of the area they are trying to test. They are certainly all valid benchmarks. I'm not really sure what you want him to do? These bencmarks are designed to be easy to compare across systems. When people and companies go to buy computers these are the numberse they look at. He can't test games becasue the 512 RAM would be a bottleneck, so all that reall leavs are things like WINRAR and DIVX that would take way to long to test since hed also have to test it on many other systems with the same files to give us a fair comparison. Pretty much hes done the majority of hte standard CPU benchmarks, and all the ones he can do easily,and this is on a CPU thats 3 months away how can you not be happy about that?

Anand already has the other stuf btw, so what more could you possibly want?

 

[Alaa]

i like anands' results they r so impressive bcuz another AMD system is compared ,some games were tested, i dont even luk at 3Dmarks or anything with numbers, i like frames per second--seconds..sumthing i can feel/see!!

 

[Keysplayr]

Originally posted by: Alaa
i like anands' results they r so impressive bcuz another AMD system is compared ,some games were tested, i dont even luk at 3Dmarks or anything with numbers, i like frames per second--seconds..sumthing i can feel/see!!

Try to count to 10 seconds without using numbers. LOL.
Look at the super Pi 32M score of 23 minutes. A few AMD users said they would have to clock their Opty's to somewhere in the neighborhood of 3.5 GHz to get there. Avalon I think said that he has his opty 148 clocked to 3GHz and can only manage 26 minutes in 32M tests. 26 minutes is fast, no doubt about it, but this 2.4 (and sort of crippled) Conroe is unbelievable. It's kind of like what BrownTown said. This is a role reversal. Now AMD has to have a 1GHz clock advantage to get closer to Core. Like the P4 had to have at 1GHz plus higher clock to even get near A64's. Different architectures, I know. But, well, LOOK!!!! hehe.

 

[Absolute0]

Yeah i've looked at the numbers. The Conroe at 2.4 Ghz is faster than my Opteron 170 at 3.15 Ghz. It's terrifying.

 

[FelixDeCat]

I still cant get over that 120,000 multimedia number! Amazing! Three times faster than the AMD/Pentium 820 number!

 

[Shenkoa]

Originally posted by: Cooler

Originally posted by: 996GT2
I know this was on page 5 of the "Intel's conroe in the hands..." thread, but I posted this just to make sure everyone can easily see it

http://www.xtremesystems.org/forums/showthread.php?t=95021&page=1

EDIT: FIxed link

These numbers are way to high to be realistic, I would not be suprised if Intel has created the Conroe with a special unit inside of it that makes it produce high numbers in benchmark application.

Imo their marketing team are scum.

 

[RaynorWolfcastle]

Originally posted by: Shenkoa

Originally posted by: Cooler

Originally posted by: 996GT2
I know this was on page 5 of the "Intel's conroe in the hands..." thread, but I posted this just to make sure everyone can easily see it

http://www.xtremesystems.org/forums/showthread.php?t=95021&page=1

EDIT: FIxed link

These numbers are way to high to be realistic, I would not be suprised if Intel has created the Conroe with a special unit inside of it that makes it produce high numbers in benchmark application.

Imo their marketing team are scum.

The really sad part, is that I believe that you might actually be serious

 

[lopri]

Originally posted by: keysplayr2003

Originally posted by: Alaa
i like anands' results they r so impressive bcuz another AMD system is compared ,some games were tested, i dont even luk at 3Dmarks or anything with numbers, i like frames per second--seconds..sumthing i can feel/see!!

Try to count to 10 seconds without using numbers. LOL.
Look at the super Pi 32M score of 23 minutes. A few AMD users said they would have to clock their Opty's to somewhere in the neighborhood of 3.5 GHz to get there. Avalon I think said that he has his opty 148 clocked to 3GHz and can only manage 26 minutes in 32M tests. 26 minutes is fast, no doubt about it, but this 2.4 (and sort of crippled) Conroe is unbelievable. It's kind of like what BrownTown said. This is a role reversal. Now AMD has to have a 1GHz clock advantage to get closer to Core. Like the P4 had to have at 1GHz plus higher clock to even get near A64's. Different architectures, I know. But, well, LOOK!!!! hehe.

26 mins PI 32M w/ Opteron 3GHz? That's impossible.

 

[Absolute0]

No it's not. I got 26 minutes at 3.15 Ghz, but with my 2gb set of ram with bad timings. with 3 Ghz and very good memory it shouldn't be too hard.

 

[dmens]

Originally posted by: Shenkoa
These numbers are way to high to be realistic, I would not be suprised if Intel has created the Conroe with a special unit inside of it that makes it produce high numbers in benchmark application.

Imo their marketing team are scum.

LOL, I'd like to see the reaction if anyone tried to pitch that idea to the designers. Real CPU circuitry that detects extremely specific codeflows in frontend then generates preordained results stored in.. microcode?!? WOAH!!!

Tinfoil hats, please.

 

[BrownTown]

yeah, clearly there is a microcode ROM of the first million values of PI, adn when it detects that SuperPi is being run it simply spits out the numbers! Intel has fooled us once again! Certainly 4MB of 5.8ns latentcy L2 cache has nothing to do with it?

 

[Keysplayr]

Originally posted by: Shenkoa


These numbers are way to high to be realistic, I would not be suprised if Intel has created the Conroe with a special unit inside of it that makes it produce high numbers in benchmark application.

Imo their marketing team are scum.

Fellow AT'ers, can I get a "Just Wow!"

 

[lopri]

Sorry I thought I was being sarcastic,

 

[allies]

This fall might be my first Intel build

 

[the Chase]

The thing that blew my mind in addition to the great perf. #'s he's getting with the slow ram, is that he is using a stock heat sink (2 heat pipes) with NO FAN and saying the heatsink is just SLIGHTLY WARM. WOW!! I realize he's not overolting the thing but still- Imagine if you slapped on an aftermarket heatpipe cooler with a slow speed fan!( or no fan at all). No need for watercooling or loud fans even when OC'ing with some extra volts!!

 

[Absolute0]

yeah and the scary thing is i'm gonna put my Conroe under my single stage lol, if the heat output is that low i'll be maintaining evap temps around -40c which is pretty good for my unit

 

[Cooler]

Sites still down they must be geting hits like crazy.

 

[designit]

Dont be so optimistice about overclocking this new "Core Architecture" (conroe I mean).
Because of Advance shared cache, wide Dynamic execution, and smart memory access Intel has implemented, these new chips are unclockable. Yes you will be able to overclock them, but nothing will be stable, and nothing will pass. Another word corrupted OS and system. Why do you think no one has been able to overclock conroe as we have seen a few tests already (by the exterme overclocking experts)?
Every time they tried to overclock it, system crashed. They blame it on BIOS, but dont count on new bios to fix it. IT CAN NOT OVERCLOCK as stated the reasons above.
Who is going to put a wager on my prediction?

 

[dmens]

Originally posted by: designit
Dont be so optimistice about overclocking this new "Core Architecture" (conroe I mean).
Because of Advance shared cache, wide Dynamic execution, and smart memory access Intel has implemented, these new chips are unclockable. Yes you will be able to overclock them, but nothing will be stable, and nothing will pass. Another word corrupted OS and system. Why do you think no one has been able to overclock conroe as we have seen a few tests already (by the exterme overclocking experts)?
Every time they tried to overclock it, system crashed. They blame it on BIOS, but dont count on new bios to fix it. IT CAN NOT OVERCLOCK as stated the reasons above.
Who is going to put a wager on my prediction?

Overclocking success depends on timing margins available on a given split and electrical/environmental conditions, architectural features have nothing to do with it.

 

[designit]

Originally posted by: dmens

Originally posted by: designit
Dont be so optimistice about overclocking this new "Core Architecture" (conroe I mean).
Because of Advance shared cache, wide Dynamic execution, and smart memory access Intel has implemented, these new chips are unclockable. Yes you will be able to overclock them, but nothing will be stable, and nothing will pass. Another word corrupted OS and system. Why do you think no one has been able to overclock conroe as we have seen a few tests already (by the exterme overclocking experts)?
Every time they tried to overclock it, system crashed. They blame it on BIOS, but dont count on new bios to fix it. IT CAN NOT OVERCLOCK as stated the reasons above.
Who is going to put a wager on my prediction?

Overclocking success depends on timing margins available on a given split and electrical/environmental conditions, architectural features have nothing to do with it.

Oh yes it does. Have you seen how the smart memory or Dynamic execution works?
if the cpu is overclocked, those 2 will go to an endless loop---->CRASH

 

[dmens]

Originally posted by: designit
Oh yes it does. Have you seen how the smart memory or Dynamic execution works?
if the cpu is overclocked, those 2 will go to an endless loop---->CRASH

Can you clarify the chip you are referring to?

Regardless, if an overclock results in a crash, then it is because there exists timing paths in the logic of that particular function, and one or more of those paths did not have enough timing margin to accomodate a given overclock. The fact that it was functional at a lower frequency means the function is validated and working.