Conroe benchmarks....w000000t!

[PeterJensen]

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

(nana2) (nana2) (nana2) (nana2)

http://vic.expreview.com/attachment/sis+media.jpg

[Toss3]

linky no worky

[slaWter]

Conroe is on top of my list this summer! I hope there are some good SLI mobos for overclocking!

[a12ctic]

wow look at that voltage thats like twice what my amd chip uses :P

[Superfly]

wow - maybe I'll be able to afford one next year. :-(

[retsam]

wow, there server is dead..damn ..

[MUYA]

wow look at that voltage thats like twice what my amd chip uses :P
CPU-Z must be reporting it wrong. Everest reports it as 1.18v ..

[Telomerase]

wtf@SSE4 or am i just out-of-the-loop?
how different is conroe from dothan? is the core still very similar, i know it has SSE3/4, EM64T, die shrink to 65nm, etc, but what about the key performance issues. such as 14-stage pipeline that a dothan has. i guess the new macro-op execution and 128-bit wide bandwidth for SSE instructions make conroe a bit faster than dothan. i'm still going to be happy with my 2.26Ghz dothan, though.

[Daneel Olivaw]

I don't trust that integer score...

[MUYA]

Conroe has 4MB l2 cache? I thought it was 2MB for each core?

[Daneel Olivaw]

Conroe has 4MB l2 cache? I thought it was 2MB for each core?

Maybe since its a shared 4MB cpuz sees it as 4MB each...?

[fellix]

I don't trust that integer score...
For 1 (one) cycle, Conroe is able to perform the following SIMD op's:

128-bit PADD;
128-bit PMULL;
128-bit LOAD;
128-bit STORE;
macro-fused comp&jump condition;


...

not counting the new fancy memory load/store "disambiguation" which can save a lot of precoius wait cycles - useful for FP code mostly.

[Subtestube]

I'm really curious to know what the cache size effect on this CPU is... how're those 2.16GHz 1MB/core ones that are considerably cheaper (dailytech posted pricing and revised pricing recently) going to perform? Hmmm.

[mike686]

That test also shows the X2 4400+ getting beaten by a good margin by a Pentium 4 Xeon...and we all know how great P4's are for gaming. ;) It must be a very Intel-biased test. Sill probably a little faster than the X2 in games though.

[hemmy]

I think the 4MB of L2 is shared btwn the 2 cores, (dont quote me) instead of having 2mb per core

Diff from dothan: dual core, 65nm, more instructions, 64bit

[hemmy]

Intel had a press release on their site about 3 years ago saying they would have 1nm 20GHZ cpus with 1 BILLION transistors by 2008. Guess that didn't turn out true

I wrote a paper on this article, but i am sure it no longer exists on their site

[Richteralan]

I take these synthetic pre-release benchmarks as a truck of salt.

I'll wait and see.

[hemmy]

I take these synthetic pre-release benchmarks as a truck of salt.

I'll wait and see.

These + all the other websites that showed a 2.66GHZ conroe beating a 2.8 FX60 isnt proof enough it is an awesome chip?

[Richteralan]

These + all the other websites that showed a 2.66GHZ conroe beating a 2.8 FX60 isnt proof enough it is an awesome chip?

I would like to see more real-world testing plus non-Intel benchmarks.

I wouldn't deny that it LOOKS awesome from graphs.

[Subtestube]

REALLY good [technical] article on Core (Conroe/Merom) architecture on Ars technica:

http://arstechnica.com/articles/paedia/cpu/core.ars

[jAkUp]

Yup, the L2 cache on Conroe are different, it is shared across the CPU's, so you actually get 4MB's out of it. Not like X2's or P4D's.

[Superfly]

REALLY good [technical] article on Core (Conroe/Merom) architecture on Ars technica:

http://arstechnica.com/articles/paedia/cpu/core.ars

Great article, thanks.

You can almost see were the huge boost in perfomance comes from now, a much cleaner architecture and getting rid of netbursts SSE bottleneck is going to make things really interesting.

[watever]

from what i heard.. its 2mb cache for each core. but if only one core is used for a program, then that one core uses both caches, giving it 4mb.

[Ninja Prime]

Ehh that's not nearly as impressive as all the hype I heard. I'll wait and see, I'm not expecting better than a 15% improvement as far as gaming goes over the current top of the line. We'll see when the next AMD proc shows up.

[SH64]

WOW! now thats a CPU for Oblivion!