Intel NDA Desktop Platform Roadmap WW03 2012

[knopflerbruce]

So, how does this compare to sandy bridge for example?

[Christian Ney]

Okay ChipHell removed pics, I managed to recover the pics and upload them again in my gallery. Enjoy:

http://hwbot.org/forum/showthread.php?p=142060&postcount=22

Edited December 13, 2011 by Christian Ney

[Christian Ney]

So, how does this compare to sandy bridge for example?

 

[TaPaKaH]

so it's 1841 degrees hotter in 06 than 2700K, very informative

[Christian Ney]

Oops ,

 

I took the diagram of my latest review and changed the values, missed the x line info

 

EDIT: Fixed

Edited December 13, 2011 by Christian Ney

[knopflerbruce]

So basically it's just an IGP upgrade. Unless it does 5.5 on air, no reason to upgrade for a 24/7 user:D I wonder if it's got the same walls as SB/SB-E.

[flanker]

nice Christian...If I remember right, 3770K has 3.5/3.9 GHz?

[Christian Ney]

First post snipped...

 

Come on Intel, you are not funny

 

2 weeks late though

[Christian Ney]

Blablablabla....

Please note that this information was obtained illegally.

Blablablabla...[/Quote]

 

Prove it!

[sin0822]

sad they snipped it

[Christian Ney]

Over the next few months, Intel will be releasing several server and workstation class processors based on Sandy Bridge. Entry level workstation offerings include E3-1200 v2 and E5-1600, both reliable alternatives to regular desktop offerings. Dual socket processors include the E5-2400 series and the higher performing E5-2600. Finally, we have the quad socket E5-4600 series.

 

We have now seen some initial performance comparisons between Sandy Bridge-EP and Westmere-EP. A document leaked by a Chinese website (PDF file) revealed what sort of performance increase we can expect in Sandy Bridge-EP over Westmere-EP. Various benchmarks were run using an X5690 (with 6 cores/12 threads at 3.46 GHz base clock, 3.6GHz Turbo Clock) and an E5-2690 (8 cores/16 threads at 2.9 GHz), in a dual-socket setup with comparable configuration, except that the X5690 had 40 GB of DDR3-1333 in triple channel configuration and the E5-2680 had 32GB DDR3-1600 in quad channel configuration.

 

Giving the X5690 a baseline score of 1 in each test, we see some interesting results. Both chips have a similar TDP. The X5690 has a core speed 24% higher than E5-2690, but the latter has 33% more cores. Using the synthetic Linpack benchmark to measure matrix multiplication, E5-2690 scored a respectable 2.2. Using OLTP Database (TPC-C Oracle), E5-2690 scores 1.5, while using a middle tier java test (SPECjbb 2005) the score is 1.56. Integer throughput using SPECint_base2006 gives a score of 1.69. Moving to some technical computing tasks, Floating Point Throughput (SPECfp*_rate_base2006) gave a score of 1.82 and Memory Bandwidth (STREAM_MP Triad) scored 1.88. The average performance over these 6 applications is approx 1.78, indicating a projected performance increase of approximately 80% when moving from the high-end Westmere-EP offering to the equivalent Sandy Bridge-EP processor.

 

Comparing the Xeon 5600, E5-2400 and E5-2600 we can see the benefits of the newer systems, and why the E5-2600 series is so powerful.

 

 

 

 

 

Source: CPUWorld

 

Gallery: http://imup.se/g/BVUoc1

 

pdf available: https://docs.google.com/uc?export=download&id=0B0BeZFb-M9wCZWJjMGMwMjItNDNlZi00MGI1LWJiNDctZTliOTEwMDhhNGE0

 

Download all pics : .rar file

Edited December 23, 2011 by Christian Ney

[Christian Ney]

[|ron]

Uhm.. Chris, cpuz seems edited in the cache area, you see that bright spot around the number 10?

Anyway, probably a 3dm11 monster at a friendly price of 1999USD :lol:

[Christian Ney]

Yeah saw that