Which Xeon is used in the 6-core system?

hajime

macrumors 603
Original poster
Jul 23, 2007
5,807
807
Hello, anybody knows which Xeon is used in the 6-core system? On the product website, it only mentions the 12-core and Quad-core systems. I want to know the cache size and its clock speed under turbo boost.
 

Inconsequential

macrumors 68000
Sep 12, 2007
1,977
1
Thanks. It is relatively outdated compared with the E5-2687w.
Stil performs extremely well, the new E5 chips are between 1 and 2 times faster, depending on what your doing. But the average performance difference is about 15%.
 

steveOooo

macrumors 6502a
Jun 30, 2008
737
87
UK
Stil performs extremely well, the new E5 chips are between 1 and 2 times faster, depending on what your doing. But the average performance difference is about 15%.
So the mp2013 will be a whopping 30% faster ! Well duh apple - u did miss out on a processor generation!
 

hajime

macrumors 603
Original poster
Jul 23, 2007
5,807
807
Stil performs extremely well, the new E5 chips are between 1 and 2 times faster, depending on what your doing. But the average performance difference is about 15%.
If I forget about the MacPro and get a HP Z820, the configurations would be like: E5-2687w (8 cores, 3.1GHz, 3.8GHz under TurboBoost) or E5-2643 (4-core, 3.3HGz), Nvidia Quadro 5000. Will that be just 15% faster than the 6-core MacPro?
 
Last edited:

El Awesome

macrumors 6502
Jul 21, 2012
471
0
Zurich
Is there any possibility to get this E5-2687w working in a MP4.1?
If Sandy Birdges are working in all other Macs, there must be somehow a possibility to include that in a MP firmware, or not?
 

Umbongo

macrumors 601
Sep 14, 2006
4,931
54
England
Is there any possibility to get this E5-2687w working in a MP4.1?
If Sandy Birdges are working in all other Macs, there must be somehow a possibility to include that in a MP firmware, or not?
No they use a diffrent socket and chipset.
 

thekev

macrumors 604
Aug 5, 2010
6,721
1,820
Is there any possibility to get this E5-2687w working in a MP4.1?
If Sandy Birdges are working in all other Macs, there must be somehow a possibility to include that in a MP firmware, or not?
There is 0 chance of this. It's not a firmware thing. These use a completely different socket and chipset. Westmere and Nehalem were compatible. Anything after that is 100% non compliant.

Stil performs extremely well, the new E5 chips are between 1 and 2 times faster, depending on what your doing. But the average performance difference is about 15%.
Even if they perform remarkably well today, I have to wonder how long term support will look given that it has retained a large number of the same components since 2009. Once the 2013 model hits, will this one receive sub-par support?
 

El Awesome

macrumors 6502
Jul 21, 2012
471
0
Zurich
There is 0 chance of this. It's not a firmware thing. These use a completely different socket and chipset. Westmere and Nehalem were compatible. Anything after that is 100% non compliant.
*facepalm*
Thanks, sometimes I'm too stupid to live.
 

thekev

macrumors 604
Aug 5, 2010
6,721
1,820
*facepalm*
Thanks, sometimes I'm too stupid to live.
I'm really not sure about the intended tone of that response :p. Anyway I probably could have worded it differently, but I'm 100% certain on this one. The new ones use LGA2011.
 

El Awesome

macrumors 6502
Jul 21, 2012
471
0
Zurich
No, you're right. Unless there is a chance to get the CPU board from the 2013 MPs running in the older ones, you can forget it.

I didn't even check the socket type before I asked. That's what I just facepalmed myself. (You know, I'm not awesome all the time :p )
 

derbothaus

macrumors 601
Jul 17, 2010
4,060
4
Stil performs extremely well, the new E5 chips are between 1 and 2 times faster, depending on what your doing. But the average performance difference is about 15%.
1-2 times faster? Am I missing something? Wouldn't that be 100-200% faster! Or is that 1x = nothing/ same, 2x = 100% faster? All are wrong except the 15% faster average.
 

frabber

macrumors regular
Nov 28, 2008
119
8
2.5 years old, sitting on 300 billion dollar and still not be able to hire enough personell to upgrade these other than iphone/tablet/macbook computers .
as if everything has to be blessed by john ive first.
i tell you apple is severely bottlenecked in the wrong way
kill it, or upgrade it, but don't leave it half-dead
 

Inconsequential

macrumors 68000
Sep 12, 2007
1,977
1
1-2 times faster? Am I missing something? Wouldn't that be 100-200% faster! Or is that 1x = nothing/ same, 2x = 100% faster? All are wrong except the 15% faster average.
When I was looking at the benchmarks there are a few specialist cases, for example, something like AES encryption/decryption performance which is MUCH faster than the W3680 chips due to dedicated circuitry.

Example: http://www.tomshardware.co.uk/core-i7-3930k-3820-test-benchmark,review-32336-6.html - Graph 4.

In cases where pure clock speed counts and the app uses areas of the chip which haven't been optimised, the E5s have a small advantage.

Go and have a look at the benchmarks yourself.

More here: http://www.anandtech.com/bench/Product/142?vs=552

Even my mates mobile Sandy Bridge 2Ghz chip slams my W3680 in AES benchmarks.
 
Last edited:

derbothaus

macrumors 601
Jul 17, 2010
4,060
4
^^^ Thanks for the clarification. I was initially like. No freakin way. Then I was like "OK, specialty apps". Now your post makes sense.
 

deconstruct60

macrumors G3
Mar 10, 2009
8,490
1,474
When I was looking at the benchmarks there are a few specialist cases, for example, something like AES encryption/decryption performance which is MUCH faster than the W3680 chips due to dedicated circuitry.

Example: http://www.tomshardware.co.uk/core-i7-3930k-3820-test-benchmark,review-32336-6.html - Graph 4.
And directly under graph 4 is the text ....

" ... In my Sandy Bridge-E review, we figured out that the AES256 throughput of Intel’s AES-NI-equipped CPUs is tied directly to memory bandwidth. With only four cores mated to a quad-channel DDR3-1600 subsystem, the -3820 rises right to the top of our Cryptography results. ..."

The W3680 has the dedicated circuitry.

"AES New Instructions YES"
http://ark.intel.com/products/47917/Intel-Xeon-Processor-W3680-(12M-Cache-3_33-GHz-6_40-GTs-Intel-QPI)

The fact that E5's have 4 memory controllers , better/faster internal ring bus, and higher RAM speeds are the primary differentiating factors. The instructions work so fast that if you have more cores lit up with them you can outstrip the memory bandwidth.

It is not too hard to compose AES circuitry to en/decrypts as fast as you can pass bytes through the encoder. Basically "live streaming". So it boils down to how fast can stream bytes from memory.
 

Inconsequential

macrumors 68000
Sep 12, 2007
1,977
1
And directly under graph 4 is the text ....

" ... In my Sandy Bridge-E review, we figured out that the AES256 throughput of Intel’s AES-NI-equipped CPUs is tied directly to memory bandwidth. With only four cores mated to a quad-channel DDR3-1600 subsystem, the -3820 rises right to the top of our Cryptography results. ..."

The W3680 has the dedicated circuitry.

"AES New Instructions YES"
http://ark.intel.com/products/47917/Intel-Xeon-Processor-W3680-(12M-Cache-3_33-GHz-6_40-GTs-Intel-QPI)

The fact that E5's have 4 memory controllers , better/faster internal ring bus, and higher RAM speeds are the primary differentiating factors. The instructions work so fast that if you have more cores lit up with them you can outstrip the memory bandwidth.

It is not too hard to compose AES circuitry to en/decrypts as fast as you can pass bytes through the encoder. Basically "live streaming". So it boils down to how fast can stream bytes from memory.

Ok, it is still twice as fast, proving my original comments as there is nothing you can do with the W3680 in a Mac Pro to make up the difference.