Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Where does it say they are ahead of the schedule? LGA 1155 parts were already announced on 3rd of January and quad core models should be available already (or late this week/early next week).

Sandy Bridge-E (i.e. LGA 2011 parts) are still Q4 2011 according to AnandTech
 
since they have been hacked and shown to work with os x. The article implies a need for apple to speed up a bit maybe 3rd q instead of 4th q. I am reading into it a bit.
 
since they have been hacked and shown to work with os x. The article implies a need for apple to speed up a bit maybe 3rd q instead of 4th q. I am reading into it a bit.

What does that show? New hardware has always been hacked pretty soon to work with OS X but it doesn't affect Apple's update cycles. It's not Apple who is slowing down here. If Intel can't release SB-E before Q4, then there won't be a SB Mac Pro until Q4 unless Apple gets early access.

I don't know how you are making conclusions that because OS X has been installed in SB PC, Apple should release the Mac Pro earlier.
 
Wishful thinking, I'd say. We all know how Apple treats Mac Pro users - with a year and a half between each new rig :(
 
at the rate we're going with mp updates, i wouldnt be surprised if maybe 2011 or 2012 gets skipped
 
at the rate we're going with mp updates, i wouldnt be surprised if maybe 2011 or 2012 gets skipped
As it's been mentioned, they can't use chips that aren't available.

Now if they decide to use the LGA1355 parts, we might see a new MP at the tail end of this year (they'd have to put sufficient human resources into it to make this year, and assuming Intel's not late in shipping parts). But it's likely that the MP won't see an update until 2012 as the consumer market is making them a lot more money, so those projects will get development resource priority.
 
As it's been mentioned, they can't use chips that aren't available.

Now if they decide to use the LGA1355 parts, we might see a new MP at the tail end of this year (they'd have to put sufficient human resources into it to make this year, and assuming Intel's not late in shipping parts). But it's likely that the MP won't see an update until 2012 as the consumer market is making them a lot more money, so those projects will get development resource priority.

What LGA 1355 parts? AFAIK that socket doesn't even exist, it's just LGA 1155 for mainstream and LGA 2011 for high-end. Only couple of ancient PC-Watch articles mentioned it. Its Wikipedia article has been removed as well.
 
What LGA 1355 parts? AFAIK that socket doesn't even exist, it's just LGA 1155 for mainstream and LGA 2011 for high-end. Only couple of ancient PC-Watch articles mentioned it. Its Wikipedia article has been removed as well.
Last I checked, the Sandy Bridge EN parts where LGA1356 (not 1355, sorry :eek:), not LGA2011 (EP versions for sure). It's no longer listed in the wiki page however (32nm section), so that information was either wrong, or Intel may have requested it be removed. :confused: There's no socket information for the EN versions (to be marketed as Xeon E5 it seems; Xeon E7 for the EP, and a single processor server chip as the Xeon E3), and the different pics have different information (one lists a 4x channel memory controller, another, Server section, still lists it as 3x channel (which is what it was listed at previously).

It was nice, clean, and easy to understand. Now, not so much.

At any rate, here's a source (.pdf file) of the socket data (LGA 2011, LGA 1356, and LGA 1155). The reduced pin count is due to 1x QPI lane (vs. 2x), PCIe lanes (24 instead of 40), and one less memory channel (3x vs. 4x in the EP's).
 
Last I checked, the Sandy Bridge EN parts where LGA1356 (not 1355, sorry :eek:), not LGA2011 (EP versions for sure). It's no longer listed in the wiki page however (32nm section), so that information was either wrong, or Intel may have requested it be removed. :confused: There's no socket information for the EN versions (to be marketed as Xeon E5 it seems; Xeon E7 for the EP, and a single processor server chip as the Xeon E3), and the different pics have different information (one lists a 4x channel memory controller, another, Server section, still lists it as 3x channel (which is what it was listed at previously).

It was nice, clean, and easy to understand. Now, not so much.

At any rate, here's a source (.pdf file) of the socket data (LGA 2011, LGA 1356, and LGA 1155). The reduced pin count is due to 1x QPI lane (vs. 2x), PCIe lanes (24 instead of 40), and one less memory channel (3x vs. 4x in the EP's).

I guess we'll have to wait and see as nobody seems to have an idea what the EN parts will use :p IMO it wouldn't make much sense to use LGA 1366 and 2011, especially because old LGA 1366 can't be used due to different chipset (X68 vs X58, or can they?). It's already confusing enough with two sockets, let alone what it would be with three :eek: Anand is saying that it will be just LGA 2011 (no mention of 1366) though that doesn't guarantee it will be so. I can try to ask him after CES when he should return my email (applied and was accepted to write for AT :))
 
IMO it wouldn't make much sense to use LGA 1366 and 2011, especially because old LGA 1366 can't be used due to different chipset (X68 vs X58, or can they?).

Wouldn't be the first time... The Pentium 4, Core Duo, and Core 2 Duo all used the same sockets, but different chipsets.
 
I guess we'll have to wait and see as nobody seems to have an idea what the EN parts will use :p IMO it wouldn't make much sense to use LGA 1366 and 2011, especially because old LGA 1366 can't be used due to different chipset (X68 vs X58, or can they?). It's already confusing enough with two sockets, let alone what it would be with three :eek: Anand is saying that it will be just LGA 2011 (no mention of 1366) though that doesn't guarantee it will be so. I can try to ask him after CES when he should return my email (applied and was accepted to write for AT :))
LGA1366 = EOL (there's a few parts that are still due out, but they're not new architecture).

Wouldn't be the first time... The Pentium 4, Core Duo, and Core 2 Duo all used the same sockets, but different chipsets.
Exactly. ;)

But in this case, it does make sense IMO;
Consumer = LGA1155
Entry/mid-level server = LGA1356
Highend/MP server = LGA2011

For the Xeons, the QPI channels, PCIe lane counts, and memory controller configurations are different, which would be better suited to different chipsets.

Just a thought anyway. :)
 
LGA1366 = EOL (there's a few parts that are still due out, but they're not new architecture).

My turn to screw up with the numbers :D That explains why I couldn't find the Wikipedia article, I had wrong numbers, the article for LGA 1356 still exists.

But in this case, it does make sense IMO;
Consumer = LGA1155
Entry/mid-level server = LGA1356
Highend/MP server = LGA2011

Well, it makes some sense but I don't think Intel has done that in the recent past at least (LGA 771 was for DPs but LGA 775 was for UP and mainstream CPUs). Why couldn't they just use LGA 2011 for everything? Another thing that sounds weird is that why does basically the same CPU with just two QPIs require 2011 pins while the same CPU with single QPI would only need 1356?

I'm still saying there is only one or two old PC-Watch articles supporting it, no other sources ;)
 
My turn to screw up with the numbers :D That explains why I couldn't find the Wikipedia article, I had wrong numbers, the article for LGA 1356 still exists.



Well, it makes some sense but I don't think Intel has done that in the recent past at least (LGA 771 was for DPs but LGA 775 was for UP and mainstream CPUs). Why couldn't they just use LGA 2011 for everything? Another thing that sounds weird is that why does basically the same CPU with just two QPIs require 2011 pins while the same CPU with single QPI would only need 1356?

I'm still saying there is only one or two old PC-Watch articles supporting it, no other sources ;)
The information available is a bit confusing, and it's reflected on Wikipedia it seems.

If I had to guess, it's due to the interest right now is in the consumer parts that have just been announced (they've had more press than the enterprise parts from day one, which isn't uncommon).

So I don't expect we'll see much on the enterprise parts anyway (Anand, Tom's, ...), and whatever further information releases, will occur H2 of this year. ;) There may be more in industry publications, but a subscription is usually required (not commonly made available to the non-paying public).
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.