Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
E5-2600V2s will have up to 12 cores each. 24-core Mac Pro is possible.

It seems pretty unlikely though given what the price will likely be.

If Intel is offering everything from 6-12 cores then, I'm guessing the 12 core versions will be unreasonably expensive. I think we'd be lucky to see Apple use the 10 core varients even.

Of course, we'll know more when intel releases details about the pricing across E5-2600V2s, but I'd guess at a configuration looking like:

2x 6-core high clock rate
2x 8-core low clock rate
2x 8-core high clock rate

And by high, I'm thinking 2.5-2.8 GHz, and low more like 2.0-2.3 GHz.

Maybe the 10-core sneaks in to the top spot, but that would come with a base GHz drop. Though turbo boost may allow similar single/low threaded speeds.
 
It seems pretty unlikely though given what the price will likely be.

If Intel is offering everything from 6-12 cores then, I'm guessing the 12 core versions will be unreasonably expensive. I think we'd be lucky to see Apple use the 10 core variants even.

Ivy Bridge is implemented on a pretty substantial process shrink from 32nm to 22nm ( -33% )

Intel is likely splitting off the 1600 series from the 2600 so the expectation over core count range should be split.

First the 1600s will likely go from 4 to 6 cores and targeting the market for affordable "high GHz" solutions. There are still going to be 4 core offerings inside of that subset. The 2600 is the subset where "core count" is going to be the primary value offering.

Of course, we'll know more when intel releases details about the pricing across E5-2600V2s, but I'd guess at a configuration looking like:

2x 6-core high clock rate
2x 8-core low clock rate
2x 8-core high clock rate

And by high, I'm thinking 2.5-2.8 GHz, and low more like 2.0-2.3 GHz.

On the transition from Westmere to Sandy Bridge the 4 core xx20 offering disappeared (the majority of the line up got bumped up 2 cores). That was with no process shrink. Seems more likely if there is a 12 core offering that the E5 2620 v2 2640 v2 and 2665 v2 would be respectively

2x 8-core 2.1 GHz ( ~$425 processor )
2x 8-core 2.6 GHz (~$890 processor )
2x 10-core 2.5 GHz (~$1500 processor )

And closer to same ranges at the Sandy Bridge models ( a 100MHz bump on base and broader dynamic (Turbo) range; add another 200-300MHz to top end. ) The jump won't be a 400-500MHz base rate bump in speed but more so 2 core bump. With better power management, they could crank up speed only when 1-4 cores really very active.


The point would be that a two 2620 v2 set-up would just cost around $850. The 24 core set up would be "crazy" priced, but the primary objective is to make the 16 core set up far more affordable. AMD may be having problems, but they aren't stumbling that badly. Haswell is going to take a while to arrive and Intel needs some gap while AMD evolves in the mean time.



Maybe the 10-core sneaks in to the top spot, but that would come with a base GHz drop. Though turbo boost may allow similar single/low threaded speeds.

If 12 is the top spot then 10 isn't gong to be all that rare (or priced in the stratosphere). With Sandy Bridge there are eight 8 core offerings in the 2600 series. There are just five 6 core offerings in the series. There are more 8's than 6's. It is quite likely to be the same ratio with a -33% process shrink to support it. There likely will be more 10's than 8's and maybe two or three 6's ( and maybe one odd ball 4 core that does make much sense in a Mac Pro like current E5 2643 : 4 cores 3.3GHz 130W and $885 price tag. )

The fact that Intel is even trying to drop a 12 count on the series shows where the focus is.
 
Last edited:
Apple usually skips the most expensive ones. Somehow I doubt they are going to double core count this round.

They offered the most expensive Xeons from 06 to 09 and the second most expensive after that. They only have to offer what is standard, but either way if there is an Ivy Bridge DP mac pro then people can have 24 core systems
 
If the Xeon doesn't gain Thunderbolt support, I think the chances of the Mac Pro having Xeons becomes slimmer (sadly).

Yeah, I know, what's the point of Thunderbolt on a desktop, blah blah blah... I just don't see Apple doing a Mac Pro without Thunderbolt. I just think it's a little too early to be looking at the Xeon roadmap.

My understanding is that the E3 would probably support Thunderbolt, but I'm not sure on that.
 
They offered the most expensive Xeons from 06 to 09 and the second most expensive after that. They only have to offer what is standard, but either way if there is an Ivy Bridge DP mac pro then people can have 24 core systems

I thought they stopped in 09. I know they adjusted the retail pricing relative to cpu choices considerably in 09. A 24 core mac pro would be cool, but I hope that more of the software I use turns towards CUDA for highly parallel tasks.
 
I thought they stopped in 09. I know they adjusted the retail pricing relative to cpu choices considerably in 09. A 24 core mac pro would be cool, but I hope that more of the software I use turns towards CUDA for highly parallel tasks.

06 Through to february 09 I mean :)
 
If the Xeon doesn't gain Thunderbolt support, I think the chances of the Mac Pro having Xeons becomes slimmer (sadly).
....
My understanding is that the E3 would probably support Thunderbolt, but I'm not sure on that.

Again it is deeply unclear whether folks are conflating integrated graphics for "Thunderbolt support". It is far easier to resource for Thunderbolt if have a Xeon E3 with a integrated GPU. That solves what flavor of embedded GPU ( integrated or soldered discrete component ) going to use to provide the required DisplayPort signals to motherboard traces. Intel has largely done all the "homework" needed to do a logical board implementation in the reference design. Just copy and tweak it with some modifications.

Intel hasn't done other folks homework for them on Xeon E5 reference designs. The implementation also has to select and embed a discrete GPU.
Most server motherboard experience these days with GPU is with the ancient Matrox GPU. There really isn't someones homework to crib from.


Eventually integrated GPU will likely creep into at least the E5 1600 (or future equivalent ) series. But there no firm evidence that Apple needs to wait that long. Pull the Intel GPU from the iMacs implementation and still have the core of a solution that would work for a Mac Pro.
 
Apple usually skips the most expensive ones. Somehow I doubt they are going to double core count this round.

Probably more accurate is that Apple skips the more expensive ones than they used in previous iterations (unless forced to move up because no other choices).

The upper 10% of the Xeon line has had price creep. Especially after AMD pragmatically went uncompetitive. Apple is not tracking that increase with their systems (even in BTO). There is already enough moaning and groaning about Mac Pro costs anyway. Tracking that trend really doesn't fit Apple's general product strategy. Nor would it be very effective after Apple tacks on their additional 30% mark-up on the CPU package option. Intel's nonlinear pricing + 30% on top is just going to largely suppress demand.


Apple's general Mac strategy for last 4-5 years has been to fill the box with value to hit the generally the same price points at the same margins. ( some exceptions for new technology form factors, but for mature products the prices don't move. ).

When folks pick $500 more expensive processors and want Apple to keep the price point the same.... might as well be asking for a lunch date with Halle Berry. It isn't likely going to happen.
 
Ivy Bridge is implemented on a pretty substantial process shrink from 32nm to 22nm ( -33% )

Intel is likely splitting off the 1600 series from the 2600 so the expectation over core count range should be split.
...
2x 8-core 2.1 GHz ( ~$425 processor )
2x 8-core 2.6 GHz (~$890 processor )
2x 10-core 2.5 GHz (~$1500 processor )

And closer to same ranges at the Sandy Bridge models ( a 100MHz bump on base and broader dynamic (Turbo) range; add another 200-300MHz to top end. ) The jump won't be a 400-500MHz base rate bump in speed but more so 2 core bump. With better power management, they could crank up speed only when 1-4 cores really very active.

I suppose if there is 6, 8, 10 and 12 core packages, a $1500 10-core with modest clock speed is likely. It is certainly a lot of core counts and clock speed combinations that intel will have to play with, and it will be very interesting to see the final structure of E5-2600 V2.

If that base 8-core model can turbo to 3.0 GHz, it would certainly be a nice entery level machine that won't get too crushed in single/low threaded parts of a work flow. And I hope Apple would make the choice to use such a processor, instead of the a 6-core at similar price, an extra 400 MHz, but a lower dinamic range (similar to how the current 2640 and 2650 sit, just cheaper). I suppose much will depend on the pricing structure from Intel, but if that 8-core 2.1 GHz is more like a ~$500, and Intel gives us a 6-core 2.4GHz for $400, I can see Apple choosing the 6-core.

I guess its more likely that the Ivy Bridge 6-core goes the way of the Sandy Bridge 4-core though, and we may just see an expensive, very high clock rate 6-core, which Apple would not pick. But hopefully we get more details from intel soon.
 
If that base 8-core model can turbo to 3.0 GHz, it would certainly be a nice entery level machine that won't get too crushed in single/low threaded parts of a work flow.

Intel still wants to sell those 1600s to users with a higher mix of low threaded work flows. So probably won't. At least this iteration. Haswell has far more fine grained power management. Probably will get closer to 3 but the 4 core mainstream offerings will have moved well past that by that time.


The 2620's range is only from 2.0-2.5. A v2 range of 2.1-2.7 is more likely. Maybe 2.1-2.8. A dynamic range of 500 is mainly used across the line up in Sandy Bridge. Pushing that to 700 is probably as far as it goes this iteration. (more likely only 600).

Here are v2 bumps from the Sandy bridge -e
http://www.cpu-world.com/news_2013/...idge-E_extreme_CPUs_to_launch_in_Q3_2013.html

It isn't huge dynamic range shift. But these are clocked much higher too at higher base rates.




. I suppose much will depend on the pricing structure from Intel, but if that 8-core 2.1 GHz is more like a ~$500, and Intel gives us a 6-core 2.4GHz for $400, I can see Apple choosing the 6-core.

I expect Intel to go after AMD. There is a chart here of AMD's two socket versus Intel's.

http://www.anandtech.com/show/6508/the-new-opteron-6300-finally-tested/2

If the 2620 v2 was 8 core and $420 priced against the AMD 6376 ( $703 ) then AMD would feel some pain. Intel may though. 6 core at $400 is a bigger margin markup for Intel at this new process level. Right now Intel is leaving heavily on server processors to keep the upward growth on profits.

I don't think Intel is going to goose the base speed on the 2620 because that leaves them less flexibility to goose the speed on the top end 130W parts. Something like the 2643 will be around but it won't be priced in the $400 zone. More like double that.


But hopefully we get more details from intel soon.

late May - early June maybe the leaks will be more detailed. I don't think Intel is going to officially tell until August or September.
 
Last edited:
Prosumer

I have a sinking feeling Apple's going to release some "prosumer" garbage. If you look at their desktop strategy (or lack thereof) its clear they are taking cues from content consumption devices (iPad) and awkwardly trying to roll them back into content creation devices (iMac). This is why it has taken them so long to update the Mac Pro -- they have no idea what it is as it no longer fits into their strategy. They have nothing better to offer because it is fine as is, and whatever "redesign" they will throw at it won't generate the kind of excitement of the iDevices because it is a professional's tool. All of Apple's hardware has transitioned to the most wasteful, non-user upgradeable junk imaginable. The kind of thing you buy, use, and throw away. I say get a Mac Pro while people are selling them off and throw a new graphics card in it. A few people are even selling new towers. http://tinyurl.com/dym45fj I hope I'm wrong but I don't think it looks good for the Mac Pro.
 
I was looking at a refurb 12core 3.06 the other day.. the thing is the E5 and even E3's are a good clip faster than that proc (x5675?). Add in ram, and ssd, and 680/7950 gfx card and you are back to $7k for a 2+ yr old box.
 
Last edited:
Your thinking has proven correct! There is no market for the Mac Pro as we know it because Apple direction is the consumer and prosumer. If anything I expect the so called New Mac Pro to be limited in expandability options and also no optical drive...

It will be nothing like the mac pro we all know and love... especially for its 4 hard drive bays and excellent dedicated slot for graphics.. I expect it will by a hybrid of the mac mini or iMac with PRO stuck on it..

Since the PowerPC era... Apple's stuff has been nothing to write home about in terms of quality.. and OS X has gotten worse since the end of Snow Leopard..

I have a sinking feeling Apple's going to release some "prosumer" garbage. If you look at their desktop strategy (or lack thereof) its clear they are taking cues from content consumption devices (iPad) and awkwardly trying to roll them back into content creation devices (iMac). This is why it has taken them so long to update the Mac Pro -- they have no idea what it is as it no longer fits into their strategy. They have nothing better to offer because it is fine as is, and whatever "redesign" they will throw at it won't generate the kind of excitement of the iDevices because it is a professional's tool. All of Apple's hardware has transitioned to the most wasteful, non-user upgradeable junk imaginable. The kind of thing you buy, use, and throw away. I say get a Mac Pro while people are selling them off and throw a new graphics card in it. A few people are even selling new towers. http://tinyurl.com/dym45fj I hope I'm wrong but I don't think it looks good for the Mac Pro.
 
I have a sinking feeling Apple's going to release some "prosumer" garbage. If you look at their desktop strategy (or lack thereof) its clear they are taking cues from content consumption devices (iPad) and awkwardly trying to roll them back into content creation devices (iMac). This is why it has taken them so long to update the Mac Pro -- they have no idea what it is as it no longer fits into their strategy. They have nothing better to offer because it is fine as is, and whatever "redesign" they will throw at it won't generate the kind of excitement of the iDevices because it is a professional's tool. All of Apple's hardware has transitioned to the most wasteful, non-user upgradeable junk imaginable. The kind of thing you buy, use, and throw away. I say get a Mac Pro while people are selling them off and throw a new graphics card in it. A few people are even selling new towers. http://tinyurl.com/dym45fj I hope I'm wrong but I don't think it looks good for the Mac Pro.

You're right and you hit the nail right on the spot. Apple's thrust is into smaller devices now.
 
Given Apple's current direction with its Macs... Most of its Macs are hard if not impossible to expand and I fear the New Mac Pro or whatever replacement machine its gonna be is going to be a very closed system with little to really no internal expandability compared to what we have now..

Our current Mac Pros could be the LAST truly expandable machines.

Um...no it hasn't. It might be proven correct, but asserting 'proof' when the new Mac Pro or whatever hasn't even been released yet is just silly.
 
Given Apple's current direction with its Macs... Most of its Macs are hard if not impossible to expand and I fear the New Mac Pro or whatever replacement machine its gonna be is going to be a very closed system with little to really no internal expandability compared to what we have now..

Our current Mac Pros could be the LAST truly expandable machines.

Except that has been true for some time. Looking at the iMac and Mac Mini has never been a particularly useful barometer for how expandable the Mac Pro will be.
 
What about an iCube
I immagine an external box connected via thunderbolt to your iMac, MacMini MacBook etc that is just stuffed with Cpu power and a few pci slots (for graphic cards (or other special cards you might need) plus maybe ports like the Belking Thunderbolt) ?
This will comply with Mac "small is beautifull" and "consumer is beautiful" policy. The Os would be built around such an architecture and the user experience would be completely transparent. Work at home with huge speed, power and realtime rendering then just unplug one cable and go out with your MacBook
 
What about an iCube
I immagine an external box connected via thunderbolt to your iMac, MacMini MacBook etc that is just stuffed with Cpu power and a few pci slots (for graphic cards (or other special cards you might need) plus maybe ports like the Belking Thunderbolt) ?
This will comply with Mac "small is beautifull" and "consumer is beautiful" policy. The Os would be built around such an architecture and the user experience would be completely transparent. Work at home with huge speed, power and realtime rendering then just unplug one cable and go out with your MacBook

Not enough bandwidth or speed..

Why does everyone wnat to shrink the thing? What exactly is wrong with it the way it is? It's quiet, cool, and fits under the desk.
 
I was sceptical about a mac pro refresh particularly because of the xeon timing. However i can see it being previewed at WWDC then released later. The new beta's of OSX apparently contain new graphics drivers. This must mean a mac pro release is imminent surely ?

http://architosh.com/2013/05/evidence-of-new-mac-pros-at-wwdc-titan-drivers-in-latest-os-x-builds/

I'd like to get excited about a new Mac Pro too but haven't we been through new graphics card drivers with multiple previous OS X beta builds before? I just don't think this is the evidence of a new Mac Pro unfortunately :(
 
I think it will be a bouncy Mac Castle which can be inflated to meet the needs of any mortal. Need some more room for HDs? Just inflate it. Need to take it somewhere? Deflate it and put it in your pocket. Need to take it accross the pond? Use it as a boat. Just $1999.

Built in USA and floated accross the world using the Oceans natural currents.

Pump sold separately for $1000.

On a slightly more serious note, I don't think there will be a huge change, just a bump in specs and maybe thunderbolt but nothing to get exicted about.
 
Hmmm

I'd like to get excited about a new Mac Pro too but haven't we been through new graphics card drivers with multiple previous OS X beta builds before? I just don't think this is the evidence of a new Mac Pro unfortunately :(

I suppose your right really my hypothesis was flawed. However the fact remains they cannot keep supporting new graphics cards ad infinitum. The Titan card has PCIe 3 capability so eventually the Mac Pro will refresh.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.