Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Mac Pro 2011 "Sandy Bridge E" everything we know
- Thread starter Amethyst
- Start date
- Sort by reaction score
Unless someone provides proofs of OCing and alternating Mac Pro's Turbo speeds, all posts about those will be deleted. This is not relevant to Mac Pro because it doesn't seem to be possible, hence it will only create confusion. Making claims without being able to back them up counts as trolling too.
This is a final warning, next step is a clean-up.
This is a final warning, next step is a clean-up.
Oh. I get it. Claiming without proof is the offense. I was going to say. Why would you get booted from Macrumors when there are threads on GPU hacks, SSD hacks and proc replacements? But a CPU overclock will get you ousted?
So do you have any proof tutor or should we look for a new "mystory"?
Proof is not bench figures as they can easily be altered.
It is possible, because everything is hackable.
But obviously we need proof.
I think a video of a known benchmark would be the proof, can't adjust the time then.
What happened to Moore's law. Is it me or do cpu speeds (not overall performance) seem stuck for 3 years now? Just saying . . .
Err, Moore's Law says that number of transitors will double about every two years. You are counting MHz when you should be counting cores. Moore's Law is working just fine. It is about to hit the wall but there is nothing in the shipping Xeon's that point to that.
As a side-effect of making the transistors smaller there has usually been an uptick in clock speed. That isn't what the law is about. There are clock distribution and leakage issues that have popped up but those don't mean you can't put more transistors to work. There are limits to how big you can make a single core. But having a 500M transistor single core isn't necessarily a good idea anyway.
For the E5's if you turn off 50+% of the transistors you can go faster ( the 4 core E5 1600 series has the higher clock speed over the previous versions. ). The clock speed junkies just need to look for E5 1600 boxes.
Last edited:
Moore's law is about transistor count, not about frequency. There has definitely been an increase in the transistor count, although it hasn't always been 2x like Moore's law states.
Not necessarily. The issue lies in silicon which meets its wall at 7-8nm. Graphene and germanium are potential successors. The former can go as small as 0.5nm and simple transistors have been clocked as high as 155GHz.
It will hit the wall well before then. One of the significant problems all along has been lithography.
older IBM paper.
"... Lithography. This has long been the technological driver of shrinking circuits, ..."
http://domino.watson.ibm.com/comm/wwwr_thinkresearch.nsf/pages/cmos398.html
while folks have moved onto EUV sources. E.g. TSMC..
"...At that node, TSMC is also evaluating other lithography candidates, namely EUV and maskless. The foundry provider will initially go with 193-nm immersion at 20-nm production, but it may also deploy EUV or maskless in 2013-to-2014, depending on the viability of those technologies, he said. ..."
http://www.eetimes.com/electronics-news/4088580/TSMC-skips-22-nm-rolls-20-nm-process?pageNumber=1
You are also going to run into physical problems that are above the size of the atoms. Same article:
"... Kevin Nowka, senior manager of VLSI Systems at the IBM Austin Research Lab, described the physical design challenges beyond the 22-nm node, emphasizing that sub-wavelength lithography has made silicon image fidelity a serious challenge. ... "
For the last several years there have been a series of work arounds for lithography. The list isn't getting shorter. However, at some point will run out of hacks. What is needed is something that can do high volume, cost-effective "printing" of circuits.
http://www.electroiq.com/articles/s...phy-with-complementary-ebeam-lithography.html
but just being able to "draw it" is just a start.
Similar issues highlighted here:
"... At 22nm and beyond, manufacturing can no longer cleanly abstract away the underlying physical challenges of semiconductor scaling. This drives the need for co-optimization between process technology and chip design. The inescapable conclusion is that the physical design of integrated circuits is becoming ever more critical at smaller geometries. ..."
Also Note that we have a shrinking number of foundries at the upper limits.
http://www.cpushack.com/2011/02/01/shrinking-process-size-shrinking-foundry-selection/
Fewer foundries means fewer chip making vendors ( because fewer customers to sell to. ). One of implicit forces at work in Moore's law is not just that physics are not in the way (haven't till now), but that there is a strong competitive market to remove barriers. A club of 4 (or 5-6 ... I suspect IBM may join just a bit later. )
Stuff like graphene works great in isolated control labs. I still haven't seen anything saying it works well out in the real world.
" ... That means that unlike silicon, which can be switched off, graphene continues to conduct a lot of electrons even in its "off" state. ... "
http://www.technologyreview.com/Nanotech/20119/
What happens when your power, battery, or minor static discharge sends some 'extra' charge through your device ? There is a very good reason silicon has been the preferred material so far.
Again there are economics of decoupling from the more mainstream CMOS processes. Everybody moving down (at different rates) helps drive the Moore's law effect. If some folks fork off into something now the effect to drop off.
I've also seen a good bit of info about AMD's bulldozer processors and the A7s from IBM that have clock speeds approaching 4.5-5.0GHz.
I think we were in a bit of a rut with clock speeds because the value of more cores was seen as more important and for us to get more cores on the same sized chip meant clock speeds suffered a bit.
I think now we're seeing that change and clock speeds will make a jump here pretty quickly.
I think we were in a bit of a rut with clock speeds because the value of more cores was seen as more important and for us to get more cores on the same sized chip meant clock speeds suffered a bit.
I think now we're seeing that change and clock speeds will make a jump here pretty quickly.
I predict Apple will use.....
I suspect Apple will use these processors in the new Mac Pro
Originally Posted by Amethyst View Post
Dual Core
Xeon E5-2637 / 2 Core (4 Threads) / 3.0 GHz / L3 5 MB / DDR3-1600 / TDP 80 Watt
Quad Core
Xeon E5-2603 / 4 Core (4 Threads) / 1.8 GHz / L3 10 MB / DDR3-1066 / TDP 80 Watt
Xeon E5-2609 / 4 Core (4 Threads) / 2.4 GHz / L3 10 MB / DDR3-1066 / TDP 80 Watt
Xeon E5-2643 / 4 Core (8 Threads) / 3.3 GHz / L3 10 MB / DDR3-1600 / TDP 130 Watt
6 Core
Xeon E5-2620 / 6 Core (12 Threads) / 2.0 GHz / L3 15 MB / DDR3-1333 / TDP 95 Watt
Xeon E5-2630 / 6 Core (12 Threads) / 2.3 GHz / L3 15 MB / DDR3-1333 / TDP 95 Watt
Xeon E5-2640 / 6 Core (12 Threads) / 2.5 GHz / L3 15 MB / DDR3-1333 / TDP 95 Watt
Xeon E5-2667 / 6 Core (12 Threads) / 2.9 GHz / L3 15 MB / DDR3-1600 / TDP 130 Watt
Xeon E5-2630L / 6 Core (12 Threads) / 2.0 GHz / L3 15 MB / DDR3-1333 / TDP 60 Watt
8 Core
Xeon E5-2650 / 8 Core (16 Threads) / 2.0 GHz / L3 20 MB / DDR3-1600 / TDP 95 Watt
Xeon E5-2660 / 8 Core (16 Threads) / 2.2 GHz / L3 20 MB / DDR3-1600 / TDP 95 Watt
Xeon E5-2665 / 8 Core (16 Threads) / 2.4 GHz / L3 20 MB / DDR3-1600 / TDP 115 Watt
Xeon E5-2670 / 8 Core (16 Threads) / 2.6 GHz / L3 20 MB / DDR3-1600 / TDP 115 Watt
Xeon E5-2680 / 8 Core (16 Threads) / 2.7 GHz / L3 20 MB / DDR3-1600 / TDP 130 Watt
Xeon E5-2690 / 8 Core (16 Threads) / 2.9 GHz / L3 20 MB / DDR3-1600 / TDP 135 Watt
Xeon E5-2687W / 8 Core (16 Threads) / 3.1 GHz / L3 20 MB / DDR3-1600 / TDP 150 Watt
Xeon E5-2650L / 8 Core (16 Threads) / 1.8 GHz / L3 20 MB / DDR3-1600 / TDP 70 Watt
Although possibly not the E5-2630 (too slow)
I suspect Apple will use these processors in the new Mac Pro
Originally Posted by Amethyst View Post
Dual Core
Xeon E5-2637 / 2 Core (4 Threads) / 3.0 GHz / L3 5 MB / DDR3-1600 / TDP 80 Watt
Quad Core
Xeon E5-2603 / 4 Core (4 Threads) / 1.8 GHz / L3 10 MB / DDR3-1066 / TDP 80 Watt
Xeon E5-2609 / 4 Core (4 Threads) / 2.4 GHz / L3 10 MB / DDR3-1066 / TDP 80 Watt
Xeon E5-2643 / 4 Core (8 Threads) / 3.3 GHz / L3 10 MB / DDR3-1600 / TDP 130 Watt
6 Core
Xeon E5-2620 / 6 Core (12 Threads) / 2.0 GHz / L3 15 MB / DDR3-1333 / TDP 95 Watt
Xeon E5-2630 / 6 Core (12 Threads) / 2.3 GHz / L3 15 MB / DDR3-1333 / TDP 95 Watt
Xeon E5-2640 / 6 Core (12 Threads) / 2.5 GHz / L3 15 MB / DDR3-1333 / TDP 95 Watt
Xeon E5-2667 / 6 Core (12 Threads) / 2.9 GHz / L3 15 MB / DDR3-1600 / TDP 130 Watt
Xeon E5-2630L / 6 Core (12 Threads) / 2.0 GHz / L3 15 MB / DDR3-1333 / TDP 60 Watt
8 Core
Xeon E5-2650 / 8 Core (16 Threads) / 2.0 GHz / L3 20 MB / DDR3-1600 / TDP 95 Watt
Xeon E5-2660 / 8 Core (16 Threads) / 2.2 GHz / L3 20 MB / DDR3-1600 / TDP 95 Watt
Xeon E5-2665 / 8 Core (16 Threads) / 2.4 GHz / L3 20 MB / DDR3-1600 / TDP 115 Watt
Xeon E5-2670 / 8 Core (16 Threads) / 2.6 GHz / L3 20 MB / DDR3-1600 / TDP 115 Watt
Xeon E5-2680 / 8 Core (16 Threads) / 2.7 GHz / L3 20 MB / DDR3-1600 / TDP 130 Watt
Xeon E5-2690 / 8 Core (16 Threads) / 2.9 GHz / L3 20 MB / DDR3-1600 / TDP 135 Watt
Xeon E5-2687W / 8 Core (16 Threads) / 3.1 GHz / L3 20 MB / DDR3-1600 / TDP 150 Watt
Xeon E5-2650L / 8 Core (16 Threads) / 1.8 GHz / L3 20 MB / DDR3-1600 / TDP 70 Watt
Although possibly not the E5-2630 (too slow)
Last edited:
http://nextbigfuture.com/2011/08/will-graphene-replace-silicon-in.html
I can't say I know too much about the possible replacements. Graphene seems to be one possibility but it has its shortages (like any element).
I'd be somewhat surprised if they took away a the 4-core SP bases. I suspect a lot of people will like that 3.3GHz model. Especially people that tend use more single threaded programs.
What is the motivation for selecting five different E5 26xx models? The dual package Mac Pro line-up has consisted of 3 different models for the last several upgrade releases.
There is very little reason to pick any of these for the single package models ( again, likely at most 3 of them. )
For me Apple now plans to quit production of the Mac Pro
For me Apple now plans to quit production of the Mac Pro.
It is a shame that all Apple products have been updated except for the MAC PRO, this may indicate an intention to cease selling the Mac Pro
For me Apple now plans to quit production of the Mac Pro.
It is a shame that all Apple products have been updated except for the MAC PRO, this may indicate an intention to cease selling the Mac Pro
For me Apple now plans to quit production of the Mac Pro.
It is a shame that all Apple products have been updated except for the MAC PRO, this may indicate an intention to cease selling the Mac Pro
It's not Apple's fault that Intel has not yet shipped the new Xeons.
On the other hand, I really don't think it will take that much time (March 2012) to see the new Mac Pro computers if Intel starts shipping in batch.
I'm pretty sure Apple has had these processors in their testing labs for quite some time, I think the logic boards and everything else are ready. It's just a matter of time before they can pop the cpu's into them.
I might be wrong though.
For me Apple now plans to quit production of the Mac Pro.
It is a shame that all Apple products have been updated except for the MAC PRO, this may indicate an intention to cease selling the Mac Pro
You're right, it may indicate that.
More likely, though, is that Apple is waiting for there to be something to actually upgrade them to.
Yes, they and other vendors have had access to engineering samples in order to validate their prototype boards.
Now they need to get their hands on final parts for manufacturing, then have to validate what comes off of the line (mistakes can happen, such as the wrong part value, bad solder joints, bad PCB <etching or shorts>,...). Once they clear everything, then they'll ramp up production and get finished systems out the door.
ah, okay. I didn't know that ES samples are different than what we might get as the final product.
But what you're saying makes sense
This is all very interesting conjecture and fun to read - but in the end I would find it difficult to believe that the Mac Pro will just disappear. I'm rather fond of the aluminum tower, but perhaps it will be distinctly different. Only Apple knows for sure.
I'm not holding my breath, but am waiting patiently for the new year and an announcement. Until then the 3,1 works just fine. Perhaps someone should send Tim a note and ask him directly.
I'm not holding my breath, but am waiting patiently for the new year and an announcement. Until then the 3,1 works just fine. Perhaps someone should send Tim a note and ask him directly.
A fine idea tamvly. Let us know what he says.
doubt he would reply anything other than "when it's ready..."
Intel needs the feedback on the ES chips in order to discover any issues. System vendors need the ES chips in order to validate their PCB designs (test circuit, PCB traces, component selection,...).ah, okay. I didn't know that ES samples are different than what we might get as the final product.
But what you're saying makes sense
In terms of the final system validation (what will be rolling off of the assembly line), they're looking for multiple things that may have unknowingly deviated from the final design that was approved. For example, the wrong part value could be loaded into a pick-n-place system (machine that actually puts the individual electronic components where they go).
Intel release date was March 30, 2010, and Apple announced the 2010 MP's on July 27, 2010.
So we're talking about 4 months after Intel's release date before 2010 MP systems reached users.
It's not Apple's fault that Intel has not yet shipped the new Xeons.
On the other hand, I really don't think it will take that much time (March 2012) to see the new Mac Pro computers if Intel starts shipping in batch.
I'm pretty sure Apple has had these processors in their testing labs for quite some time, I think the logic boards and everything else are ready. It's just a matter of time before they can pop the cpu's into them.
I might be wrong though.
One last time, you don't need a Xeon to have workstation performance.
Seriously, take that $2,500 for the junk base Mac unPro and build
or buy a screaming fast i7 system with 2x-3x more specs than the Mac unPro.
Any hardware options you want and a choice of more than 3 stupid two year old video cards.
Why bother with an iToy company that thinks computers are trucks?
They don't care anymore.
I wish they would do like HP and spin off the computer biz
and let somebody run it who likes computers.
