Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Status
Not open for further replies.
Anywhere here is an update to mine to show Tallest Skils idea with the handles. Thanks for the other feedback.

That design is awesome, if that was used by Apple I would seriously consider paying the extra cabbage for one over the iMac :D

or thats what that dopey English guy who designs all the products said on the Apple site when he was talking about the new aluminium unibody designs.

You mean Jonathan Ive, the guy that's been designing Macs since the original iMacs. Yeah, dopey doesn't really fly considering some of the great stuff he's penciled at Apple.
 
You mean Jonathan Ive, the guy that's been designing Macs since the original iMacs. Yeah, dopey doesn't really fly considering some of the great stuff he's penciled at Apple.

oh its all great, i love his designs, the unibody macbooks are a design feat. But come on, he's definately on weed or an extreme dose of beta blockers.
 
You totally rock, you know that? Great mockup. Here I am, at one of the best computer engineering schools in the country, and all I have this entire month is one gen-ed class, because that's all they let us take (one class)...

LOTS of down time, lots of boringness, no major-related classes, and I was in Ireland last semester, so I don't KNOW anyone in my field that has the passion I do for this stuff... I've been trying to teach myself this stuff (CS4, 3D mockup software, etc.) but it isn't going too well...

So which of the four Tylersburgs are they going to use?

DP/36 in the twin quad core arraignments and possibly either SP 24 or 36 if there's a lower end single quad core configuration.

Can you site a source for this?

Everything you need to know about the tylersburg chipsets
http://xtreview.com/addcomment-id-3366-view-Intel-Tylersburg-chipset.html

List of Xeon 3500/5500 CPUs and prices
http://amdintel-cpu-review.blogspot.com/2008/11/intel-is-planning-to-launch-xeon-5500.html

I had to. I love the design, but I just had to. :D

Well, Mac Pro reverse oreo/ The Mac Pro Oreos have have blacks side and silver middle. Choasbunny did a great job on that mockup.
 
Bloomfield/Gainestown. With the exception of a second QPI link for dual CPU configurations, these two chips are identical. Both will work in either SP or DP tylersburg motherboards, though bloomfield will be limited to a single CPU in DP motherboards. Gainestown (Branded Xeon 5500), however will be produced in a wider array of frequencies than Bloomfield (branded core i7 and Xeon 3500)

ECC memory support: There is some miss information about the Nehalem platform not supporting ECC memory. This is both true and false. Bloomfield CPUs with the consumer Core i7 branding have ECC memory support disabled. Bloomsfield CPUs with the Xeon 3500 branding and Gainstown Xeon 5500s have full ECC-support up to DDR3 1333.

Single CPU configuration:
When Apple moved to the PowerPC 970 (G5) they created a revolutionary new northbridge to accompany it called U3. U3 was designed be modular. U3 and U3H were designed for use in dual CPU workstations and servers. It was designed to connect to the K2 southbridge with an optional PCI-X controller between them at 1.6, 3.2, or 4.8 GBps through 8 and 16-bit Hypertransport buses. This allowed Apple to offer its G5s with single or dual CPUs, PCI or PCI-X, with 4 or 8 DIMM slots. The family also consisted of U3L designed for desktops. It was paired with a southbridge called shasta. U3L used a physically smaller motherboard and showed up in both the 1.8ghz single CPU PMG5 and the AGP iMac G5.

What does an old PPC G5 chipset have to do with the next revision of the Mac Pro? Perhaps quite a lot. On the intel platform, there was a divide. Desktops and servers/workstations used slightly different chips, with different sockets, different memory, and different chipsets. These factors have kept the base single CPU MacPro as a very expensive proposition as an entry-level Core 2 PowerMac would have been a completely different beast. With Nehalem/Tylersburg, this is no longer the case. The situation is much the same as G5/U3, same basic CPU core, variations of the same chipset, same socket, same memory. Xeon 3500/5500 and Tylersburg SP/DP are for the most part interchangeable. What this means for configurations is unknown, but the hurdles that have kept the line well over the $2000 mark are potentially solved.


See? I'm not the only one who thought of this!
 
That is truly hideous... NO BLACK ON THE MAC PRO PLEASE!!!!!
It would look like s hit

Like it or not, black and silver are the new black over at Apple, so you Gainestown boys are gonna get some. At least my Penryn (and current ACD30) will remain a lot more beautiful then your black and aluminum MPs and yer black and aluminum and glass :)eek:) ACDs.

:D
 
Nothing there indicates that 5500 Xeons will work on single socket boards, or visa versa. Couldn't find anything anywhere else either. I'm not saying it isn't possible or probable, but I don't see anything that specifically says it is possible and is going to be supported.

Usually the single socket systems have a different pinout. You can have one cpu in a multicpu system with the others empty. I don't remember if you have to have a blank or not. I also am not sure it would be any more cost effective to have one (or more) of the sockets removed in a multicpu system. As you already have all the extra pins for communicating with the other cpu's.
 
I'm betting on the discontinuation of the 20", but a 40"... Ehh...

Is there evidence to support it?
There was a rumor two (?) years ago that said Apple might release larger-than-30" ACDs, but...

But I'm wishing. :p

DP/36 in the twin quad core arraignments and possibly either SP 24 or 36 if there's a lower end single quad core configuration.
Two different boards...not sure about that.
 
But I'm wishing. :p

Two different boards...not sure about that.

Wouldn't it be the same board though, just a different NB/IO hub?

Ok, does ANYONE know if DDR3 can be put on a riser card or not? I always thought that was one of the reasons for FB-DIMM, it allowed that since it removed the need to directly wire every pin of every DIMM slot to the NB. In any case, if DDR3 does require a direct connection, that would be a 1440-pin connection for each 6-slot riser! :eek:

Also, if there is a cap of 9 DIMMs per CPU, does that mean 12 slots could only be done when running with 2 CPUs?
 
Nothing there indicates that 5500 Xeons will work on single socket boards, or visa versa. Couldn't find anything anywhere else either. I'm not saying it isn't possible or probable, but I don't see anything that specifically says it is possible and is going to be supported.

This should help. The second quicklink is dormant if if there is not another CPU present.
http://www.hardwarezone.com/articles/view.php?id=2516&cid=2&pg=3

Usually the single socket systems have a different pinout. You can have one cpu in a multicpu system with the others empty. I don't remember if you have to have a blank or not. I also am not sure it would be any more cost effective to have one (or more) of the sockets removed in a multicpu system. As you already have all the extra pins for communicating with the other cpu's.

Plus when's the last time you heard of two related CPUs with the same pin layout not being compatible with two related motherboards? Intel used the same pin layout on purpose.

There was a rumor two (?) years ago that said Apple might release larger-than-30" ACDs, but...

But I'm wishing. :p

Two different boards...not sure about that.

PMG5 had three.

Wouldn't it be the same board though, just a different NB/IO hub?

Actually same NB/IO hub, just slightly different variants. Saves intel a bundle on hardware development and driver work while offering flexibility to the consumer.
 
This should help. The second quicklink is dormant if if there is not another CPU present.
http://www.hardwarezone.com/articles/view.php?id=2516&cid=2&pg=3

Plus when's the last time you heard of two related CPUs with the same pin layout not being compatible with two related motherboards? Intel used the same pin layout on purpose.

Right, but when was the last time Intel had DP and UP processors work on the same board? Logically the possibility makes sense. I see where you are comming from and I can see where the benfits could be for Intel, vendors and consumers. However I'd have expected someone to have specifically stated that UP processors will work on a DP chipset and vice versa by now.

Maybe it's just assumed by tech writers who are aware of the feature that it's common knowledge so relaying the information has been overlooked.
 
Also, if there is a cap of 9 DIMMs per CPU, does that mean 12 slots could only be done when running with 2 CPUs?

Yes. Without a second CPU there is no way for the other one to access the extra memory branch as each branch is connected to the system via the intergrated memory controller on the processor. So if Apple offer a single processor on a dual socket board (as they currently do) it would be limited to 6 DIMMs.
 
Very nice mockups there. Tallest Skil, how about a post or a section in your first post that links to all the mockups, if such a section is needed?

And I don't know if any Xserve discussion is appropriate in this thread (because it and the Mac Pro have similar CPUs and release dates), but if Apple wants to keep the 80 W TDP of the current Xserve Harpertowns, the maximum speed that can be used is 2.53 GHz ($744 as opposed to the $915 of the 3.0 GHz Harpertown). That 2.53 GHz CPU is also the top of the 1067 MHz RAM Gainestowns.

I hope it doesnt look like that...
 
I wonder how well these new CPU's are going to handle heat?

I always wondered why Apple doesn't look towards some kinda of liquid cooling to regulate temperature.

I also vote Apple lets us into the BIOS on the Mac Pros.

--Garrett
 
I wonder how well these new CPU's are going to handle heat?

I always wondered why Apple doesn't look towards some kinda of liquid cooling to regulate temperatur
e.

I also vote Apple lets us into the BIOS on the Mac Pros.

--Garrett

they did implement liquid cooling with the G5's

needless to say, many have leaked

there is no bios in the mac pro by the way
 
You learn stuff everyday.

What was leaking? I have yet to have a problem. . . .

The coolant.

http://www.xlr8yourmac.com/systems/G5_coolant_leaks.html


And EFI is no reason to dump the setup utility. Heck, they could put a Safari in there if they wanted to. But they won't.

MSI_Click_BIOS_01.jpg
 
Status
Not open for further replies.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.