Public institution but private grants for workforce training. Austin-area colleges are doing pretty well.
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.
From Facebook's "We Want a New Mac Pro"
- Thread starter 5050
- Start date
- Sort by reaction score
Public institution but private grants for workforce training. Austin-area colleges are doing pretty well.
Well that's definately not the norm in science/engineering across public schools.
----------
There must be data somewhere on OSX updates, or similar, that could track this right?
I can't attest for other area universities, but the one I teach at has mostly gone the iMac route for the lower level course labs along with the older Mac Pros (can't say for sure if they've ever been upgraded). But for the advanced classes and graduate labs they've gone with Boxx machines and a substantial render farm.
It just seems to me that those of us who prefer the all-in-one don't get anything out of the modular (via TB) approach except a higher price tag. The rest never really needed a Mac pro in the first place and were always looking for that missing product between iMac and MacPro.
The tower desktop market is quite soft, and getting softer there is nothing that Apple can/will do to change that fact. The home user no longer needs a "hub" to connect their devices to. If the home user needs some storage they go buy a buffalo drop their life on it and let their smart TV access it. The tower desk top has really been relegated to gamers, pro's, and stick in the muds like myself who don't know any better.
That's similar to what I've seen. Students use the university cluster on PCs, maybe iMacs, or their personal computers.
I think this is pretty accurate. You serve the people that where in the gap well, and you hit the low level Mac Pro users (those opting for the base model or close to it). Those that need the dual processor version, even the top single processor model, and high end GPUs are going to have to make compromises. Especially if we see something that needs the E3s because of TDP limits that won't allow for the E5-1600s. There is really no price reason to leave the E5-1620/50 for something like the E3-1245/80, the only reason Apple would do that is for TB and form factor. But you give up ECC RAM, over 300GB in max RAM capabilities (yes, I know the OSX issues, but those can be easily fixed if Apple actually cared), and an additional 2 cores for the 1650. So for those that really do want to make some sort of mini cluster out of this product, pricing goes through the roof for the same capabilities already possible in dual processor workstations.
My own guess is that the Mac Pro will be a new especially hairy-chested member of the Mac Mini family. Probably double-wide in a pizza-box format.
Big honking processor, lots of RAM (probably soldered for reliability), fast SSD, great GPU. And Thunderbolt for everything else including a PCIe crate and/or drive bays.
There are many good things about this. Thunderbolt is basically PCIe-on-a-wire (or fiber). So this approach lets the costs of expandability be borne by those who want it at the time they need it while allowing essentially unlimited expandability.
In one fell swoop Apple will have replaced both the pro and the XServe with something far superior to both.
There will be a lot of whining (it's starting already) but it will be awesome.
Big honking processor, lots of RAM (probably soldered for reliability), fast SSD, great GPU. And Thunderbolt for everything else including a PCIe crate and/or drive bays.
There are many good things about this. Thunderbolt is basically PCIe-on-a-wire (or fiber). So this approach lets the costs of expandability be borne by those who want it at the time they need it while allowing essentially unlimited expandability.
In one fell swoop Apple will have replaced both the pro and the XServe with something far superior to both.
There will be a lot of whining (it's starting already) but it will be awesome.
My guess is Apple has that data. Steam might too, but I doubt you could drill down like that on their site.
Yep, seems like a bit of verification of this thread.
Also, this probably means no Mac Pro at WWDC. Apple wouldn't do damage control if it's less than a week away. That does probably mean Apple is sticking with the Xeon E5.
There is a pretty big gap price wise. Two "Good" , "Better" , "Best" line ups core foundations ( v2 and v1 pricing but the v3 and v2 pricing coming up should be similar):
E3's ( Pulling iGPU xxx5 models to ease TB impelement. Is suspect v3 are bumped up on GHz. )
E3 1225 v2 3.2 GHz 4 cores $209
E3 1245 v2 3.4 GHz 4 cores $266
E3 1275 v2 3.5 Ghz 4 cores $339
http://www.cpu-world.com/news_2012/2012020701_Prices_of_Xeon_E3-1200_v2_CPUs.html
They probably need to do more differentiate on different PCI-e cards to separate these out more. (just like on the iMac line up only with desktop cards to push the overall system into higher range. )
E5 ( again three to )
E5 1620 v1 3.6 GHz 4 cores $294
E5 1650 v1 3.2 GHz 6 cores $583
E5 1660 v1 3.3 Ghz 6 cores $1080
http://www.cpu-world.com/news_2012/2012020701_Prices_of_Xeon_E3-1200_v2_CPUs.html
The 1620 almost starts off where the E3 1275 ends. Trying to limbo back down to the $2,000 border would definitely be possible making that as the foundation swap. Especially with the task of trying to bring the top end of the line up back under the $2,600 mark.
Xeon E3 has this. (somewhat similar to how the Core i 49xx has it switched off from the basic Xeon E5 architecture so to be a Core i offering).
Pragmatically an evolutionary Mac Pro with one bank of RAM won't hit that. It isn't an OS X limitation and OS X's limitation is already busted by current Mac Pro let alone the E5 series (v1 or v2).
But a 32GB cap is going to snare a few. (but apple gets to put off upgrading OS X for another couple iterations )
Probably additional cores in two slots in the line up. I think more folks will not be happy with the GHz tradeoffs. The E5 1620 is what alot of "legacy software, single threaded" folks have been looking for and had to push to the top end of the single package Mac Pro line up to get when really only have entry level money ( hence the moaning and groaning.)
If want to do a mini cluster then Mac Minis are already there. If all doing is x86 core count chasing then just start racking minis. Especially if can swing quad minis through the whole line up on this next v3 (Haswell) update.
If it is a cluster of GPU core counts than doesn't work well, but x86 only really needs to be about $800/node not $2,000/node cost if trying to stick to a limited budget. If Apple can squeeze quads as the default into the mini then it just takes 5 to get to a count of 20 ---> $4,000. Which is much lower than any two 10 core count offering is going to be. That will be about
$2,200+ just for the two CPU packages themselves before Apple mark up and rest of larger infrastructure costs are added in.
But yes a very high speed interconnect is not there ( if need to ship intermediate results off the different nodes in the cluster). And the mega expensive per node software packages don't scale. Neither one of those I don't think Apple looses much sleep over missing out on.
I don't think Apple is trying to say "stack up boxes" to get back to dual Mac Pro at all. That's more of the ultimate modular speculations that any real insightful read on Apple's direction. Either they will ship a dual to participate or drop out of that sub-market. One of the two.
Last edited:
Why would I want to do that, I already have enough clutter on my desk. And the SDI-cards, even more Thunderbolts? The GPU expansion chassis? Blu-ray burner? We already have enough clutters on our desks in this industry, grading monitors, control panels, etc... Why would anyone believe that spreading parts around like a wild medusa is a good thing, when you can have everything in a box? Is that not the point of a workstation?
Again, if the rumors are true: **** apple! ...and their horrible consumer culture made possible in slave camps. **** them. Is it just a ****** corporation like the other ****** corporations? Well, they decide. Our industries saved this company. Jobs loved content developers. **** Apple if they **** us over once more.
Samsung, Linux, HP and Micro**** are waiting.
Last edited:
And one part of this is the cost of manufacturing & stocking 3-4 variants (single - dual CPU, etc)
As has already been debated, I think it varies to the extremes. The notional "College Cluster" of stock Mac Pro is probably going to get superceded by a room full of iMacs on its next upgrade cycle, which takes it out of play. For professionals who actually workload their hardware, I'd say that probably "95%" have contemplated their contingencies for when they expect to hit headroom limits. As a matter of lifecycle example, I can recall that between the time I bought my G5 PowerMac and by the time that it was retired, its internal storage capability inceased by more than tenfold.
Yes, local drives are going to continue to be under pressure from network/cloud storage alternatives, but the Network's I/O performance also needs to be there to make it viable. The Mac Pro (and state of the shelf) is still Gigabit Ethernet (vs 10GbE), which is slower than even the antiquated SATA-1 hard drive interface that's been in the Mac Pro since it was a G5 PPC.
Apple definitely has the data, because every time in OS X that an App crashes and you authorize sendingin their automated bug report, the data sent includes hardware info. The question is if Apple bothers to data-mine the info they already have. To do so looks to be a very trivial task IMO.
-hh
Not really. They seem to run somewhat opposed to one another. Unless "extra graphics cards" means capping out at 2 (which could be since that pragmatically the cap now and two being inserted by default) then dual GPUs doesn't really match up once take Thunderbolt's requirements into account. [ still don't think going to get two standard PCI-e cards if Thunderbolt is primary part of the equation ]
Large memory upgrades and the timing does seem to knock out the E3 though if large means over 32. However, muddled use of "expansion" if talking about swapping 4 2GB DIMMs for 4 8GB DIMMs. It is exactly the same number of DIMMs and hasn't really expanded at all. An standard config E3 would probably ship with 2 DIMMs and 4 slots. Technically that can be expanded. Just saying expanded without a size context doesn't really move and E3 in or out of the picture.
Yeah appears as managed leak qued up manage exceptions. And yet the conversation was placed months ago. If look closely nothing really waves off WWDC unless latch onto E5 has having to be the foundation.
Only if shifting away from WWDC. If they are then also only slightly less probable they are sticking with one box to service both single and dual CPU package setups.
A single individual doesn't comprise a market.
If looked at linked the context is usage coupled to a laptop. There is no desk concept where it is already requested and deployed. Incrementally, adding other uncluttered desks is entirely possible.
Even in the current state the Mac Pro never did go after "have it all in one box" like the HP z820 and Dell 7600 do. Can even trace back to the ancient "can't survive without the 6 PCI slots in Power Mac 9600 " threads. Apple dropped down to 4 slots and the sky didn't fall.
Narcissistic hand waving didn't save the company. Apple stop doing some stupid things. They canceled a completely muddled product line up. They exited some markets altogether (Newton ). Microsoft loaned the a ton of cash to limp by on. Frankly the first major move they made was iMac. Which was the correct one since Mac OS X market needed to be large enough to be a viable target for software developers. A broad spectrum of users buying Macs saved Apple; not some privileged 1% .
That is fundamentally different that wedded to a specific hardware deployment. Jobs directly supervised the transition down to 4 PCI slots. Complete spin doctoring that Jobs was a mega lover of max slots and biggest to do everything boxes.
----------
Can keep retreating into a smaller, more skewed subset that fits but then detach from the larger issue of whether there is a viable market anymore. Toss all of these folks overboard and the pressing question is there enough people left to sell Mac Pros to.
Instead of measuring what is there, it is act to measure what want to find.
It couldn't possibly be actually expanding the number of drives over that lifecycle. That significant amount of that is technically not expansion but upgrades of the same fixed number of drives. Those are two different things.
What I didn't consider is that time capsule and airport extreme etc. all have SKUs too. Turns out they are also in line for MEXXX generation updates. So you know what.... it's probably those
----------
You can't have dual E3s though. Isn't that dealbreaker? 4 cores max
I'm pretty much an average pro user.
No loyal customers did. Loyalty is a two way phenomenon.
Spin doctoring that this is about numbers of slots...
THUNDERBOLT SUCKS!!!! it's only the equivalent of a SINGLE pcie x 4 slot!!
It can handle one or two peripherals at a time well that's it!!!
Thunderbolt 2 isn't much better.
----------
"In numerous thread you and other xMac fans have clamoured that all you need is a headless iMac. Along comes what is largely a headless iMac and all of sudden it is bad."
It's only bad as a Mac Pro, not as a headless iMac! 2 separate needs
Just a note: People seem to forget that this is also about OSX as a viable platform. Supported hardware at both and all ends of the spectrum is important to keep OSX and the Mac a full alternative to Windows. The delay of a new MacPro has already made the Mac platform vulnerable.
Yep. Could be for some.
The only thing that really needs more than PCIe 4x is a GPU.
A lack of internal expansion room is a non-starter. How would you run an external storage array that requires an adapter card? How would you install SAS disks? Graphics cards? Audio/Video capture cards? Not everything can be done with a one-button approach to market compatibility. The only way that would work is if Apple intends to deliver some sort of server platform to compliment a desktop system.
