Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Money drove Power PC also. If Apple had paid IBM to make what they wanted, it would have been built. That Sony, Microsoft, and Nintendo did it showed it was really a matter of putting your money on the table. Apple kept their money in their pocket.

The decision was driven by money in that yes, infinitely much money would probably have gotten IBM to pick up the pace on PowerPC development. But there were also technical merits to the switch - fairly substantial ones. I'd argue that the technical merits from a switch from standard-form GPUs to bespoke ones is largely confined to "A nifty case".

If Apple largely designed and built the card that isn't likely going to happen. It is in the same class as the current Mac Pro's daughtercard.

The current Mac Pro was a very different process. Apple evaluated reference design and there was boot issues to tackle. Sharing a heatsink with the rest of the Mac not likely something Apple is going to let some 3rd party vendor (with its own agenda) do.

Having not gotten any closer than a picture of the Mac Pro, I can't say if its replaceable. I'm hoping it's closer to blade-type SSDs in the Macbook Pro - proprietary, but duplicatable, rather than the Mac Pro daughtercard. But its a faint hope, and one I'm certainly not banking on.

Proprietary software is a big driver of that. CUDA arrived first but so did Adobe Flash. It is taking HTML5 a while but the proprietary solution is being pushed out of the over time.

This is new Mac Pro design is one that Apple probably plans to live on over the next 5-7 years. Where things will be 2-3 years from now is a bigger driver of whether it is a good match or not. CUDA is loosing ground. Inertia may keep it around a bit longer in some HPC subsets (among some others), but with 3 of the 3 largest GPU vendors backing OpenCL and only one vendor behind CUDA the long term prospects for CUDA aren't that good. Up in the same category of "Android is never going to catch iPhone" , "Windows on open hardware is never going to catch Mac OS on proprietary hardware " , etc.

We'll see. I think the wind is certainly blowing in that direction, and if I had to write something new I wouldn't write it in CUDA. But for right now, I'm annoyed at the loss of flexibility.

It is a trade-off. I think that is way they are going to try to go "overkill" (relative to the software alignment to the design) so that it will take a while for a broader spectrum of the software to catch up and take advantage.

Some subset of the rapid GPU trading is trying to brute force bad (at least unoptimized ) software to go faster. Throwing hardware at it to flog it into going faster when it is doing distinctly not particularly optimal things.

May not be able to sit on a box infrastructure 5-8 years but won't be every 2 either.

I agree that it seems Apple is trying to way overshoot "What you need now" so that their box isn't obsolete in a year or two, which is probably a good idea. I'm just irritated that we're drifting more toward closed, bespoke boxes rather than away from them. It's a tradeoff I'm generally willing to accept in laptops, but less friendly to in desktops.
 
I am only finding 8gb parts in the new ram speed.

Ram speeds available at different sizes seems to catch up with what is allowed in the boxes pretty quickly.

The max RAM speed supported by Sandy Bridge E5s is 1600, and 16GB sticks of that showed up pretty quick after E5s release, from what I recall. So, I'm betting we see 16GB sticks at 1866 commonly available by the time the new Mac Pro actually rolls out (which I'd guess is atleast a month or two behind E5 v2's release).

32gb for now, maybe 64gb by years end. 128? That would require 32gb ram chips. Good luck.

The 32GB sticks are bit of special case I think, because of the large price jump from 16GB. If you really need a lot of RAM, its a lot easier to just buy a workstation that can fit 2x the DIMMs to acheive the memory needs. Ie. if you really need 512GB of RAM, you can probably spring for a 4 socket system with 32 DIMM slots. Just some quick math 512GB of RAM will cost you about $11K at $700/32GB stick, but at 16GB sticks will get you there for just about $5K. And that $6K is about the difference in price between a RAM-less, but otherwise fully configured, 2 socket vs 4 socket intel system. If you're talking about AMD, the prices shrink quite a lot.
 
Last edited:
I've already posted this once today, but I'll bung it in here for a laugh:

Image

The Mac disconnect with that is where does the person who is interacting with it sit? A personal computer with no people involved in the system's set up.

I know there are people attached to these somewhere else. I don't think that Apple believes though detaching a person from the computer is operating in the area where they have competitive advantage.

If it is a box in the "cloud" providing services..... It isn't going to matter much as much that it is OS X.
 
The Mac disconnect with that is where does the person who is interacting with it sit? A personal computer with no people involved in the system's set up.

I know there are people attached to these somewhere else. I don't think that Apple believes though detaching a person from the computer is operating in the area where they have competitive advantage.

If it is a box in the "cloud" providing services..... It isn't going to matter much as much that it is OS X.

The room next door, it's the machine room for a studio. You don't want noisy fans and nasty heat in there, not to mention using up space for coke and hookers, so you shove them elsewhere.

No cloud, all real-time local hardware. Welcome to music and A/V post production.
 
No cloud, all real-time local hardware. Welcome to music and A/V post production.

The cloud is really expensive for always-on, non-burst workloads. And yeah, you don't want your users sitting in the same room as all that - they'll go deaf and cook.
 
making a difference

I will be buying one if it can take in upgradeable GPUs, even if they have to be "transformed" by Apple. I mean, it's not the best solution, but I've lived those times and it was alright.

However, everything else in the machine, makes me wow. When I was younger I couldn't afford the Cube even though I loved that thing. Now that I can pay for it, I will certainly buy one, even if it's just to support a beauty like this.

But I keep my fingers crossed for 3rd party GPU market, hopefully enough people buy these to at least provide a top consumer GPU each year concurrently with it's PC counterpart.




+1

"Now that I can pay for it, I will certainly buy one, even if it's just to support a beauty like this."

Exactly that !!

But what you pay for is this right here:

http://www.youtube.com/watch?v=PGtP6ZQ6Lt8

I will get a Mac Pro at all cost. I love that thing already... and a nice Phantom Drone from DJI to carry my GoPro Hero3.

:apple:
 
There is no custom hardware needed. It is largely drivers that is missing piece. A Thunderbolt driver for a Tesla Card would solve the CUDA problem for those that can afford one. The external Thunderbolt device would only need to be able to handle the load, but would not techically be custom for that. Any Xeon Phi , FirePro , Tesla card would work with drivers.

But yes getting folks to write IOKit drivers for OS X is going to be a challenge. However, custom hardware is not the hurdle at all to doing local development to bubble up to bigger boxes for production runs.

As long as their is a PCI driver, I don't think you need to write a special Thunderbolt driver. It'll just work.
 
oh wait one more whine.
Positively not rackable.

sure its not rackable, thats why we have the mac mini, anything server class is not going to need 12 cores. but if your building a super computer out of these things,



imagine a room with nothing but clear tubes with mac pros in the center of each of them and wires running down each side of the tube being used for research computers
 
I posted this on another thread but I thought I'd share my thoughts here as well...

So, I've been reading through the forums across multiple discussions of the new Mac Pro (or whatever name you want to give it) and I've yet to form an opinion of my own as to whether or not this machine is a worthy successor to the beautiful silver tower or even if it does, in fact, represent the "wave of the future" or whatever. By itself it is a beautiful, powerful piece of machinery. But is it a Mac Pro? I don't know. I think we should all attempt to reserve judgment until such time as the final product is announced and until we see prices and options. Also, I have heard whispers on the wind that there might be "one more thing."

More to the point of this post: I've seen a lot of numbers thrown around about Thunderbolt speeds and PCIe speeds and so on and so forth and the one thing that keeps ringing clear is that the numbers are a bit fuzzy, especially when math is involved (converting gigabits/s to gigabytes/s, for example). The problem with this is that numbers don't tell the whole story. The system itself is only going to have so many PCI express lanes available, some of which have to service the GPUs and some of which have to service the TB ports and the SSD. Now, looking at the Mac Pro 4,1 and 5,1 the PCIe slots were hardwired to provide two 16x ports and 2 4x ports. So the argument that the new Mac Pro does not have internal slots and so is crippled is actually a little bit specious; like the old MPs you still have 2 hardwired 16x ports serving the first two graphics cards. An old MP with more than 2 graphics cards is going to provide at most PCIe 4x to each of the remaining two cards. Not to mention that without external power support your options for that were limited anyway. And then at 4 PCIe cards you had to go external anyway.

My point is that one can go and look at all the workstations in the world, including the old Mac Pros, and see that no matter how many internal PCIe slots you have you are still only going to have a couple at full 16x and the rest have to share the remaining lanes. That is the same thing the new MacPro is doing... just making those lanes external on the TB bus.
 
+1

"Now that I can pay for it, I will certainly buy one, even if it's just to support a beauty like this."

Exactly that !!

But what you pay for is this right here:

http://www.youtube.com/watch?v=PGtP6ZQ6Lt8

I will get a Mac Pro at all cost. I love that thing already... and a nice Phantom Drone from DJI to carry my GoPro Hero3.

:apple:

We're here to do that, it escapes a lot of people, but that's what some of us were born and exposed to by being around this lifestyle since kids. Apple and Steve have, in many ways, make me want to make a difference and excel where I could.

This Mac Pro is just the tip of a huge iceberg.

Sorry if it sounds like total fanboyism to anyone.
 
Last edited by a moderator:
Positively not rackable.

Give me a drill, a standard rack shelf, some nuts/bolts, a bunch of large-ish diameter pipe clamps, and a bicycle inner tube (to protect the finish of course) and I could probably get 3 of these in a 4U space, and a shallow one at that. I'm not certain I'm in the market for this product (depending on price) but Apple is thinking outside the box (literally) on this - we must do the same.

People are smart - give them design parameters and a goal, and they'll make it work. I bet there will be rackable solutions within a year on the market.
 
sure its not rackable, thats why we have the mac mini, anything server class is not going to need 12 cores. but if your building a super computer out of these things,



imagine a room with nothing but clear tubes with mac pros in the center of each of them and wires running down each side of the tube being used for research computers

Did you see the picture I posted? Being able to rack mount a computer has nothing to do with server class and everything to do with practicability in any number of business applications.
 
No cloud, all real-time local hardware. Welcome to music and A/V post production.

Cloud doesn't necessarily mean internet. The computer on the other end of a long wire.... at that point whatever sends back the right signals with good service can be substituted.
 
Cloud doesn't necessarily mean internet. The computer on the other end of a long wire.... at that point whatever sends back the right signals with good service can be substituted.


Ah okay, my apologies, I was assuming the layman's definition :eek:

In this instance, the whatever needs to be running MacOS ;)
 
As long as their is a PCI driver, I don't think you need to write a special Thunderbolt driver. It'll just work.

No it won't. Standard pragmatic practice is that people ignore the hot-plug option requirements that can be added to drivers. There are lots of assumptions through into the software that the card is always there.

That was part of the Thunderbolt spin and kool-aid. No new drivers required... it looks 100% just like it was on a internal PCI-e switch. Well no when it disappears in the middle of a transaction.

There may be some Tesla card hot plugging. there probably are a few folks who leave their supercomputer on 24/7/365 and expect you to yank out broken parts while still traveling at a 70 mph down the road. But I doubt it. It is probably "shut down that node and reboot it." .
 
All this huge positive changes, but somehow folks still complaining. I am not sure i fully understand this...My only concerns at this point are: price and upgradable parts(and price for).

What don't you get? They bricked their graphics card shortening the lifespan of the computer and pushed all expandability towards the more expensive external add-ons.

But i need to say this: Apple had only 2 choices ahead of them. Either they killed the line, or they do it less upgradable so we buy more often. Otherwise the project will be unprofitable. Just how many of us change the graphics card and postpone a new machine? Let's face it: Apple needs to sell a machine to us every 2 years or less to maintain the line productive. My guess is we will not be able to upgrade much onto the machine, but we will see an amazingly low starting price.... Think of under 2K. I am so excited that i started to save for one! Yuuuuupiiiii! :p

Looks like you do understand what they did though your defense of it is completely inane. I'm not sure how you rationalize the planned obsolescence of a professional product as a good thing for the product line. Certainly no professional would see any value in it. Here's a product that's really powerful and will have half the lifespan of the former product and cost you twice as much to upgrade. Who's ready to sign up?

Don't be surprised when there's not a line for them. The move Apple is making is betting that the professional market that is already pissed after waiting 3yrs for a meaningful update won't just abandon ship now. They're tapping their brand loyalty extra hard here and they will hopefully get burned for doing so. It's completely disrespectful to the professional community that has kept them alive in their dark days.

As I've said before true innovation from Apple would be listening to the needs of this community and actually trying to address them. This new Mac Pro is a complete failure in that regard.
 
WolfPackPrime0 and WolfPackPrime1 = 32 + 32 actual cores of destruction and 128 w/HT.

Wow I'm actually shocked, that board is amazing

Not only are Supermicro motherboards amazing, rendering in Cinema 4d on those particular boards is fantastic - mine get Cinebench 11.5 xCPU scores in excess of 48. You can get the barebones system from here: [ http://www.provantage.com/supermicro-sys-8047r-7rft~7SUPB08K.htm ] for $2228. It includes 2x 1200W PSUs and has 4x PCIe x16 slots and 32 ram slots. Slip a GTX Titan or GTX 690 or two in there and use Octane render and get ready to be blown away. It's like rendering on many workstations at once.
 
Last edited:
High Performance Computing ("HPC")


Maybe, if you exclude sectors such as HPC and the like, and include only things like file-serving in a small business, he might be somewhat correct, but understating the case of many others. My render farm has 166 CPU cores (w/o including Hyper Threading), 9,600+ ATI Stream Processing Units and 39,840 Nvidia CUDA cores and I need more. So would < 12 CPU cores make my day? Not!
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.