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.
David Pogue says new desktops in 2013
- Thread starter macrumorsuser10
- Start date
- Sort by reaction score
Current iMac runs so hot it fries your eggs already.
It doesn't run that hot in terms of internal temperatures. Possibly a little bit hotter than I would like to see on a desktop CPU. Touching the back end (the heat sink) and trying to gauge whether it's running hot or not is just as smart as touching a car's engine to determine whether it's overheating. Call me crazy, but I would use the water and oil temperature gauges.
Yes, that is a substantial part of it too. Parts at the right price points to no change the margins. However, the longer the stretch they run with 3500 series parts, the more that portion of the Mac Pro market will shrink. If it shrinks too small they may decide that just don't need to persue it anymore. Especially, if the iMacs soak up the vast majority of the "slack" there.
The problem is that isn't likely to happen at the desktop level. Sandy Bridge to Ivy Bridge to Haswell the mainstream processors have stopped at 4 cores.
That is largely because they now how to compete with the GPU for additional transistors and Intel is not relegating the GPU to the "extra left over" transistor budget anymore.
" ... Actual graphics performance was perceived as an afterthought, and the unused area (so called white space) in the chipset was available for the GPU, but rarely a millimeter more. The imperative was to provide a free graphics solution, without adding to the cost of the platform. The investment in software and drivers was similarly limited, leading to products that were good for multi-media but wholly inappropriate for 3D graphics and games. ... "
http://www.realworldtech.com/page.cfm?ArticleID=RWT042212225031
That was the old Intel approach. The new one is GPU cores are on more equal footing. With the GPU is going to come increased memory I/O and that also is going to soak up some space to alleviate.
Even without the GPU budget pressures the mainstream market is still hooked into "bigger GHz is better". Yet another reason to avoid bumping the core count higher to clock the capped 4 even higher.
All of that points to the likelihood that even at Broadwell the mainstream line will still be capped at 4 and that other functionality will move on-board the mainstream die. Not just x86 "core count for core count" sake.
I just don't see where the iMac is prune off the 6-8 core folks.
The only chips still on the complex core count race are the E5 and E7. Even the E5 1600 is a good candidate for a smallish GPU that was more finely tuned for OpenCL work than the Ivy GPUs cores are. At Haswell, Intel could merge some GPU into at least the 1600's . The 2600's don't need it.
If Intel goes the route where the 1600's are primarily complex core pumped , but leaves the PCI-e lane count limited then I can see Apple walking away.
The problem in the OS X space is that Apple has been bending over backwards to jumpstart this. Grand Central Dispatch, OpenCL , etc.
That 3-4 years after introduction developers still aren't leveraging this in their software is deeply at odds at where Apple wants to go.
If Apple can't jump start wide spread software development to leverage more than 4 cores then neither the single or dual package E5 has traction.
Either the software starts to scale or whole thing is going to go.
It is also not just a single app issue. I believe leveraging visualization is going to be come more prevalent on the desktop workstation market. If put multiple machines in a single box it is much easier to soak of more than 6 cores if set those machines to working on multiple things are once with deadlines.
The issue is that the iMac could be extended to cover part of the single CPU package space. It is not likely to be extended to cover the previous generations dual one. Let alone the current or future ( 2 * 10 cores ) one.
The only 20+ core in a single package is the MIC offerings.
Only if the software workload doesn't scale. The cost ratio is jacked up because the software (if primarily just running one application) is jacked up.
Yeah software lags behind hardware..... but just how long can that be used as an excuse? That's akin to saying the current Mac Pro bump is OK because lagging is excusable.
the bigger element of the problem is getting affordable > 4 core development platforms to more developers. OpenCL helps with that if some develop more chops using more widely available GPGPU to test out code base on. That can be re-routed to x86 cores if they are a glut around on a high core count box.
For the mainstream consumer parts, true.
But to keep the MP customers by transitioning them to a replacement system, the enthusiast parts would be suitable to meet performance targets. Basically a Xeon minus ECC under the hood.
Quantity pricing isn't any different, but it cuts the DP systems entirely, which reduces inventory and costs if done properly (stock management to simplified assembly processes). It also addresses the IGP performance issues by sticking to a dedicated GPU (embedded or discrete solution).
No, it won't cover DP users, but it doesn't appear that this sub-segment is nearly as substantial as SP users since Nehalem.
I'm seeing this in other spaces as well (Dell, HP,...). DP's still have their use in situations where a cluster is out of the question, but the SP's are capable of a lot more.
As you've mentioned, OSX or otherwise, software is significantly behind. Both in leveraging GPU's for GPGPU processing as well as true n core multi-threading. Software developers have been dragging their heels as they seem to be waiting to see what happens first before committing the funds for development.
I'm not making excuses, just stating that they're cheap and want everything handed to them on a silver platter. Unfortunately, there's nothing to really motivate them to speed up their time-tables it seems.
If this were to happen, then a DP for the slot configuration would make sense, as it could be loaded with GPU's to leverage the software capabilities.
This isn't the case ATM though, which is another reason I've come to the previously stated conclusions.
I could see Apple going to an iMac Pro, lets say it has retina display, an SSD, 1-2 HDD bays, desktop GPUs and allows E3-1200 with ECC Ram.
The thing is that if Apple wants a 27" retina display in an iMac, aren't they going to need better GPUs than these mobile ones anyway? Then wouldn't the heat/size from desktop GPUs require them to increase the thickness anyway? If that's true, they might as well jump to Xeons, potentially E5s that allow for >32 GB of RAM.
Then, if the iMac Pro comes around once 27" retina displays can be made, I could see the Mac Pro being available only in the DP configuration, potentially with a case redesign, which might explain this silly spec bump update at such a late date. I still think its a dumb thing to do and that Apple is going to lose some business because a lot of people are tired of waiting already. New Xeons are out and business needs to get done. Then once people convert to a different OS, it will be difficult to justify to go back to, by then, year old technology in early 2013.
The iMacs are fine with Sandy bridge for now. Ivy Bridge doesn't bring many advantages for the desktops, but the workstations are not OK with westmeres. This is dumb.
It's time apple to give guidance to its most loyal customers
I believe Mac Pro users are some of the most loyal long term Apple customers in the market. This secrecy from the company is now about to greatly effect the pocket books of its highest end customers.
It's time that Tim Cook, or someone stand up and provide some quidence to its customers. If they are building a new Mac Pro for some later date, simply stand up and say it. If the strategy is to move more to thunderbolt and clustered Mac mins, iMacs, MacBook Pros,,,, fine. Just give some guidance. This is not a$700 iPad or $400 iPhone purchase.
It's not just about buying the single computer. When I look at the peripherals I have connected to my system I see over $10,000 sitting under my desk.
I'd not be happy, but would continue to be loyal if apple released these small CPU bumps, a better GPU solution and a thunderbolt card for the Mac Pro telling me they have a new design for the Mac Pro inthe works. OR tell me that architecture is dead and they are moving to clusters of thunderbolt devices.
It is high time Apple provide some guidance to its most loyal customers.
I believe Mac Pro users are some of the most loyal long term Apple customers in the market. This secrecy from the company is now about to greatly effect the pocket books of its highest end customers.
It's time that Tim Cook, or someone stand up and provide some quidence to its customers. If they are building a new Mac Pro for some later date, simply stand up and say it. If the strategy is to move more to thunderbolt and clustered Mac mins, iMacs, MacBook Pros,,,, fine. Just give some guidance. This is not a$700 iPad or $400 iPhone purchase.
It's not just about buying the single computer. When I look at the peripherals I have connected to my system I see over $10,000 sitting under my desk.
I'd not be happy, but would continue to be loyal if apple released these small CPU bumps, a better GPU solution and a thunderbolt card for the Mac Pro telling me they have a new design for the Mac Pro inthe works. OR tell me that architecture is dead and they are moving to clusters of thunderbolt devices.
It is high time Apple provide some guidance to its most loyal customers.
Nope they support 4K output which is the maximum a 27" iMac could have in the next few years.
It is even worse than that. The Retina display actually consume more juice as the backlighting problem gets harder.
If Apple is on a "Retina or nothing" track with the iMac then it too is on a equally doomed path for the next year or so.
The 27" iMac already has 2560 by 1440 pixels . The new MBP hiDPI has 2880 x 1800 . Frankly that's really not all that much of an "advantage" to the 15" screen. Larger pixels are useful in non photo/video work.
If Apple is just trying to double for doubling sake then 5120 x 2880 screen is kind of ridiculous. It would be technologically cool, but I doubt the yields and price at 27" are going to be affordable.
Generic desktops perhaps. For the all in ones though being able to run pedestrian graphics loads off the integrated GPU allows the box to run cooler. Cooler over a longer period should lead to longer lifetime with fewer problems.
The 50W GPU is in the new MBP 15" hiDPI but I bet they turn it off then they can. The iMac is in the same boat, just with a bigger display attached.
Has it occurred to anyone that maybe they simply don't have a plan? This is likely just swept to the side. Maybe they are a bit afraid of the backlash to outright cancel it. What surprises me more is the complacency of the iMac line. To me that is much more drastic a marketing decision, even if it is just delayed. It speaks volumes. And for anyone to be banking on what night be available a year from now computer wise is just ridiculous in my opinion. Anything can happen in a year.
Last edited:
Not having a plan is a plan in itself when it comes to a product that needs constant development by nature. In this case, the message sent is either, "we're neglecting you because you're not that important to our business," or "we plan to discontinue the Mac Pro line, but we'd still like you to buy our unsold inventory."
Like Rush says, "If you choose not to decide, you still have made a choice."
Displays are a weird issue, but the engineering problems with desktop gpus that significantly outpace the imac's top current option is more a problem with surface area and heat than machine thickness. A thicker computer means you can fit bigger fans and potentially dissipate more heat, but it doesn't solve the other problem. Beyond this Apple has been running with their thinner is better thing even where it makes no difference. In the imac it's just a gimmick as you don't actually reclaim that space, and people are not using these things in wall mounted configurations.
I suspect the rumor on the imacs and Ivy Bridge will be disproven. We may very well see new imacs and minis during the summer months. Imacs don't gain a lot in raw cpu power, but the tdp is lower. The figures are quite conservative. They would gain a new gpu generation which IS needed. Having all but the top slot outpaced by the macbook pro updates is bad. USB3 is overdue here.
They do. 4k always sounds a bit misleading, as the actual resolution doesn't quite line up that way.
In photo/video work if you're planning to pan around an OpenGL based application with a true 5120x2880 display, that would be a recipe for major lag. I thought they'd just use a high resolution panel and double everything.
If it's a retina or doom thing, expect them to scale back to a 24" or something around there. Even 30" panels have become less common and relegated to niche products. I'm aware of the 31" versions on the market. They're just not mass market items that users would typically leave on their desks.
The issue of cooler = longer life seems to come up quite a lot. Heat always reveals problems like if a part is headed toward failure. What is your take on some of the portables? Does running them at high temperatures often reveal noticeable problems with manufacturing or lead to lower machine lives? Say one of your machines like a macbook pro is spending too much time in the 170-190F cpu temp range, is it likely to significantly increase the chance of major failure within a few years?
It's a product that has been on life support development for a long time. Drop in new parts. Make firmware tweaks. Ship.