The iMac Pro is a marketing machine.

[Chancha]

Maxing out the iMac 2017 with Apple's BTO price is substantially higher than you would with 3rd party components. RAM alone will save you half the cost. And then by the time the iMac Pro becomes available, the price of 2TB SSD out there will also probably go down a bit which further inflates the BTO SSD price (which probably won't change until the next model).

A realistic price difference between an iMac (maxed but at 1TB with 3rd party RAM at 32GB) vs iMac Pro (base model which is set at 1TB and 32GB) is approx. $3300 vs $5000. That's almost 50% higher price for same storage and amount of RAM, but with higher calibre CPU, GPU, RAM speed, and I/O bandwidth. Probably not bad value per se but being an AIO with no 3rd party RAM option available you will likely add to the 5000 price and more, while still suffer the same reliability and longevity issues you get with a regular iMac.

 

[CWallace]

People really are donning the rose-tinted specs over this "modular" and "upgradeable" thing. Read the transcript of that press conference and try and deduce what the Apple folk mean by those words, not what you'd like them to mean. Schiller uses "modular" to distinguish current offerings from laptops and all-in-ones, while Federighi uses "upgrade" to refer to Apple's ability to keep their product up-to-date. The fault they acknowledged with the trashcan is that it is dependent on a three-way distribution of heat between the CPU and two GPUs so even Apple can't offer a version with a single powerful GPU or a better choice of CPUs. There's certainly no mention of user swappable GPUs or PCIe slots. Now, obviously, I'm just speculating like everything else, but I don't see anything in that interview that would make Federighi or Schiller's pants catch fire (at least by executive-speak standards) if the new Mac Pro was no more user-upgradeable than the trashcan.

Apple don't really need to create a Mac Pro that is another "sealed box" designed only to be upgraded by Apple themselves or Apple-certified repair shops - they have that now with the iMac Pro, which I am convinced was meant to replace the current Mac Pro at the top of the Mac line-up and clearly was designed to handle a "hot" CPU and GPU now and in the future.

 

[Chancha]

Apple don't really need to create a Mac Pro that is another "sealed box" designed only to be upgraded by Apple themselves or Apple-certified repair shops - they have that now with the iMac Pro, which I am convinced was meant to replace the current Mac Pro at the top of the Mac line-up and clearly was designed to handle a "hot" CPU and GPU now and in the future.

I kind of buy that conjecture, judging by the tone of the roundtable, and the timing of iMac Pro roll out vs modular Mac Pro's, it seems likely that the iMac Pro was conceived as the sole continuation of the "Pro desktop line" by converging the two previous lines together. The only thing where the iMac Pro is substantially different from trash can form factor is with a monitor permanently attached, and then the trash can in many "pro's" eyes was far from being modular enough anyway. If the roundtable never happened, judging by Apple's lackluster-ness towards the Mac line up, an iMac Pro made perfect sense, it can upsell the maxed out iMac crowd, keeping some of the trash can crowd, all the while streamlining Apple's product lines so they could focus even more on iOS.

But everything has changed. If the modular Mac Pro turns out to be versatile, usable, and of great professional value like the Cheese Grater did, by then the iMac Pro will just become an odd child.

 

[theluggage]

Nonsense.

If you actually need 2TB/64GB on an iMac, odds are that you'd need 2TB/64GB just as much on an iMP.

You won't make an iMac 27 any faster by removing memory and SSD.

...and you won't make it significantly slower, either unless you have a workflow that actually uses 64G RAM or 2TB internal SSD. If you do then trying to run it on a 32GB/1TB iMac Pro will introduce a bottleneck that could easily drag it down to below iMac i7 speeds.

In fact, you might actually need more RAM and SSD on the iMP to keep those 8 cores and super GPU fed with data. As The single-thread performance of the iMP probably won't be spectacularly faster than the i7 iMac. The big performance boosts will rely on those extra 4 cores and GPU capacity being made to earn their keep.

Sorry, but the only "apples vs. apples" price comparison is $3700 for a 32GB/1TB iMac vs. $4999 for a 32GB/1TB iMP.

 

[mathemabeat]

Nonsense.

If you actually need 2TB/64GB on an iMac, odds are that you'd need 2TB/64GB just as much on an iMP.



...and you won't make it significantly slower, either unless you have a workflow that actually uses 64G RAM or 2TB internal SSD. If you do then trying to run it on a 32GB/1TB iMac Pro will introduce a bottleneck that could easily drag it down to below iMac i7 speeds.

In fact, you might actually need more RAM and SSD on the iMP to keep those 8 cores and super GPU fed with data. As The single-thread performance of the iMP probably won't be spectacularly faster than the i7 iMac. The big performance boosts will rely on those extra 4 cores and GPU capacity being made to earn their keep.

Sorry, but the only "apples vs. apples" price comparison is $3700 for a 32GB/1TB iMac vs. $4999 for a 32GB/1TB iMP.

Finally. Someone who gets it.

 

[Miguel Cunha]

I don't care about upgradability, I'll just sell it in a couple of years and get a newer model. It's really not that complicated.

Exactly. Between Craigslist, eBay, and now Facebook Marketplace, selling a well cared for Mac is easier than ever. I don't understand why people feel like they are locked into a machine for all eternity if it doesn't meet their needs in a few years.

Probably because those commodities aren't available in many countries/regions.

 

[PatriotInvasion]

Probably because those commodities aren't available in many countries/regions.

Good point, but if you're in relatively urban or suburban U.S., pretty easy to sell Macs.

 

[joema2]

Nonsense. If you actually need 2TB/64GB on an iMac, odds are that you'd need 2TB/64GB just as much on an iMP...

That is not correct, as can be seen from all the discussion on this forum about how to squeeze out a little more iMac performance. E.g, using 3rd party RAM with faster timing, or SSD plus an external drive vs Fusion Drive, etc. These often have limited real-world benefit but they may improve things some -- so can be worth the cost, complexity or risk. They are attempts to push iMac 27 performance higher.

My documentary film group faces this issue right now due to the huge performance burden of 4k. We can get maxed out 2017 iMac 27s or get a current 12-core nMP (only $6k on the Apple refurbished store) or wait until December and get a few base iMac Pros.

We don't *need* an iMac 27 with 2TB SSD or 64GB RAM but this would improve certain workload performance a bit. Nonetheless an iMac Pro with 1TB SSD or 32GB RAM would likely be much faster -- despite the lower memory and SSD. Our main storage is on external Thunderbolt arrays but 2TB SSD would enable an iMac to hold a few more things there, so this would help some. But that would be a desperate "back door" effort to wring more performance out of the iMac -- it's not needed from a storage workflow standpoint, so it wouldn't be needed on an iMac Pro.

Likewise the idea that 64GB RAM on an iMac 27 mandates that on an iMac Pro is not correct, and that's not how most software works. There are diminishing benefits to increasing RAM but it generally helps some. It is related to cache hit ratio and locality of reference of the main workflow tasks. It is not a rigid, black-and-white thing. This can be seen from this study of performance benefits of various Mac RAM configurations: http://www.macworld.com/article/203...ifications-of-additional-memory-on-a-mac.html

It can also be seen from Intel's L3 cache design for Skylake X, which reduced cache size by 1/2 yet provides better overall performance: http://techreport.com/review/31986/intel-core-x-series-cpus-and-x299-platform-revealed/2

In a CPU-bound workload you can easily tolerate modestly lower memory or CPU cache hit ratios if the payback is higher CPU performance due to other optimizations. At a fixed level of optimization, higher cache sizes usually helps performance but the benefit diminishes with increasing size.

The serious professional work like ours, the urgent challenge is how high on the performance scale an iMac 27 can be pushed before upgrading to an iMac Pro or Mac Pro. The iMac can be pushed higher but at a significant dollar cost. The amazing thing is a top-spec BTO iMac 27 overlaps with the cost of an entry-level iMac Pro yet is not nearly as fast.

Unlike the title of this thread, in the real world the iMac Pro is not a "marketing machine" but it is badly needed for professional work. Trying to push the regular iMac that high only results in a very expensive iMac that is far slower than the base iMac Pro. Pulling SSD and RAM out of the iMac will make it less expensive but also slower, when the challenge for this work category is to go faster.

 

[Chupa Chupa]

The iMac Pro isn't mean't for anyone specifically, it's a tool for marketing and reignite enthusiasm back to the Mac (and iMac specifically) until they come up with an all new design.

Agree it's a marketing tool, and nothing wrong with that. But heartily disagree it does much to reignite enthusiasm -- at least mine. For that Apple should have announced a Mac mini Pro with user upgradable RAM and storage + eGPU support and a wide selection of processors and GPU CTO to satisfy everyone from enthusiasts to giant post production and photography companies. Pricing starting at $1399 -- essentially headless entry level 27" iMac and going up to iMac Pro specs. THAT would excite me. iMac Pro is too niche to move me.

 

[theluggage]

That is not correct, as can be seen from all the discussion on this forum about how to squeeze out a little more iMac performance. E.g, using 3rd party RAM with faster timing, or SSD plus an external drive vs Fusion Drive, etc.

...but we're not talking about faster 3rd Party RAM or SSD vs. Fusion. We're talking about Apple's eye-wateringly overpriced 32 to 64GB RAM upgrade and equally expensive (but maybe less overpriced for high-speed PCIe SSD kit) 1GB SSD to 2GB SSD upgrade, which is the only way to push the iMac price up to iMac Pro levels.

There are diminishing benefits to increasing RAM but it generally helps some.

The important thing about diminishing returns is to know when to stop throwing money at them. The 64GB and 2TB upgrades on the iMac may shave a few seconds off your time, but are so expensive that they're unlikely to "pay their way" unless your workflow is genuinely RAM or disc-limited... and in that case, I'm sorry, but a 32GB/1TB iMP will probably be RAM- or disc- limited too.

This can be seen from this study of performance benefits of various Mac RAM configurations: http://www.macworld.com/article/203...ifications-of-additional-memory-on-a-mac.html

All that article proves is you'd be bonkers to try to Photoshop large files - or allocate 4GB of RAM to a VM - on a MacBook with 4GB and a spinning HD (of the none-too-fast variety that Apple like to fit). The latter will absolutely guarantee that things will grind to a halt as soon as you start swapping to disk - something that SSDs help with enormously, not because of their size or sustained transfer speed, but because any SSD has order-of-magnitude faster seek times than spinning rust.

The closest test they did to your 4k editing was the Handbrake test - which was barely affected by quadrupling the RAM. Probably means that it was limited by that slow HD. Now, I know that Handbrake can happily max out all virtual cores on an i7 so that's one thing that the iMP's 8 real cores should eat for breakfast - if your disc storage can keep up.

Our main storage is on external Thunderbolt arrays but 2TB SSD would enable an iMac to hold a few more things there, so this would help some.

If your workflow is really being throttled by external HD storage or lack or RAM on an iMac it will still be throttled by external storage on an iMac Pro and those extra cores will go unused. It might be that improving your external storage is the key to improving your workload...

The amazing thing is a top-spec BTO iMac 27 overlaps with the cost of an entry-level iMac Pro yet is not nearly as fast.

...as I said, be very careful assuming that the per-core performance of the iMP will be night-and-day better than an i7. Some of that premium will be for stability and extra PCIe lanes and the rest depends on making efficient use of the extra cores and GPU features.

We really know squat about the iMP's performance at the moment. Probably there will be some jobs where a 32G/1T iMP leaves a 64G/2T iMac in the dirt, and others where the extra RAM and/or SSD turns the tables. Until we start to see some actual tests on the iMP its anybody's guess and the sensible comparison is to compare the same RAM/SSD configurations.

 

[askunk]

Regarding the price...

 

[fig]

Interesting to hear the variety of opinions here.

For 96% of graphic designers and editors the iMac Pro will be massive overkill, you're just fine with a nicely speced iMac.

In chatting with my pro friends who actually do have the need for a high powered machine, the iMac Pro won't have the graphic card options they want and it'll be really pricey to spec out. They'll go PC.

It's beautiful, I'm just not sure it's got much of a market.

 

[dawnrazor]

what's the problem here? If you're not in the market for an iMac Pro... don't buy one... the last upgradable Mac was made in 2012... pinning after an upgradable Mac of any flavour is an exercise in futility...

 

[Miguel Cunha]

Good point, but if you're in relatively urban or suburban U.S., pretty easy to sell Macs.

I'm sure it is.
Unfortunately, that's not my case.
Otherwise I would save tons of money and switch my machine more often.

 

[Georgio]

Bottom line on the iMac Pro is that if you want it and have the money spare you'll buy it. I can see a lot of design companies buying these partly out of prestige and partly because 3-5 years down the line they'll still a get good trade in whereas PC's are just worthless.
The key issue will be heat management, if they can be designed to run quietly & relatively cool without sounding like a jet on take off then they'll be a success.

 

[Joe The Dragon]

Regarding the price...

that board can do dual cpu! also do an AMD build

 

[Lesser Evets]

The iMac Pro isn't mean't for anyone specifically, it's a tool for marketing and reignite enthusiasm back to the Mac ... until they come up with an all new design.

INCORRECT...AND CORRECT!

"ISN'T MEANT FOR ANYONE" is a heavy super-duper-overexaggeration ranking 10.1 on the Richter scale of butthurt-hata discontent. So let's just label your venom for what it is.

On the other hand, you could be hitting a bullseye by toning down that remark and claim the iMac is "crippled niche". The second part of that quote is right-on-the-mark--the iMac is a palliative bit of PR to bridge a dangerous span Apple's highest managing and marketing allowed to open up and ruin their reputation as a computer company for the creme-de-artiste (and richie-rich geeks).

There is no doubt that the iMac Pro will indeed blow away all other Macs coming before... mostly due to the processors with their endless cores churning out data 10x faster than garden-variety computers of last year. These iMacs aren't going to be made for "no one in particular", but rather for serious video/cgi workers and video game goofs that want Macs.

If Apple gets their act together, there might be a serious MacPro late next year; that computer will probably be the antidote to your venomous exaggeration. Maybe.

 

[fig]

Regarding the price...

You think the fact you've included a $1700+ video card in that build might have a have a bit to do with that price?

Also worth noting the fact that the iMac won't have the ability to run Nvidia GPUs is probably it's primary issue.

 

[guzhogi]

I wonder what a maxed out iMac Pro would cost? I also wonder how it'll work with distributed computing apps like Folding@Home

 

[bplein]

I wonder what a maxed out iMac Pro would cost? I also wonder how it'll work with distributed computing apps like Folding@Home

At whatever rate the final Xeon CPU is capable of. There's nothing special about Macs for this type of work, and in the case of Folding@Home, there isn't even GPU support on Mac. So it's actually one of the most expensive ways to participate.

 

[askunk]

You think the fact you've included a $1700+ video card in that build might have a have a bit to do with that price?

I didn't include anything The image was posted on another website and I found it interesting. Of course, there are many differences at the hardware level, let alone the fact that the iMac will be stuck with it.

Vega FE costs around 1200 USD, so there is still room to say that at least Apple is not ripping people off with a huge markup on the parts. I think somebody did the same job with the tcMP, showing that - at the end of the day - the cost was pretty much worth the effort, at least when it comes to pure "present" value of the parts.

Peace

 

[fig]

I didn't include anything The image was posted on another website and I found it interesting. Of course, there are many differences at the hardware level, let alone the fact that the iMac will be stuck with it.

Vega FE costs around 1200 USD, so there is still room to say that at least Apple is not ripping people off with a huge markup on the parts. I think somebody did the same job with the tcMP, showing that - at the end of the day - the cost was pretty much worth the effort, at least when it comes to pure "present" value of the parts.

Peace

Apologies, thought that was yours

I don't think the cost of the iMac Pro is absurdly high for what you get, but I also don't think there's a huge market for it. Missing on the video card was pretty critical if they're trying to go after high end 3D artists.

 

[theluggage]

Regarding the price...

I don't think the iMP will be bad value when you compare it like-for-like with other Xeon/ECC machines, with OpenCL or CUDA-optimised workstation GPUs. Especially when you include a 5k display in the spec - at a time when there are only a couple of hideously expensive options (I suspect that the PC world is waiting for DisplayPort 1.4 capable 5k displays that can drive 5k over a single wires - people with PCIe GPUs don't want to connect their display via Thunderbolt).

In fact the Mac Pros have (at release - maybe not after they've gone 3 years without an update) often offered reasonable bang-per-buck when compared like-for-like with Xeon/ECC/FireGL/Quadro PCs.

The question is whether people who actually need that horsepower will accept it in a sealed unit with no choice of displays (the 5k may be nice, but what if you prefer a pair of 30" or larger 4k screens?)

 

[Coyote2006]

The iMacPro looks cool and seem to be an extremely powerful machine. But I still have to wait for the new MacPro to compare.

I never liked the iMac concept as a professional customer. I wan't to upgrade my Mac and choose my own monitor.

It's rediculus to pay $6000 (or even more) for a computer for which you can't upgrade the GPU or CPU. Apple has still not understood the pro market.

 

[askunk]

As much as I agree with the first part of your statement, I don't reach the same conclusion. Perhaps you don't know the full story. Apple itself admitted that there is a portion of the Pro market which finds an AIO suitable for their workflow. After all, there are people happy about the tcMP. You are just the other portion which will be addressed by the mMP, like many of us.

Think about it. At the end of the day, you are at the border between the two groups, since you stated you want to compare the machines first, before buying one of them.

Finally: 4.999$, not 6k. And the value of the parts is quite worth it, although I do agree that it doesn't suit the needs of some professionals.