madan

melgross said:

madan said:

I'm not trying to make it hard on anyone.  But I am trying to clear things up so people know what they're getting into.  Buyers remorse sucks.  It would be a shame to spend 8k on a computer and find out that it competes unfavorably with a 5k iMac Pro.

Except that other in your own mind, it doesn’t.

Maurizio said:

madan said:

You know, come to think of it.  You could get an iMac Pro, with 64 GB of RAM, a base Xeon, more storage and a Vega 56, stone the base Mac Pro over the head and it STILL COMES WITH A 5K LG MONITOR BUILT IN.  How nuts is that?

Sure it doesn't come with support with 12 TB 3 lanes but srsly, you're probably not going to need that.  A base Xeon wouldn't be able to handle that throughput anyways.  So honestly, the iMac Pro is a better deal because in 5-6 years you just buy another iMac Pro and you get a whole new system, PLUS A WHOLE NEW MONITOR, to boot, for the same price.  Like I said, the Mac Pro only makes "sense" once you start cracking the 20,000 USD threshold. Once you start putting in gpus and cpu configurations that can handle the crazy bandwidth and performance than an iMac Pro just can't touch. But the base system? A Vega 56 is 30% faster than the Mac Pro's base Radeon 580.  And the system costs LESS and BRINGS  A MONITOR.

No, i do not think you get what a new Mac Pro is.
I have a Mac Pro 2009 running in my home studio; in 2009, i paid 3000 euros for it; it had Sata 2, USB 2, a few hard disk, and a GT120 graphic card, and an 8 core double cpu running at 2.16 Ghz. The lowest possible end.
Today, it run with nvme SSD, has USB3, an RX580, and two 6 core 3.4Ghz CPU, and it still current.

All this was massively less expensive than buying the 3 iMac that has become obsolete in the same timeframe.

The Mac Pro is a PCI machine; it is evolutive, that is the whole point; of course, the stellar point is when you spend more than 20K$, but it fully make sense in
a context where needs and gains evolve.

But as of today, if i had the kind of needs and money, i wouldn't buy an iMac Pro, i would buy a new Mac Pro, low end, and let it evolve in base of my needs; it would
be massively less expensive, and upgradable to technologies that today not yet exists, like USB4. In ten years from now it would be still useful.
An iMac Pro bought today, before a refresh, will be obsolete in about 3 years, and not upgradable.

Anyway, the point about the Mac Pro, as many poster said here, it is not a machine for gamers, it is not a machine for the masses; it is a machine for those that need it; they will reconise it.

Maurizio

foregoneconclusion said:

madan said: I think the camp that benefits from the Mac Pro is probably smaller than the camps positively and negatively affected by it. 

That's the reality of personal computing today vs. ten years ago. In 2009, Photoshop was viewed as software that needed high-end desktop hardware. In 2019, your new smartphone could run Photoshop. Software hasn't really kept up with the advances in hardware. Applications that used to need top of the line machines have slid down comfortably into the mass consumer area. It's only the heaviest of lifters that need the brute force of a 2019 Mac Pro. Personally, that's worked out quite well for me. I don't have to spend close to $3,000 on a desktop anymore. A low-end standard 5K iMac from 2017 blows past my old 2009 Mac Pro for my own professional work, takes up less space, costs 1/2 as much, and is far quieter and energy efficient.   

dewme said:

madan said:

Unfortunately, the Mac Pro has a distinct issue on its value curve.  It's a horrible value system at its base price, that quickly ramps in value as the price becomes astronomical.

At its 6000 USD base price tag, the computer is a joke.  The base Xeon it has was about 1200 bucks (on release).  It was blessed with 240 dollars of ECC RAM (on release).  It had a nice, airflow-centric case to be sure.  Good cases that are solid steel/aluminum are, often, 200-300 USD.  Even if we counted the Mac Pro's case as a 500 dollar case, and counted its M.2 storage in the default model as 240 dollars, we'd still be sitting at 3000 dollars for the system.  The Radeon 580 is a naught 200 dollar card (even on release).  

That means you're paying effectively ~ 3000 dollars for a power supply and motherboard.  Which is kinda nuts.  I mean the power supply itself is about 200 bucks at most (actually less) and the fans can't be more than 100 bucks.  So you're buying a, albeit ultra bleeding edge, motherboard for 2700 USD, which is highway robbery.

Yes, the special component of the Mac Pro isn't the CPU or the GPU (although the Mac Pro can top out with sky-high Xeons and absolutely monstrous Arcturus-precursor dual Vega 2s), it's the motherboard.  The base system doesn't ship with any of that super hardware though.  Yes, the motherboard accommodates 1.5 TB of ECC RAM.  Yes it has the ability to run almost a dozen bus lanes for TB 3.  Yes, it accommodates both power via the port and via adapter for gpus.  Yes the Pro Vega 2 is a beast of a card, dwarfing the Radeon VII's already ludicrous 16 GB of HBM2.  But you get NONE of that with a 6000 dollar base system.

With a 6000 dollar base system, you get an amazing motherboard, that might never be used.  You get a low-end Xeon that is outperformed by most Core i9s (Xeon reliability is worth 800 dollars?!).  You get a gpu that is budget by today's standards (the MacBook Pro's Vega gpu is about as fast as a 565-570 which itself is only 10-15% slower than the Mac Pro's 580...).  And a bunch of super components like psus and the like that may never be used unless you upgrade them yourself down the line.

You could build a DIY computer with pretty much identical performance for less than 1500 dollars.  No, I'm not kidding.  Sure, it's not upgradeable with ECC RAM. Sure, it doesn't have 12 TB 3 lanes or 10 gigabit ether.  No, it doesn't have a ridiculously overpowered psu for a system that draws under 300 Watts.  But still, you're buying a system with such low specs all those upgradeable touches are pointless unless you spend thousands more upgrading the system anyways. 


Sure, you can get a great high end Xeon and push the RAM to 1.5 TB.  Yes, 2 Pro Vega 2s are absolutely nuts, with a max of 128 GB of HBM2 RAM.  But that system costs 50k.  The base system gets you NOTHING.  And it's 6000 USD.  For workflow alone, a computer 1/4 the price will do the job.  

So yes, the Mac Pro may be a great machine at the high end but anyone that buys it in the low end better not convince themselves they're getting a super computer because it's a budget system, at most and they're paying between 4-10x as much for the privilege of the Apple emblem.

What you're describing is a recurring problem with well-architected products and solutions, i.e., products designed to support specific quality attributes such as modularity, modifiability, upgradability, performance scalability, etc. Everyone wants all of the values that a well-architected product or solution provides, but they don't want to pay for it when the base-level implementation is really a starting point for acquiring the potential value that the product's quality attributes can deliver. But just like potential energy, potential value is not realized until it is exploited to provide a benefit, which in the case of the Mac Pro is when you start exercising the potential by upgrading components, scaling up the performance, adding massive storage, etc. So yeah, you're paying for the architecture at the entry level but if you don't need the architecture or don't plan to exploit its attributes you may end up spending a lot more than you need to.

It usually comes down to making intelligent and informed decisions about what you're buying while taking into consideration the intended lifecycle of the product or solution. Too often people, teams, and organizations will make the wrong decision because they're applying short term considerations to longer term problems. Or vice versa. They'll look at the price of the architected solution, balk at the price in terms of their current budget, and cheap out on the purchase. A year later, or when the regime changes, they'll realize they didn't buy what they really needed for long term value and revisit the whole process and end up spending more in the long run and inciting churn. Of course it works the other way too. It's not an easy decision, but for people and organizations that apply sound economic justification for their purchases, taking into all factors like depreciation and salvage value, it SHOULD be a data-driven decision and not an emotional one. These are exactly the kinds of decisions that organizations make every day around all manner of personnel and capital expenditures from computers to upgrades of production machinery. I imagine many buyers of Mac Pros will apply these same sort of decisions.