Real pros don’t give a crap about gaming…unless you are developing games, in that case this machine is up to the task
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.
Adobe and Other Developers Are 'Incredibly Excited' About New Mac Pro
- Thread starter MacRumors
- Start date
- Sort by reaction score
Oh maybe now Adobe can actually develop software that doesn’t stutter and run like **** on macOS.
Maybe... I would much rather get a 32 core/64 thread thread-ripper for cheaper than the base price though. Apple’s hardware continues to be good quality, but the pricing just doesn’t make sense to me anymore and I feel like MacOS has really stagnated/regressed in recent years. Perhaps it still makes sense for graphic designers but as a data scientist, I’ve slowly been moving off of Apple and onto System76. It’s still more expensive than if you build the systems yourself, and I’ll admit the laptop build quality is not up to the MacBook. However, the customizability of Linux and the higher performance is a way better fit for scientific computing in my opinion. You should try Ubuntu or PopOS if you haven’t in a while. It really is quite impressive.
I payed $4,000 for my 2008 Mac Pro and LED Cinema Display. It has lasted me 11 years, and blows my mind when I think about it like that. My wife has spent more money on laptops over the same time frame.
Apple forgot that the point of having an upgradable computer is to let consumers upgrade it. This is unfortunately not a prosumer device.
While I was able to upgrade with SSDs, RAM, USB cards, and hard drives over the years, this machine kind of starts you off maxed out with nowhere to go.
I know this computer would last me another 12 years, but I also know I don’t need all that power right now. I’m barely working with 4K footage. I just want a Mac that I can upgrade in 4 years rather than need to replace outright.
The prices are not out of line for me at all. If you need this kind of power, this is what it costs. I had no issue buying a maxed-out iMac Pro when it came out and it has served me well.
so ,to have a balanced stand should be priced $999 ???? really ?
I think Mac Pro's target are either creative types (video, audio, photography) or medical or research.
I am a software developer and for me, all Apple needs to do to complete the ideal setup are:
- MacBook Pro: dependable keyboard, good thermal design (I don't really need discrete GPU), lots of RAM (ideally user upgradable), and good screen. Bonus point for optional cellular.
- Display: I don't need ultra fancy 6K3K display. Apple designed 5K display would be totally fine, ideally with HDR.
- Keyboard and trackpad: Touch ID on the keyboard. Switch to USB-C or Qi wireless for charging.
- Mac mini: 2018 Mac mini is pretty good, but just make the RAM slots more easily user serviceable and fix the wireless issues.
My point is that they could have a lower spec entry machine at lower price. Maybe 4 core, 512GB and 32RAM.
The entry level price point is too high. While it might be in par with the specs (taking out the ridiculous 32GB Ram and 256GB storage) it just won't reach the mass appeal that could generate with a lots of Pros that do not want to invest so high on a Mac that offers expandability.
Not true. For those of us doing GPU scientific computing (e.g., "big data" statistics and machine learning), Apple's new offering does not provide a "yes" answer to "IS the problem solved." CUDA is currently the leading game in town for solving these problems (really, the only game for some important algorithms), and by failing to support NVIDIA, Apple is effectively abandoning an important segment of the academic and industrial market. Moreover, many of us already working on these problems already have NVIDIA cards we'd like to use in a new Mac Pro (speaking for myself, I have a $1k card and two $2.5k Tesla cards). Even if there were an AMD solution for our problems, someone like me must effectively throw away the cost of the new Mac Pro in order to move to it. (I currently use Mac for much of my non-computing work, but for GPU computing I have to deploy to a Linux tower with NVIDIA GPUs, a very awkward and underproductive way to have to work, but forced by Apple and NVIDIA refusing to come to terms and serve their customers.)
The GPU card is a tool, and for a pro machine, Apple needs to let developers choose their preferred tools as much as is feasible. Supporting NVIDIA cards is technically feasible; it's only lack of will that's preventing it.
FYI, for background on the underlying issues (which are frustratingly political and vague), this is a good read: https://appleinsider.com/articles/1...in-macos-and-thats-a-bad-sign-for-the-mac-pro
Last edited:
The new Mac Pro might be useful for audio/video editing, but it's an expensive doorstop for those of us working in AI/deep learning. The only usable GPUs for deep learning are from Nvidia at this time. I just built my own awesome Linux workstation for < $10k, including NVIDIA Titan RTX GPUs.
Apple cutting support for both CUDA and OpenCL meant that I switched immediately from Apple to Linux, and the new Mac Pro doesn't solve the issue; I can load it up with 4 GPUs but they are completely useless for my code. OS X is now crippled by requiring use of its proprietary Metal framework that simply doesn't work for state-of-the-art machine learning.
[doublepost=1559610054][/doublepost]
This may be true for the use case of "processing multiple 8K video streams in real time", but that's just one problem. There's another large pro market of people who do machine learning work. Nvidia is the only commercially-available GPU viable for a large class of machine learning problems; it has hardware features unmatched elsewhere, along with software libraries (PyTorch and Tensorflow) that currently only work on Nvidia.
The GPUs in the Mac Pro simply can't run my machine learning code.
Video-processing may work across a variety of GPU manufacturers, but for better or worse Nvidia is the only game in town for deep learning, so Apple has decided that it doesn't one any of us in the deep learning field to be its customers.
I switched to Mac around 10 years ago due to its developer/research-friendly OS. I stopped using Mac last year when Apple became hostile to those of us in the scientific/AI community. (And as an aside, they also dropped "Pro" support from their laptop line by introducing a broken keyboard that is also missing the "escape" key that I use so much while coding. So I can't even use the "pro" laptop easily as an SSH dumb terminal.)
The new Mac Pro is completely useless to me for my work.
Last edited:
I’ll wait until I can see if it can compete at all with cuDNN and with the mixed-precision/quantized training I can do on V100 cards. I don’t know if AMD even has support for these sorts of things in their hardware.
Also, pyTorch support is more important to me.
Dude, they’re not cheap by any means. Of course you’re right in that we can’t get a 1:1 comparison because we’re dealing with two different operating systems. But internals alone, you can get similar for less.
The problem is their entry price is so high. Other workstation manufacturers have a much lower entry point. Yes, the hardware is specced lower at that level but on anecdotal experience alone, there are a ton of professional users where those PC entry level workstations are more than enough.
Oh god no. The baseline configuration specs are embarrassing (as they are on most macs). With some upgrades the hardware shines, but we have no idea what those desirable upgrades will cost. And there is ample reason to think they will be outrageously priced, even for pro hardware. Today Apple introduced a $1000 "pro" stand...
g\
Well for 6K you can have a 32 core beast with 128MB RAM, 2TB Nvme, RTX 2080Ti and room for 3 more GPUs.
https://pcpartpicker.com/list/FP2ztg
https://pcpartpicker.com/list/FP2ztg
I see two sata ports on the motherboard picture, but don't see it mentioned in the tech specs. Maybe there will be space for off the shelf 2.5" ssd.
Look, prices are due to steel and aluminum tarriffs. And forced US manufacturing.
I am really liking the WTC cage design hiding... the cheese grating. Very elegant. Like a cross between the original WTC stainless steel cage and a flamingo's feather patterns. The feet bugged me at first, but the clunky look is reminiscent of bird feet I guess. Not as practical as the original chesse grater which you could hang off the end of a desk if you needed to (like with the most used ubiquitous $499 music production desk). They never consult me on these things. Or the need to handle Nvidia cards.
Can't wait to win the lottery so I can upgrade. And get a bigger desk.
I am really liking the WTC cage design hiding... the cheese grating. Very elegant. Like a cross between the original WTC stainless steel cage and a flamingo's feather patterns. The feet bugged me at first, but the clunky look is reminiscent of bird feet I guess. Not as practical as the original chesse grater which you could hang off the end of a desk if you needed to (like with the most used ubiquitous $499 music production desk). They never consult me on these things. Or the need to handle Nvidia cards.
Can't wait to win the lottery so I can upgrade. And get a bigger desk.
Here’s the picture that I’m referring to. The SSD blades look shorter than standard M.2 NVMe blades, which is all I’m personally familiar with.
Last edited:
Fair enough. From an expandability standpoint I can certainly appreciate your view. I must admit, I'm a little close-minded because all I care about is the raw power and higher-end specs to cover the work I do. I know there are great lower cost options in the form of the Mini and the iMac, but obviously expandability starts to become an issue there.
Hey, since you were so cool and stuff, check this out: an internal HD RAID MPX module being made for the new Mac Pro. A solution for hard drive peeps!
https://www.theverge.com/circuitbre...tb-pegasus-mpx-module-storage-hard-drive-raid