Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I work in the music business and spend a lot of my time in tracking and mixing studios, sometimes crossing over into the area of film scoring and dubbing (which is by far the most lucrative part of the industry).

Bar one studio which seems to be talking about buying a new Mac Pro in order to keep a client happy (at a loss), everybody has started looking towards Hackintoshes. Most of the film mixing studios have already made the move. Even the businesses which have the money to pay for one of these machines (which is virtually none) are turning their backs.

All you people who say ‘Pros will see it as an investment’ are also forgetting that the ~£10k price difference between a well specced out MacPro vs a Hackintosh will also buy you a very nice set of monitors.

But if they are truly using those machines for work purposes, what happens if there's an update that breaks the Hackintosh OS and they have to get work done?
 
  • Like
Reactions: 724699
Understandable.
No problem. 👍🏻

I’m not even defending Mac Pro (some aspects are completely wrong, like a 256 Gb SSD on base configuration or an 8 core base CPU).

the base units are a joke and it's clear Apple is attempting to use some sort of upsell here.

at the base / bottom 2 tiers of configuration, the price point is very poor value. I attribute this to an over-engineered design and Apple padding the higher volume sku's with massive margins to cover those costs/

As you get up in the sku's, the higher end ones, the Mac Pro (still with Apple tax) is a far better value for the money.

As I said, this indicates that they've over-engineered the thing again. They've tried really hard to differentiate themselves, while using "standard tech" but making it different enough. but in doing so, have driven costs up.

the new MOBO for example is really good, but it doesn't fundamentally offer anything new to the desktop tower solution. Other than allowing Apple to provide proprietary modules in addition to standard PCI-E modules. Proprietary modules that don't need to be proprietary.

The new Mac Pro is a classic definition of a parity product trying to "reinvent the wheel" just to stand out, but not truly offering anything that is uniquely new or better. Just different. essentially "change for change sake"


its still a great machine. It's still the Mac Pro that the 2013 trash can should have been. but it still has that Apple hubris of "we're better just because we're apple" attached.
 
  • Like
Reactions: IG88
My $700 5,1 Geekbench multi core score is 6429. Entry level Mac Pro benches at 7644. Love the build quality of the new machines but wtf Apple?
 
Multi-core CPU performance is ~10% better than a $3,000 MacBook Pro. Price difference: 100%.

The base model is such terrible value. Probably Apple's lowest value-for-money Mac ever.

If you’re only using the CPU, and you’re not using the advanced features of the Xeon CPU, and you don’t care about expandability, you’re probably right!

Apple’s not using quite standard Xeon chips (the Mac Pro 8-core has 24.5 MB cache, while the W-3223 has 16.5 MB), but you can compare Intel processors pretty easily with ark.intel.com. So looking at the i9-9900K (top of the line iMac) vs. Xeon W-3223 (entry-level Mac Pro), there are some interesting differences.

* The iMac CPU has a 20% higher clock rate for single-thread performance (5.0 GHz vs. 4.0 GHz).
* The Mac Pro CPU has 50% more cache (24.5 MB vs. 16 MB).
* The Mac Pro CPU has TDP of 160 W vs. 95 W. This is rather fuzzy, but basically leads to less throttling under heavy load.
* The Mac Pro CPU supports 8x more memory (1 TB vs. 128 GB).
* The Mac Pro CPU supports 3x more memory channels (6 vs. 2). The bandwidth isn’t given, but should be at least >2x, too.
* The Mac Pro CPU supports 4x more PCI (64 lanes vs. 16).
* The Mac Pro CPU supports AVX-512 (2–4x throughput vs. the iMac for vectorized code).
* The Mac Pro CPU supports new instructions targeted towards deep learning (AVX-512 INT8 instructions).
* The Mac Pro CPU supports advanced error detection & recovery. (This isn’t just ECC; but I don’t know how much OS X implements.)

If you don’t run software which needs lots of memory bandwidth or use PCI cards, and you don’t run vectorized code or your applications aren’t optimized for newer Intel CPUs, the Mac Pro is Not For You. (Well, maybe if you can effectively use 16+ cores.)

There’s not a lot of systems with the W-3223 processor yet, since it’s quite new, but quickly configuring an HP Z6 G4 workstation with that processor, 32 GB of RAM, and dual 10 Gb Ethernet brought it to $4293, though the HP has only 2 TB ports, 2 USB-C ports, and no WiFi. Apple’s pricing isn’t unreasonable in comparison. (Or you could argue that HP’s is too high.)

Oh — someone suggested that 256 GB SSD is too small for a base configuration. Not if you don’t want to pay for Apple’s SSD! There are tons of PCI slots, you can buy your own and install it. For instance, you could install a 12.8 TB Samsung PM1725b and have 6300 MB/s sequential reads for $4K or so.
[automerge]1576795064[/automerge]
My $700 5,1 Geekbench multi core score is 6429. Entry level Mac Pro benches at 7644. Love the build quality of the new machines but wtf Apple?

Primate Labs actually describes the list of CPU workloads they run at https://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf . You can decide for yourself how representative they are of your work. Note that Geekbench runs quickly even on older systems, so the individual tasks are not very large.

Don’t get me wrong — it’s certainly useful to have benchmarks like this which give you a picture of how basic desktop applications might run. But they aren’t representative of what most users actually use high-end systems for.
 
Last edited:
  • Like
Reactions: Digital Skunk
Actually, if you compare the Mac Pro to other workstations, it's not really that expensive.

Yeah I think that's the same video I found last week. Had to go all the way to Puget before a workstation got significantly less expensive than Apple.

I'm willing to bet that Apple is already testing an A series version of the Mac Pro. The numbers Intel is getting out of their chips seem to be minor advancements, at least compared to how quickly Apple has progressed with their A series chip and more recently their GPU's. And you know that Apple hates having to rely on a company like Intel.
Economies of scale is not there. When you sell 100M iPhones per year, you reap the benefits of your R&D investment on your custom CPU 100M times. Especially when all 100M of them have identical CPU.

They won't sell enough Mac Pros (considering $6000 base price) to recoup the investment on an Apple designed CPU / motherboard / chipset. Now think about having multiple SKUs (8 core, 12 core, 16 core, 24 core, 28 core) and the idea falls on its face even harder.

This machine is not for you!
That take has been pretty much dismantled by this point. Several professionals have already spoken up and said Nope.
 
  • Like
Reactions: ssgbryan
That take has been pretty much dismantled by this point. Several professionals have already spoken up and said Nope.

Hunh? Just because a few "pros" say they won't buy it doesn't negate that it isn't a machine for everyone, nor does it imply it's not the *right* machine for others.
 
I don't see anything special regarding Mac Pro's sustained performance. This has been something normal in the PC land for as long as I can remember.
Agreed. My 9700K will run Handbrake, all 8 cores @4.8 GHz, for all eternity under 65 DEGC. Apple's 8 core Xeon is nothing special in that regard.
[automerge]1576810666[/automerge]
Hunh? Just because a few "pros" say they won't buy it doesn't negate that it isn't a machine for everyone, nor does it imply it's not the *right* machine for others.
Anyone that complains about price gets the "IT's NOT FOR YOU!!!!!" treatment. Yes, that nonsense has been debunked many times over and it needs to STOP.
 
  • Like
Reactions: SeattleMoose
The point you're making is irrelevant.

Ditto.

You're making an entire subthread about how workstation parts aren't really worth the significant markup. Which is a valid point of view. But it's irrelevant in a thread about the Mac Pro, because everything about the Mac Pro screams high-end workstation parts. It's not particularly fast, not particularly cheap, not even particularly well-specced at the baseline. It's just meant to be a reliable workstation that also can be specced all the way up to the high end. Discussing that consumer parts are a lot cheaper, that AMD Ryzen is sometimes faster, etc. is all moot.
[automerge]1576835056[/automerge]
Anyone that complains about price gets the "IT's NOT FOR YOU!!!!!" treatment. Yes, that nonsense has been debunked many times over and it needs to STOP.

Debunked?

Apple isn't gonna make the Mac Pro significantly cheaper and change the parts from Intel Xeon to AMD Epyc (much less Ryzen) just because a lot of people would like a $2k tower. The Mac Pro isn't that, never will be that, and hasn't been that in a long, long time (and when it was, people were whining that it wasn't a $1k tower).
[automerge]1576835309[/automerge]
If you’re only using the CPU, and you’re not using the advanced features of the Xeon CPU, and you don’t care about expandability, you’re probably right!

Apple’s not using quite standard Xeon chips (the Mac Pro 8-core has 24.5 MB cache, while the W-3223 has 16.5 MB), but you can compare Intel processors pretty easily with ark.intel.com. So looking at the i9-9900K (top of the line iMac) vs. Xeon W-3223 (entry-level Mac Pro), there are some interesting differences.

They're standard chips. The Xeon W-3223 has 1 MiB level-1 cache, 8 MiB level-2 cache and 16.5 MiB level-3 cache. Intel only counts level 3 on Ark, and Apple counts level 3 and level 2.
 
Last edited:
I'm not going to read ten pages of trenchant analysis but it's pretty obvious the issue isn't Apple or Intel here: It's the developers of Geekbench altering their number scaling to provide more easily parsable results for people who can't deal with large numbers with higher end machines.

I bought and paid for copies of Geekbench 3 and 4 back in the day too and the same iMac 15,1 model I own iMac benches 4928 single core and 16819 roughly on those versions.

With version 5 everything has changed: single core score is 1063 and multicore 3796.
Open CL bench was formerly 80,000 and is now 23,000.

Their marketing blurb indicates that "Geekbench 5 to include more ambitious benchmark tests with larger data sets and longer running times." So basically you need a Mac Pro to get respectable results now?

This kind of reminds me of when the developers of X-Plane created a graphics system that no current
computer could run on maximum settings until GPUs had improved 18-24 months later.

I believe the perception that these processors are underwhelming can be laid squarely at the feet of the
devs of this "meaningless indication of processor speed" app.

The scaling has changed:
Why change the scaling just to accommodate people at the top of of the scale whose tiny brains can't deal with six or seven digit numbers? Was this done to try and simplify and keep things to a scale from zero to
100,000 or what?

They should have been more transparent In doing so you totally changed the scale of performance for older system to accommodate new ones and your new tests provide the appearance of far worse performance for new systems if you were used to their old system.
 
Last edited:
I'm not going to read ten pages of trenchant analysis but it's pretty obvious the issue isn't Apple or Intel here: It's the developers of Geekbench altering their number scaling to provide more easily parsable results for people who can't deal with large numbers with higher end machines.

I bought and paid for copies of Geekbench 3 and 4 back in the day too and the same iMac 15,1 model I own iMac benches 4928 single core and 16819 roughly on those versions.

With version 5 everything has changed: single core score is 1063 and multicore 3796.
Open CL bench was formerly 80,000 and is now 23,000.

Their marketing blurb indicates that "Geekbench 5 to include more ambitious benchmark tests with larger data sets and longer running times." So basically you need a Mac Pro to get respectable results now?

Err.


If you are comparing your results in Geekbench 5 to results from Geekbench 4, please note that the updates in Geekbench 5 and the change to its baseline score make it impossible to compare scores between the major versions. A score of 1000 in Geekbench 5 and a score of 1000 in Geekbench 4 do not represent the same performance.

They changed the baseline, is all.

This kind of reminds me of when the developers of X-Plane created a graphics system that no current
computer could run on maximum settings until GPUs had improved 18-24 months later.

Except that X-Plane isn't a benchmark?

I believe the perception that these processors are underwhelming can be laid squarely at the feet of the
devs of this "meaningless indication of processor speed" app.

No. The low end of the Xeon W-3200 series just isn't that great at performance. It's easily beaten by the high end of Coffee Lake-SR. However, at the higher end, you get quite a few more cores, if you need that.

I don't understand the confusion here. It's a benchmark. Numbers are relative measures, not something absolute like "instructions per second", etc.

The scaling has changed:
Why change the scaling just to accommodate people at the top of of the scale whose tiny brains can't deal with six or seven digit numbers? Was this done to try and simplify and keep things to a scale from zero to
100,000 or what?



They should have been more transparent aIn doing so you totally changed the scale of performance for older system to accommodate new ones and your new tests provide the appearance of far worse performance for new systems if you were used to their old system.

They are plenty transparent. Just look at their browser, and you'll see that Geekbench 5 and 4 results are separated from each other for this very reason. (Charts, OTOH, are all 5 now.)
 
  • Like
Reactions: Digital Skunk
Debunked?

Apple isn't gonna make the Mac Pro significantly cheaper and change the parts from Intel Xeon to AMD Epyc (much less Ryzen) just because a lot of people would like a $2k tower. The Mac Pro isn't that, never will be that, and hasn't been that in a long, long time (and when it was, people were whining that it wasn't a $1k tower).
[automerge]1576835309[/automerge]

Apple sold a $3000 workstation forever, and now suddenly doesn't anymore. Did something happen recently that made $3000 workstations no longer relevant?
 
  • Like
Reactions: SeattleMoose
Looked at the price tag and the fact that Apple “cheaped out” on the legacy (is the standard USB 3 port really legacy?) ports on the 2019 MacBook Pros and I will be keeping my mid 2012 MBP for a while longer.
Soldered in memory (non expandability) on the latest MBPs is a deal killer.
Expansion was an economical visit to Amazon now it’s a trip to an Apple dealer and hundreds of dollars plus having you computer “booked off sick” for weeks.
 
  • Like
Reactions: SeattleMoose
We've known that all along. I replaced my old Mac Pro with a Mac Mini because it offers adequate performance and because the new Mac Pro would be prohibitively expensive.

The new Mac Pro is not for you or me, but for people who need to crunch serious data. You can't install 1500 GB RAM on the MM and you can't get the same kind of throughput. You can't expand the MM.

Xeon processors (diesel) have always lagged Core processors (gasoline) in performance, but they are more reliable, expecially in parallell.

For photography, casual video and everyday use, it would not make any sense buying the new MP. The MM will serve all your needs. But if you are to edit several channels of 4K/8K footage, apply effects and encode longer movies, the MP will just blow other Macs away completely. If you have special gear and hardware needs and need the expandability, the MP is for you.

The new MP is not comparable to the old cheese graters. They were prosumer machines, for you and me. Apple could still fill that hole by offering a classic Mac Pro with the performance of a Mac Mini + GPU, but with more expandibility in a chassi rather than loose gadgets via TB.
Thanks for the insight. I have the 2012 quad-core Mac mini. I’ve maxed out the ram to 16 GB. Have a ssd for it but haven’t installed it yet.
 
err.. sucks for me. I want to buy the new Mac Pro, but 7500 for the model I'm looking at seems to perform very close to the same as the base iMac Pro. I just do NOT want a screen attached to my computer. So I can pay a few thousand just to have a screen removed and have upgradeability, or I can get the iMac Pro. In reality, I will probably do none of the above and just wait. This has been the same game for years with Apple. I and other co-workers of mine have held on to older models of Macs waiting for a reasonable one for years longer than we want while using PCs as our option when our needs require more power. I feel like Apple has lost a lot of money to people like us. I have skipped at least two upgrades and others I know have done the same. The 16" MacBook Pro did get a few to upgrade but I am still in need of a desktop computer that has NO screen, I already have 2 5k Monitors.

Just the opinion of a random developer who wants to run multiple docker containers, VMS, Xcode and simulators, while runnings side programs and scripts :) I do have a work provided MacBook Pro 2018 15" and it can do it, but man does it throttle and makes regular things very slow when I have a couple things from my list running at once.
 
  • Like
Reactions: mdatwood
Apple sold a $3000 workstation forever, and now suddenly doesn't anymore. Did something happen recently that made $3000 workstations no longer relevant?

Yes.

Modern computers have become more than adequate for the majority of tasks that people would do on them, largely negating the need for powerful workstations for all but the most demanding of workflows.

At the same time, Apple has been trying to migrate as many of their desktop users over to iMacs, and from the sound of it, they have been fairly successful, to the point where they clearly don’t think that a mid-tier Mac desktop is worth their time.


56847b4e4f226f105d1bae8addd96d63.jpg


Apple is likely also in love with the iMac form factor, which is easier to maintain and encapsulates their design-led ethos so well.

It’s a bit of everything.
 
The appropriate tool for the job.
First pre-purchase question: WHAT DO I NEED THE COMPUTER FOR?
You don’t need “HAL 9000” to send emails and create spreadsheets & text editor documents.
If I ever get into high end graphics I’ll build a monster machine for a fraction of the cost.
Bin there ... dun that.
 
I hope basic computer users understand that running Handbrake and video games is in no way a decent benchmark for any system in the $5k ballpark.
Video encoding isn't a legitimate use case for testing thermal performance? Huh.

It's a big thermal load on my CPU (uses AVX), much more than any game. It's certainly relevant to me that any desktop I use be able to do video encoding at reasonable CPU temperatures.
 
  • Like
Reactions: ssgbryan
But if they are truly using those machines for work purposes, what happens if there's an update that breaks the Hackintosh OS and they have to get work done?
Do they have to install the update. Installing new updates have not been a good idea in general lately anyways.
 
But if they are truly using those machines for work purposes, what happens if there's an update that breaks the Hackintosh OS and they have to get work done?

They aren't attached to the internet. They are set up as a dongle for the software.
 
I used to go with AMD when building my rigs, then moved to Intel once it became clear they were the performance leaders in the Core 2 and i7 years, but will be moving back to AMD for my next build unless Intel can overtake them again at some point.

I followed the same trajectory as you with PC builds. AMD, then Intel. I’d like to build my next box with an AMD chip. I think they’ve earned my business back. (And a 3800X with a PCIe4 SSD would be a really fast machine for a really long time. Probably 3 paced out GPU generations... kinda like my 3700K I guess. 🤷‍♂️)
 
Video encoding isn't a legitimate use case for testing thermal performance? Huh.

It's a big thermal load on my CPU (uses AVX), much more than any game. It's certainly relevant to me that any desktop I use be able to do video encoding at reasonable CPU temperatures.

But I didn't say Video Encoding.


I said HANDBRAKE


After a decade handbrake has gotten pretty good at decoding DVDs and other lightweight video formats.
 
But I didn't say Video Encoding.


I said HANDBRAKE


After a decade handbrake has gotten pretty good at decoding DVDs and other lightweight video formats.
I brought up HANDBRAKE in regard to "sustained performance." A big thermal load is a big thermal load.

Again, my 9700K will run a big AVX load like HANDBRAKE, all 8 cores at 4.8 GHz ALL DAY LONG and the core temperatures don't exceed 65C. The Mac Pro's "sustained performance" is nothing special.

32 core Threadripper 3970X, overclocked, can run full bore indefinitely and stay under 75C. Again, the Mac Pro's "sustained performance" is nothing special.
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.