Is there any reason to expect an iMac Pro refresh? I always assumed it was a one time spaceholder due to the inability to get the new MacPro released.
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.
MP 7,1 W5700X Release
- Thread starter darkmatter343
- Start date
- Sort by reaction score
Is there any reason to expect an iMac Pro refresh? I always assumed it was a one time spaceholder due to the inability to get the new MacPro released.
If this is based on the W5700X being a "replacement" for the 580X , then I suspect going to be deeply disappointed. AMD's W5700 Pro base prices at $799. Apple's W5700X is likely to be substantially more. There are some substantive bill of materials differences and economies of scale differences:
a. 8GB move VRAM. (more/denser stuff , higher cost )
b. the W5700X is closer to the RX 5700 XT ( 40 CUs) than the 5700. (more CUs , higher cost. )
c. two TB controllers, 4 TB PHYs ports, and switch. ( MPX connector PCI-e lanes that need to go somewhere ).
d. Apple deluxe packaging. (for instance: full MPX module so quite large , highly engineered heat sink )
pretty good chance this breaks the $1000 barrier. ( or at least is pretty close to it). Which gets into a bit of a catch-22 ( more expensive sell less so lower volumes , less scale ... smaller base to spread fixed costs over ... so higher cost. )
All of that is going to push a Mac Pro 2019 with a W5700X card and 1TB of SSD farther away from the low to median priced iMac Pro. ( and amore way from the higher end iMac BTO options. ). Going to be talking about two substantively different price points.
[automerge]1582405317[/automerge]
Apple never said anything like that. In April 2017 the pointed explicitly that there was demand in this space drive by folks who were OK with iMac (and MBPs). It isn't a one off. It has much higher development priority than the Mac Pro did. ( arrived two years earlier so hard pressed as to why there would be any doubt about that. ) There were lots of folks outside of Apple who constantly repeated this as if it was true. But Apple never said that in the slightest.
If Apple can pack more bang for buck value into the iMac Pro there is zero reason why it can't stay.
And if did a redesign so got at least "RAM door" parity with the iMac ( or iMac 27" falls into same no "RAM door" hole) then even more so. The iMac is the best selling desktop that Apple has. The Mac Pro is apparently stuffed with a "low volume tax". ( .e.g, dwarfish base SSD. super mature GPU , etc. yet $5,999 starting point. There is more than usual margin there. )
If Apple had come with a $3,999-4,999 Mac Pro then maybe. But this 100% increase in base entry price says tons about trying to avoid the the old Mac Pro entry price range and the spot where the bulk of the iMac Pro lies in now. If Apple dropped the iMac Pro there would be almost nothing there but some quirky iMac BTO variants. "the iMac Pro is just a placeholder until they roll out the mid-range xMac" ... errr probably not.
Last edited:
There has been no indication either way. Both iMac and iMacPro are due for a refresh and/or redesign. If screen is based on XDR in any capacity, it would need to support target display mode to command price. Almost “useless” without.
The screen "feature" that the iMac Pro more so needs is the calibration aspects that the XDR is suppose to eventually get more so than the 6K or "beyond HDR range" feature. The iMac Pro far more so needs calibrated pixels rather than more. It isn't like the current display is "bad". But much of the distain from the "pro" space is about control (by the user) rather than the number of pixels presented.
The XDR backlight would probably be problematical given it needs to be actively cooled to be that accurate and the iMac internals aren't going to help at all with that in the slightest. If it is just the panel without the backlight that may not help Apple all that much in terms of scale ( same part, multiple products).
The iMac Pro CPU and GPU are relatively quite expensive. Throttling those to pragmatic nothingness via target display mode isn't a broad range, high value add. If LG wanted to do a 6K monitor with a backlight of their own design and Apple wanted to piggy back on that. That isn't "target display mode" as much as a specifically targeted display. While there were multiple monitor implementers of 5K initially eventually they all faded away to leave Apple ( iMac ) and Apple proxy (LG Ultrafine pragmatically outsourced from Apple). A 5K panel that took DP v1.4 as a baseline (so single stream) seems like it would get more traction than sticking with dual stream 6k.
There was a comment previous about the iMac Pro being a stopgap, "dead end" product until the Mac Pro got out. The XDR looks to be far more a stopgap, "dead end" than the iMac Pro is. If Apple gets microLEDs working anywhere decently the current XDR will get chucked into the "killed" bucket pretty fast. "can't innovate my a__ " of 2013 is the "best pro display in the world" of 2019. '
Also Target Display Mode runs into substantive complexity when have to project two DisplayPort streams rerouted from the outside. Two streams from GPU to display have to be switched with alternative outside feed. Not a huge leap but twice as much stuff to do (for something not done all that often. ). If wanted to bring that back getting to single stream 5k displays would probably help substantially more.
P.S. in Mac Pro space a single stream 5K display would help boost the number of 5K displays could run with these base GPU cards as the GPUs start at DP v1.4 output basics (and pick up some a bit more base VRAM).
Last edited:
The Vega II doesn't have the Instinct's full ECC pipeline - I'd bet on it being a memory-expanded Radeon VII based card, rather than being Instinct derived. Much like the "FirePro" DX00, was Radeon-based, not FirePro based. Up-speccing a consumer tech, is far more Apple's style, than down-speccing a professional tech - eg XDR.
It is the exact same die with some features turned off/on. The Vega II has Infinity Fabric turned on (at a slightly narrower width (not as many 'lanes')). That makes it much closer to the MI series than Radeon VII and probably a major portion of where the substantive price jump comes from.
Radeon VII was gimped on features and sold probably close to at cost just to fill a large tactical gap that AMD had. ( Navi was way late. Mac Pro was going to be relatively late so Vega II's weren't going to consume the die starts. , etc. ). If there were no MI50/60 or Vega II products AMD would have never done the Radeon VII.
A/V data streams typically have a casual approach to bit level data integrity ( pragmatically few case because the errors typically buried unnoticeably. ). As long as the Mac Pro target user base is skewed that way Apple probably won't put much effort there. That matches up with Dx00/FirePro from the MP 2013 era. ( but once again though it is the same die with features switched off/on. )
However, there is no Infinity Fabric in AMD's consumer space at all. With the major upgrades to Metal to support Infinity Fabric set up that isn't not a minor feature that is likely going to evaporate on the next tip end MPX module.
The catch-22 with the new Mac Pro is that the only way to pack the max number of substantive GPUs into the system is by using linked MPX modules. Yes, more than a few folks are going to with 3rd party add-in GPUs for 1-2 range (and some 3 edge cases on the border or power constraints ), but getting to 4 is basically through MPX. Linking those 4 with high speed bandwidth is directly in the Infinity Fabric wheelhouse. Even more so given the current system has a PCI-e v3 constraint ( that is also relatively highly oversubscribed. )
Is "Big Navi" going to bring Infinity Fabric to the consumer space? So far little indication of that. First, AMD has made comments about the gaming space being monolithic, 'max big die' oriented. ( chiplets and MCM not the path). Second, the major sponsors of Navi ( the game console vendors MS and Sony) are even less interested in non monolithic solutions. Navi coupled to HBM2 and/or external Infinity Fabric may not happen at all. It basically ends up being primarily a Polaris upgrade path and the rest of the line up breadth is filled with alternatives.
Last edited:
Vega 20 based MI50/MI60 were introduced in November 2018 time frame. Apple didn't ship Vega II until December 2019 time frame. For Apple .... so far this isn't a long time. Folks marinated on Mac Pro 2019 initial specs from June 2019 until December 2019. "Dog ate my homework" meetings in April 2017 , 2018. Same group of folks running this MPX module roll out.
it would be May-June before they were off of their established track record. The EM reulations ( FCC , etc. ) submissions for the card suggest they were going to try to do better but their brittle supply chain they don't have total control over. So could very well be May-June.
That seriously should have extremely little to do with Apple shipping the hardware. They don't necessarily require a NAB song and dance to release something.
Major clean up of last drivers ( Numerous 5700 issues covered in release notes https://www.amd.com/en/support/kb/release-notes/rn-rad-win-20-2-2 ) it is at least as likely that there are hardware (production) and software issues holding this up. ( If they got it substantively wrong in the primary target Windows drivers then pretty good chance it was wrong in the macOS ones also. )
The MPX modules are made in China. If the Chinese source production line is substantively understaffed then keeping the 580X and Vega II output going with the reduced resources would be much higher priority than releasing a 4th option.
If the software and hardware is ready and NAB disappears that really have zero impact on releasing. Apple has a gigantic hole in price point options. Apple doesn't need NAB to generate interest in filling that hole.
I agree. I've read the "NAB" comments here quite a few times and scratched my head. While Apple sometimes appears at NAB to discuss their NLE, effects software, etc, they rarely if ever introduce hardware around it. There's no reason to. Like other "coming soon" options that have happened in past Apple products, the W5700X will just one day appear. And that'll be that. No grand announcement surrounding it or any such thing; it'll just be available on the list. We've seen it with the laptops, and this will likely not be any different.
Given Apple's recent track record with the Mac Pro - expect W5700X some time around 2025
Radeon Pro W5700 is not an old GPU, but yes, at $5999, it should be the base MPX module for Mac Pro.
I think it's more likely because it's a higher binned W5700 (yield issues for the die). Or because of software bugs (Navi drivers/firmware).
The PCB + heatsink isn't going to be the holdup on this thing. It's a full size module where everything custom about it can either be taken as-is from the Vega II, or has been proven out on the Vega II. The Vega II heatsink can be used practically as-is. The TB3 controller design has been proven out on the Vega II. Custom PCBs based on reference designs aren't exactly difficult things to get right.
I would not be surprised at all if Apple pretty much has this ready to go, but has either got caught off balance by yield issues, or lingering bugs that are taking a lot longer to fix than they expected. Neither would surprise me, considering the state of Navi drivers on both Mac and Windows since last year.
That said, since you can get a W5700 today that will work with bootscreen support. Or a RX 5700 XT. What is in it for Apple to make their own clone GPUs?
Truthfully, this is all on Apple for the delay. Good for them, positive points, whatever spin you want to put on it for telling people in December 2019 they would be announcing/releasing a model with W5700X as BTO option, but three months later there have been no updates. It does not bode well for their new PRO focused machine strategy that they were supposed to be returning to. Seems like more of the same - get what we give you and we won't give you an update to plan ahead. It's almost embarrassing.