MP 7,1 Waiting for Mac Pro 7,1 A1991 (no more)

[barmann]

But for big pro shops, disk array vendors have already thought ahead of what you are thinking and included redundant power supplies on enclosures. Something that the Mac Pro itself does not have.

Those pro shops you keep bringing up, that's a market Apple has been losing bit by bit for years .
The tcMP has been a massive nail in that coffin .

The very high end shops and big budget projects use a different kind of hardware and software than the majority of users anyways ; Apple never catered to that clientele, no matter what they used to say on one more thing day ( now known as my arse day ) .

If Apple won't provide the basics with the next MP , why would anyone sink money into a product and system that will then have been proven to be erratic at best ?

 

[deconstruct60]

The NMP maxed out the pci-e lanes so they only had 1 PCI-e SSD.

If an new mac pro shows up there should be the pci-e lanes to do 2 with 1 cpu and 2 video cards + 2 TB 3.0 buses.

Intel Xeon W has 48 PCI-e v3 lanes and a DMI link (to the PCH) that is equivalent to a x4 PCI-e v3 link. So net 'pragmatic' lanes coming in/out of CPU package is 52 v3 lanes. There is a "free" switch in the PCH chip (pragmatically up to another 10+ PCI-e v3 links. ). However, that is still hanging of the CPU package with just x4 worth of bandwidth. It is likely, (as with with iMac Pro) , the T2 SSD is hooked to the PCH so it is close to saturated if lean on the boot SSD very heavily. [ There is enough on the PCH to add Wifi , Bluetooth, maybe 1 GbE (if Apple wants three or would be willing to go asymmetric one 10GbE and one 1GbE )


Xeon E5 present in the MP 2013 only has 40 lanes of PCI-e v3. The DMI to the PCH is based on PCI-e v2. That PCH had v2 'fan out' links.

The Intel offerings have substantively changed in the 5-6 years Apple has spent in 'Rip van Winkle' mode.

Two x4 SSDs (8) , two 10GbE ( could share a x4 link ) (4) , two TB v3 controllers (8) , two x16 "big cards" (32 ) wouldn't be a problem.
8 + 4 + 8 + 32 = 52

Two easy things Apple could do. one or two x4 switches.
A x4 switch and a hardwired x4 10Gbe and a x4 ( x8-16 phys ) std socket.
A x4 switch and multiple m.2 slots on the same bandwidth. ( could go "wide" with more SSDs but not much of an increase in bandwidth. https://eshop.macsales.com/shop/express-4m2 only internal ).
They would have problem linearly increasing the bandwidth with the physical SSD slots. However, if skip the second GPU and put in a very high PCI-e card SSD could scale ( for example: https://www.anandtech.com/show/12700/intel-optane-ssd-905p-coming-soon as perhaps alternative to a RAM disk acting as an SSD. There are other more expensive add-in-cards that push bandwidth higher than x4 worth. )

 

[Joe The Dragon]

X2 10g-e only needs pci-e X4 2.0

 

[Biped]

Time to play everybody's favourite 'let's read into the marketing materials' yay ! ...

Welp, looks like another trashcan.

[picks up basketball, goes home]

 

[Simon R.]

Time to play everybody's favourite 'let's read into the marketing materials' yay ! ...

Welp, looks like another trashcan.

[picks up basketball, goes home]

LOL it's just Apple's new headquarters where the event takes place. Nothing to read into that I tell you; there will be no mention of the new Mac Pro.

 

[Mago]

I dony think Apple will sneak peek the mMP'19, neither to launch the Mac mini or updated iMac, too much mac stuff remaining this year to share an iPhone-focused event (with Watch series 4, Qi charger, 3 phones, new iPads expected in the same event too), I bet on anohter Mac Event next month as usual focused on iMac, mini, MacBooks (air, non-pro) and a fair Sneak Peek or introduction to the mMP19.

 

[deconstruct60]

I dony think Apple will sneak peek the mMP'19,

At the iPhone event? Pure delusion.

neither to launch the Mac mini or updated iMac, too much mac stuff remaining this year to share an iPhone-focused event (with Watch series 4, Qi charger,

Nevermind that Apple hasn't shared any significant Mac stuff at the vast majority of previous iPhone focused events.

3 phones, new iPads expected in the same event too), I bet on anohter Mac Event next month as usual focused on iMac, mini, MacBooks (air, non-pro) and a fair Sneak Peek or introduction to the mMP19.

if there is a Mac event coming later why wouldn't Apple "sneak peak" it at that one? If it is a Mac product sneak peeking, the Mac event would be far more focused event to do that at.

Not sure I'd bet on the iPads at the same event. iPhones (at least 3 ) + watch ( + charge pad + maybe Air Pod tweak ) and just maybe iPad Pros. I think an " iPhone SE Plus " ( a refactored 6 with a guts of 7 ) is more likely than an iPad Pro. If there is a "sneak peek" is may be the iPad Pro (if the product is 'disrupted' enough to give warning to the folks who like the current model. )
[doublepost=1535814857][/doublepost]

.....

Welp, looks like another trashcan.
....

Looks like an iPhone camera. There is a very high chance that the new iPhones have some feature(s) that Apple has been working on for a while. "Even better camera" (than your 2-3 year old iPhone) will be a selling point they are pushing.

 

[Bradleyone]

September Event program guide

Tim Cook walks out beaming, "Good morning! Things are going great!"

Keynote slide with "$1 Trillion" appears on stage.

Next Keynote slide: "(And we didn't get there selling Macs.)" No more mention of the Mac.

iPhone 9, XS, XXS announced.

20 min explanation on how to pronounce them.

iPad Pro. Face ID. Smaller bezels. The future of pro computing. Hint, hint, are you getting it?

tvOS, Apple Music sub numbers. Services, services, services.

Apple Watch 4. Thinner. Smaller bezels. Now works from orbit, measures your blood sugar levels in Zero G, displays it by changing the color of the watch band.

10 min schtick with an attractive intern aboard the International Space Station.


October Event program guide

No one walks out on stage, because it isn't held in a theater. "What, you thought we'd put on a show for Macs?! Lol, get out of here you dinosaurs. Did we mention the $1 Trillion we got to without caring about the Mac?"

Matthew Panzarino publishes a scrappy, coffee-stained, back of a napkin "Press Release".

Mac mini 2018: Black. Same size as an Apple TV. Actually, it is an Apple TV. Basically a hollowed out shell from leftover Apple TVs. 4 USB-C ports. 2 USB-A ports. Marco Arment says this makes it the best thing ever.

MacBook 2018: updated with whatever blood Intel can squeeze from their 14nm stone. "Now with 20% more turbo."

MacBook Air 2018: "Our engineers worked hard on this. We took our existing, much loved MacBook Airs, and fitted them with brand new, state of the art... keyboards."

New product: Mac Air. Apple branded cans of compressed air. Instant product shortages. Bestseller on eBay. News reports of cult gatherings holding "breathe the Apple Park air" parties. iFixIt does a teardown to reveal they're packaged in Shenzhen, you dumbasses.

ONE MORE THING

The non-existent stage (aka press release) falls silent as a small piece of paper slowly flutters down, like a leaf in the wind, watch how it soars.

On one side are the words, embossed in silver lettering, "The New Mac Pro"

On the other side is an image of a colossal, dual Xeon, DDR4-3200, 2x nVidia RTX 2080 Ti, PCIe 4.0, 16GT/s raw bandwidth, watercooled monster: think everything you've read in 700 pages of "Waiting for Mac Pro 7,1" but more.

Phil Schiller leaps onto the not-a-stage press release and says,

"Can't innovate, my ass. Well, they got that right. Anyway, here it is, everything you guys have been waiting for."

Keynote slide appears with a single line:

"Hackintosh. Apple says it's okay."

John Gruber says it's great because it can run Markdown. Well, it will, once he can find some drivers.

Epilogue: September 2019 Event

A dry wind blows through the decaying frame of Apple Park. Tumbleweeds and pages of old newspapers litter the landscape.

On one newspaper, a headline reads, "Apple stock crashes."
On another, a short article with a byline by Matthew Panzarino reads:

"Earth. Population approximately nine billion... All Borg."
"Android. I mean, all Android."

"How did this happen? How did a $1 Trillion company fall so fast?"

"After abandoning the Mac, which in the words of the late Tim Cook, was neither necessary nor sufficient to make Apple a $2 Trillion company, Apple suddenly found itself unable to compile the code to make its iOS devices, aka small slabs of glass and metal, actually do anything."

 

[Aldaris]

Phil Schiller leaps onto the not-a-stage press release and says,

"Can't innovate, my ass. Well, they got that right. Anyway, here it is, everything you guys have been waiting for."

Keynote slide appears with a single line:

"Hackintosh. Apple says it's okay."

Really need a laugh button.

 

[pl1984]

Feel free to list use cases that are common and rely on internal bays. Please do educate us idiots. I can't speak for others, but I am also not blindly advocating the removal of internal bays. I'm just saying that they're likely somewhere far down the wish list and if other compromises need to be made, so be it.

Here's one: Using an SSD for the boot / application drive and using internal storage for project files. Or how about this one: Using the SSD for boot / application drive and internal storage for virtual machines running on the system. I use both of these on my cMP and Z620.

 

[deconstruct60]

Feel free to list use cases that are common and rely on internal bays. Please do educate us idiots. I can't speak for others, but I am also not blindly advocating the removal of internal bays. I'm just saying that they're likely somewhere far down the wish list and if other compromises need to be made, so be it......

Internal storage and Internal bays aren't the same thing. One is a function and the other has more aspects of being a form ( being a commodity form factor is far more being a 'form' than being a 'function' ).


One, more common use cases that can be seen is folks with multiple OS instances inside of a machine. A current macOS instance for production use and other OS instances. That could be Windows ( on a separate drive). Could be older macOS versions. Pragmatically internal Time Machine clones for snapshots could be treated as another OS instance ( although can't boot from it, it soaks up at least as much space) .

Apple's new file/storage management system APFS starts to address this a subset of that. It supports having derivative macOS instances sharing the same (largely single ) drive. 'Copy on write' allows them to share a high percentage between OS 'updates/upgrades' but at some point that will push expansion pressure past a common single device.. ( it also only works on macOS versions going forward from now. )

Letting APFS ( in its CoreStorage aspect) consume a whole physical device is just cleaner. Yes, there is Boot Camp utilities to slice a drive up into partitions and add overhead that Windows needs, but there is more flexibility of just give the whole device to APFS policies to set up how the device is 'sliced up'. ( Same was true of CoreStorage.)

If a substantive number of users in the median/average space have multiple boot volumes then having multiple boot devices targets inside the machine is aligned with the user needs. If part of the feature set of the volume option boot screen is to select from a few OS volumes ... then having some set available inside probably is all that rare of an option.
This isn't a "My data is so humongous" issue. It is aligned with what folks do with the systems (can go look at large body of user configurations)

Second, different drives for different stuff ( this have overlap with the above, but a more general issues). Some data has significantly different access patterns and bandwidth profiles than other data. The Applications folder is a fundamentally different backing store than the scratch drive store for a large media manipulation tool. Separating, those out into different drives can help with performance and device lifetime. SSDs make highly variable, mixed use more tolerable on a single drive but it does still have trade-offs. A drive that can do 0.7-2.0 drive writes per day is going to cost different than a mainstream drive.

Moving paging/scratch drives farther away from the computation doesn't enhance performance. If have to fall out of RAM to disk (and back again) you shouldn't need to go far if it is an option. Similar issue with using a drive as a 'cache' to slower and/or high latency storage. For example, Optane as a middle tier between RAM and slower (and higher latency) nonvolatile storage. [ Going forward some other nonvolatile memory is likely going to come along and talk how. "everything on Flash" probably won't pan out. ] "More performance' will probably help the Mac Pro value proposition. ( a lot more than "has the cheapest storage". Apple isn't targeting Mac Pro primarily as the cheapest option. )

So even if Apple thinks SATA is "old" and "boring" that doesn't mean can do everything off of one and only one local drive ( especially is not very highly constrained on space. )

 

[handheldgames]

Let's just say that Thunderbolt 3 is far from a pro interface. On it's best day, it has a hard time competing with PCIe 2.0 x8. On top of that, the controller is limited by DMI 3.0 no matter how many TB3 channels you combine. If Apple provides onboard PCIe SSD ports, expect the proprietary interface and small SSD's like in the iMacPro. CPU socket and RAM are likely to be the only thing to be standard.

 

[Joe The Dragon]

Let's just say that Thunderbolt 3 is far from a pro interface. On it's best day, it has a hard time competing with PCIe 2.0 x8. On top of that, the controller is limited by DMI 3.0 no matter how many TB3 channels you combine. If Apple provides onboard PCIe SSD ports, expect the proprietary interface and small SSD's like in the iMacPro. CPU socket and RAM are likely to be the only thing to be standard.

well the imac pro does have an CPU with more lanes that can run a few TB3 buses on cpu.

 

[BlueTide]

….This isn't a "My data is so humongous" issue. It is aligned with what folks do with the systems (can go look at large body of user configurations)…

Second, different drives for different stuff ( this have overlap with the above, but a more general issues)...

Moving paging/scratch drives farther away from the computation doesn't enhance performance. … "More performance' will probably help the Mac Pro value proposition. ( a lot more than "has the cheapest storage". Apple isn't targeting Mac Pro primarily as the cheapest option. )

Thank you for this. I did mean to ask more specifically why some folks are so adamant about internal 3.5" platters - since I do see a point of having option for faster storage - blades, expansion cards or hat have you. I personally explored with multiple OSes myself, some booting from USB, but also with PC towers and hot-swap caddies to switch them easily. That was mostly to avoid human errors in erasing the other OSes, though.

I also very well understand the need for speed, such as case for scratch. Sometimes this NVME stuff just blows ones mind on how fast it is. It also sucks how little of it one can easily have. I've seen a few TB3 4 bay RAIDs for this, which do increase the storage potential a lot, but 4 bays with smallish drives is still relatively little. I also very well understand the need for more space, that one can get with platters. I still don't get why anyone would make a big deal of not having internal spinners as afaik, that can be worked around decently well in this day and age.

I don't think you're saying this, but I'll blabber it out anyway; the fact that a feature exists alone doesn't say anything about its usage. It's quite funny how much easier it is to add features to software than to take them out - when you try to take things out, you always find some small number of users that throw a hissy fit about it.

 

[pl1984]

I also very well understand the need for speed, such as case for scratch. Sometimes this NVME stuff just blows ones mind on how fast it is. It also sucks how little of it one can easily have. I've seen a few TB3 4 bay RAIDs for this, which do increase the storage potential a lot, but 4 bays with smallish drives is still relatively little. I also very well understand the need for more space, that one can get with platters. I still don't get why anyone would make a big deal of not having internal spinners as afaik, that can be worked around decently well in this day and age.

Why work around something when there was a perfectly acceptable solution to begin with? With four internal drive bays one can put 40TB inside the enclosure. Why introduce extra cost and the potential for increased failures by moving it outside of the enclosure? Space savings? If I have to purchase an external enclosure am I really saving any space?

 

[Nugget]

Why work around something when there was a perfectly acceptable solution to begin with? With four internal drive bays one can put 40TB inside the enclosure. Why introduce extra cost and the potential for increased failures by moving it outside of the enclosure? Space savings? If I have to purchase an external enclosure am I really saving any space?

Because for most users they either want zero drives or more than four. There's nothing magic or sweet about four bays other than the historical context. And, as has been explain ad nauseam, there's no performance advantage for internal platters any more so there's little incentive to live with the many downsides that arise from putting them inside a workstation.

In light of that, I'd rather Apple spare the cost, expense, and engineering tradeoffs they'd have to make in order to support internal drives. Producing a chassis that has four internal 3.5" bays comes at a cost of space, cooling, and other potential features that got dropped.

...the potential for increased failures...

External drives are unquestionably more reliable than internal drives. You're way off base with that objection.

 

[pl1984]

Because for most users they either want zero drives or more than four. There's nothing magic or sweet about four bays other than the historical context. And, as has been explain ad nauseam, there's no performance advantage for internal platters any more so there's little incentive to live with the many downsides that arise from putting them inside a workstation.

There's also no downside so why not have the best of both worlds?

External drives are unquestionably more reliable than internal drives. You're way off base with that objection.

I wasn't referring to the drives (which would be identical whether mounted internally or externally, nothing about mounting a drive in an external enclosure magically makes it more reliable.

 

[BlueTide]

Why work around something when there was a perfectly acceptable solution to begin with? With four internal drive bays one can put 40TB inside the enclosure. Why introduce extra cost and the potential for increased failures by moving it outside of the enclosure? Space savings? If I have to purchase an external enclosure am I really saving any space?

This is getting so childish/selfish that I'll just let you have this one and go chuckle by myself.

 

[Nugget]

There's also no downside so why not have the best of both worlds?

Because there are downsides, several of which I listed in the post you just replied to.

I wasn't referring to the drives (which would be identical whether mounted internally or externally, nothing about mounting a drive in an external enclosure magically makes it more reliable.

Neither was I referring to the drives themselves. Nobody with an interest in reliability, uptime, and uninterrupted workflow should care about the reliability of any single drive no matter where it is mounted. What is important is that external drives offer a wealth of options for improving uptime and reliability which are either impractical or impossible for internal drive bays in a Macintosh computer. Namely:

• External storage has the option for redundant power and network connections, eliminating single points of potential failure
• External storage offers more robust and performant filesystem options like btrfs or zfs which will outperform apfs and offer advanced data integrity protections which macOS does not natively support
• External storage scales higher than four bays, providing a wider variety of RAID options with multiple hot-standby drives and hot-swap drive bays
• External drives are easier to access from alternate or multiple workstations making it quicker and simpler to deal with a broken workstation.

All of these things can contribute to a much more reliable workflow and leave your complaints about "two power supplies" sounding myopic and hollow.

 

[mattspace]

In light of that, I'd rather Apple spare the cost, expense, and engineering tradeoffs they'd have to make in order to support internal drives. Producing a chassis that has four internal 3.5" bays comes at a cost of space, cooling, and other potential features that got dropped.

Except that removing drive bays between the cheesegrater and the trashcan wasn't used to reduce the cost of the machine, it didn't produce a machine with better cooling capabilities, and it didn't come with any significant new features that were enabled by removing the spinners.

There's no reason to believe Apple will do anything more profound than "make it thinner" when removing any feature you might care to name.

 

[pl1984]

Because there are downsides, several of which I listed in the post you just replied to.

If you want space savings buy a MacBook Pro. If you want better cooling then purchase an external enclosure with better cooling (I do not believe this is an area where the cMP design suffered). As for "other potential features" this is ambiguous and therefore meaningless.

Neither was I referring to the drives themselves. Nobody with an interest in reliability, uptime, and uninterrupted workflow should care about the reliability of any single drive no matter where it is mounted. What is important is that external drives offer a wealth of options for improving uptime and reliability which are either impractical or impossible for internal drive bays in a Macintosh computer. Namely:

• External storage has the option for redundant power and network connections, eliminating single points of potential failure
• External storage offers more robust and performant filesystem options like btrfs or zfs which will outperform apfs and offer advanced data integrity protections which macOS does not natively support
• External storage scales higher than four bays, providing a wider variety of RAID options with multiple hot-standby drives and hot-swap drive bays
• External drives are easier to access from alternate or multiple workstations making it quicker and simpler to deal with a broken workstation.

All of these things can contribute to a much more reliable workflow and leave your complaints about "two power supplies" sounding myopic and hollow.

If all of this is important to you then use external storage. Nothing about the cMP design prevents you from doing so. Meanwhile I, and others who like having everything in the system, are free to do so. You win, we win. It's a win-win situation. Fail to see the objection to such an offering.

 

[Nugget]

Fail to see the objection to such an offering.

This much is clear, you're working very hard to ignore what people are saying.

 

[JesterJJZ]

If all of this is important to you then use external storage. Nothing about the cMP design prevents you from doing so. Meanwhile I, and others who like having everything in the system, are free to do so. You win, we win. It's a win-win situation. Fail to see the objection to such an offering.

I never understand the logic of people that don't want certain features only because THEY don't use them. Will it hurt ANYONE if they include space for a couple 3.5mm drives? If you don't use them, then who cares? It doesn't affect you. It's the same thing with the folks that are glad the headphone jack is gone. Like seriously wtf?

 

[Nugget]

Will it hurt ANYONE if they include space for a couple 3.5mm drives?

Yes. Every component comes at at an increase in raw cost, plus the opportunity cost of the engineering time, effort, and compromises that have to be made in order to equip and support it. Requiring four internal 3.5" drive bays will have a large impact on the final product and there's virtually no technical justification for it.

I certainly won't be upset if the 2019 Mac Pro has bays for internal hard drives, but I would find it to be a confusing decision for them to have made and would indicate to me that they don't actually have a good grasp on the professional market. Apple clearly demonstrated terrible judgement with their design with the 2013 Mac Pro. As much as we're all waiting for a new Mac Pro we're also all waiting to see if they've come to their senses again from an engineering perspective.

Four SATA drive bays in a 2019 workstation would not be an encouraging sign in that regard.

 

[pl1984]

This much is clear, you're working very hard to ignore what people are saying.

No, they're not clear. At least I do not see them offering any benefit. I think I've been perfectly clear in stating as much. I see nothing but positives.