2022 Mac Pro said to use Intel Ice Lake Xeon W-3300 CPU

13

Comments

  • Reply 41 of 71
    crowleycrowley Posts: 10,453member
    Xed said:
    crowley said:
    Xed said:
    crowley said:
    Xed said:
    Makes sense. Professionals don’t like experiments and like to wait for a well established matured technology before they shift horses. 
    You think Apple designing and using their chips is an “experiment”? We’ll that’s one fucking  long abd established experiment.

    https://en.m.wikipedia.org/wiki/Apple_silicon
    Professionals tend to view their workstations a bit more sensitively than your average consumer electronics.
    Because someone wants Intel in their workstation it means that Apple designing their own chips for well over a decade is just an immature experiment. Got it!
    Apple hasn't shipped a single workstation with their own silicon.  They've only been shipping Macs with their own CPUs for 6 months.  I'm sure it'll probably be fine, but it'll definitely be an experiment.  The trash can Mac Pro was also an experiment, because that form factor hadn't been done before, so I imagine there's a few people still feeling a little sore from that.
    By that measure than any change becomes an experiement. Are you saying that professionals never buy new Macs? Are you saying that no one who uses their Mac for work bought a trashcan Mac Pro? How about professionals who focus on video and photos? Did no one buy the Pro Display XDR because it was an "experiment," as you put it, despite being superior in its abilities?

    The simple fact is that Apple is transitioning from Intel to ARM, and when they have all the parts ready and in place to make the transition for a Mac type they're going to fucking do it despite your foolish, anti-Apple, "professionals want real processors", derp derp argument.
    That's not my "argument" at all, calm down.  Are you a first adopter?  I'm not, and many people don't want to be first adopters, because any big shift is liable to see problems that need to be worked out.  Some people also wait to update iOS and macOS until they hear if there are bug issues.  The M1 shift has gone pretty well so far, but if my livelihood depended on my machine then I think I'd be even more reticent to buy into a branch new system until I knew that it was established and stable.  Especially since Apple's history with the pro workstations has been spotty at best.

    So no, I'm not being anti-Apple, and nor am I saying "professionals want real processors".  I'm saying that when it comes to businesses and making money, many people are risk averse, so Mac Pro customers in this area may appreciate seeing a bit more bedding in time with the M-series before they make the move.  I have no idea how prevalent that thought is, but I'm sure it exists.

    Note that Apple refreshed the PowerMac G5 in later 2005, after the Intel transition had started, and it was the final Mac to be updated, probably precisely to give a similar kind of assurance that the platform was good.
    avon b7darkvaderelijahgnadriel
  • Reply 42 of 71
    MarvinMarvin Posts: 15,310moderator
    This is just a refresh for the current Mac Pro. No big deal. It’s cost-effective for them, and likely has always been part of the plan. The insider news here is Apple has settled on which generation of Xeon-W they will use for this, and when.

    But this does indicate, I think, that the first Apple Silicon Mac Pro will have a different form factor from the current design. The “mini-me” concept shown in the illustration here seems absurd to me, almost a joke. I could see them producing a module for the current Mac Pro, but I can’t see them making that the only option. iMac Pro reborn is a possibility, but it could also be something that is the heir to the cylinder, something that goes really well with the XDR display.
    Making a shorter Mac Pro would be a simple way to redesign it. The shorter one would look better:



    Even this form factor is overkill for Apple's chips - a lot of the space still left there is for external memory chips. The highest level of performance needed in this kind of product is about 8-16x the performance of the M1 and 10x the memory capacity. Apple can fit this into a 200-300W chip, especially on 3nm in 2022.

    The current Mac Pro with the highest Intel and AMD config idles at 300W according to Apple:

    https://support.apple.com/en-us/HT201796

    The old iMac Pro could handle 500W. The cylinder Mac Pro was around 300W.

    If they want to still support PCIe cards (Afterburner etc) then this form factor makes some sense but they can support PCIe externally using a connector.

    I could see them just making this form factor EOL, it depends on how much they value rack-mountable machines. The amount of sales will be in the low tens of thousands per year.

    When people see how much power is in the iMac and even 16" MBP chips, it will encourage another portion of pro users to switch to those instead.

    Apple doesn't even have to beat Intel/AMD in performance, it's enough to offer better performance per dollar. Intel and AMD vastly overprice their workstation and server chips because it's such a low volume market and high cost investment. Apple can make these chips at a fraction of the price and they don't need to overprice them because selling chips isn't a line of business they are in, their profit comes from the sale of the whole unit. Apple Silicon allows Apple to undercut Intel/AMD in price by as much as $12k.
    williamlondonh2pwatto_cobra
  • Reply 43 of 71
    avon b7avon b7 Posts: 7,625member
    Marvin said:
    This is just a refresh for the current Mac Pro. No big deal. It’s cost-effective for them, and likely has always been part of the plan. The insider news here is Apple has settled on which generation of Xeon-W they will use for this, and when.

    But this does indicate, I think, that the first Apple Silicon Mac Pro will have a different form factor from the current design. The “mini-me” concept shown in the illustration here seems absurd to me, almost a joke. I could see them producing a module for the current Mac Pro, but I can’t see them making that the only option. iMac Pro reborn is a possibility, but it could also be something that is the heir to the cylinder, something that goes really well with the XDR display.
    Making a shorter Mac Pro would be a simple way to redesign it. The shorter one would look better:



    Even this form factor is overkill for Apple's chips - a lot of the space still left there is for external memory chips. The highest level of performance needed in this kind of product is about 8-16x the performance of the M1 and 10x the memory capacity. Apple can fit this into a 200-300W chip, especially on 3nm in 2022.

    The current Mac Pro with the highest Intel and AMD config idles at 300W according to Apple:

    https://support.apple.com/en-us/HT201796

    The old iMac Pro could handle 500W. The cylinder Mac Pro was around 300W.

    If they want to still support PCIe cards (Afterburner etc) then this form factor makes some sense but they can support PCIe externally using a connector.

    I could see them just making this form factor EOL, it depends on how much they value rack-mountable machines. The amount of sales will be in the low tens of thousands per year.

    When people see how much power is in the iMac and even 16" MBP chips, it will encourage another portion of pro users to switch to those instead.

    Apple doesn't even have to beat Intel/AMD in performance, it's enough to offer better performance per dollar. Intel and AMD vastly overprice their workstation and server chips because it's such a low volume market and high cost investment. Apple can make these chips at a fraction of the price and they don't need to overprice them because selling chips isn't a line of business they are in, their profit comes from the sale of the whole unit. Apple Silicon allows Apple to undercut Intel/AMD in price by as much as $12k.
    One of my doubts here is if we will finally see a move to cloud 'computing' in these next few years and away from needing local power both in terms of pure compute performance and the energy required to achieve it.

    I've seen white papers pointing in that direction but I always remember the 'netbooting' scenarios that were the 'future' years ago and never really materialised. 
    killroyh2p
  • Reply 44 of 71
    crowleycrowley Posts: 10,453member
    Marvin said:

    Intel and AMD vastly overprice their workstation and server chips because it's such a low volume market and high cost investment. Apple can make these chips at a fraction of the price and they don't need to overprice them because selling chips isn't a line of business they are in, their profit comes from the sale of the whole unit. Apple Silicon allows Apple to undercut Intel/AMD in price by as much as $12k.
    Seems like a whole load of assumption going on there.  How can Apple make low-volume, high-cost investment chips at a fraction of the price?  You say it as if it's a given, but there's no certainty there.

    And if their profit comes from the sale of the whole of the unit then how can they undercut Intel/AMD by as much as $12k?  That's only $1k less than the price of a Mac Pro equipped with the best processor Apple will BTO!  That top of the line Mac Pro comes with an Intel 3275M which is around $3000 retail, so your numbers don't make sense to me at all.  

    Apple might be able to achieve some economies of scale by reusing chips that they've developed for other machines, but that will mean the Mac Pro has few advantages over those other machines apart from expansion slots (at a hefty premium).  I doubt they will do that.  And if they have to invest in a higher grade of performance chips for the Mac Pro then they may be able to do better and/or cheaper than Intel and AMD, but that's yet to be seen.
    elijahg
  • Reply 45 of 71
    ajmasajmas Posts: 597member
    I suspect if they are doing this it’s because many businesses are conservative and already have a heavy investment in software that is optimised for thar chipset. We need to consider the cost of new software and the reality is that many software houses won’t have a finished stable port of their applications.    

    The real question is whether they will be offering an ARM based Mac Pro at the same time?
    watto_cobra
  • Reply 46 of 71
    MarvinMarvin Posts: 15,310moderator
    avon b7 said:
    One of my doubts here is if we will finally see a move to cloud 'computing' in these next few years and away from needing local power both in terms of pure compute performance and the energy required to achieve it.

    I've seen white papers pointing in that direction but I always remember the 'netbooting' scenarios that were the 'future' years ago and never really materialised. 
    For things like video processing and real-time feedback, cloud computing isn't practical. Video footage is TBs of data, uploading data to the cloud for processing is the biggest hurdle. For things like rendering, cloud computing works better as the processing load is much higher than the data size.
    crowley said:
    Marvin said:

    Intel and AMD vastly overprice their workstation and server chips because it's such a low volume market and high cost investment. Apple can make these chips at a fraction of the price and they don't need to overprice them because selling chips isn't a line of business they are in, their profit comes from the sale of the whole unit. Apple Silicon allows Apple to undercut Intel/AMD in price by as much as $12k.
    Seems like a whole load of assumption going on there.  How can Apple make low-volume, high-cost investment chips at a fraction of the price?  You say it as if it's a given, but there's no certainty there.

    And if their profit comes from the sale of the whole of the unit then how can they undercut Intel/AMD by as much as $12k?  That's only $1k less than the price of a Mac Pro equipped with the best processor Apple will BTO!  That top of the line Mac Pro comes with an Intel 3275M which is around $3000 retail, so your numbers don't make sense to me at all.  

    Apple might be able to achieve some economies of scale by reusing chips that they've developed for other machines, but that will mean the Mac Pro has few advantages over those other machines apart from expansion slots (at a hefty premium).  I doubt they will do that.  And if they have to invest in a higher grade of performance chips for the Mac Pro then they may be able to do better and/or cheaper than Intel and AMD, but that's yet to be seen.
    Apple can make more cost-effective high-end chips because they are only building them for their own use and only have to produce a single unit. Intel/AMD make dozens of chip models and each one has to have a separate design and manufacturing process. They have to charge customers thousands to make that process profitable.

    Apple only has to manufacture around 50k chips per year for high-end use. Even if it cost them $500 per chip to manufacture, that's $25m.

    The current high-end spec of the Mac Pro is $24k for both CPU and GPU. Apple can match that spec on Apple Silicon for over $12k less. There are PCs that are in this price range too:

    https://www.newegg.com/p/N82E16859152110
    https://www.amazon.com/AMD-Ryzen-Threadripper-PRO-3995WX/dp/B08V5HPXVY
    https://www.amazon.com/EVGA-24G-P5-3975-KR-GeForce-Cooling-Backplate/dp/B08J61SS5R/

    Apple can make something competitive with this performance-wise for a much lower price on 3nm chips. They have complete freedom to make whatever they want. They've said in the past, the Mac Pro is a passion project for them, it doesn't make a lot of money and now they don't have to pass any revenue on to Intel or AMD.

    Like I say though, the tower form factor hasn't made much sense for Apple to keep making it for at least a decade and it makes less sense now than ever with such efficient hardware. Apple has stated repeatedly their goal was always to make the hardware disappear, that's why the new iMacs are so compact. The Mac Pro has stuck around as an ugly wart on that goal for far longer than necessary and this was only due to the failings of Intel, Nvidia and AMD over the years. Now they can go it alone and build exactly the hardware they want.
    williamlondonh2pwatto_cobra
  • Reply 47 of 71
    XedXed Posts: 2,519member
    crowley said:
    Xed said:
    crowley said:
    Xed said:
    crowley said:
    Xed said:
    Makes sense. Professionals don’t like experiments and like to wait for a well established matured technology before they shift horses. 
    You think Apple designing and using their chips is an “experiment”? We’ll that’s one fucking  long abd established experiment.

    https://en.m.wikipedia.org/wiki/Apple_silicon
    Professionals tend to view their workstations a bit more sensitively than your average consumer electronics.
    Because someone wants Intel in their workstation it means that Apple designing their own chips for well over a decade is just an immature experiment. Got it!
    Apple hasn't shipped a single workstation with their own silicon.  They've only been shipping Macs with their own CPUs for 6 months.  I'm sure it'll probably be fine, but it'll definitely be an experiment.  The trash can Mac Pro was also an experiment, because that form factor hadn't been done before, so I imagine there's a few people still feeling a little sore from that.
    By that measure than any change becomes an experiement. Are you saying that professionals never buy new Macs? Are you saying that no one who uses their Mac for work bought a trashcan Mac Pro? How about professionals who focus on video and photos? Did no one buy the Pro Display XDR because it was an "experiment," as you put it, despite being superior in its abilities?

    The simple fact is that Apple is transitioning from Intel to ARM, and when they have all the parts ready and in place to make the transition for a Mac type they're going to fucking do it despite your foolish, anti-Apple, "professionals want real processors", derp derp argument.
    That's not my "argument" at all, calm down.  Are you a first adopter?  I'm not, and many people don't want to be first adopters, because any big shift is liable to see problems that need to be worked out.  Some people also wait to update iOS and macOS until they hear if there are bug issues.  The M1 shift has gone pretty well so far, but if my livelihood depended on my machine then I think I'd be even more reticent to buy into a branch new system until I knew that it was established and stable.  Especially since Apple's history with the pro workstations has been spotty at best.

    So no, I'm not being anti-Apple, and nor am I saying "professionals want real processors".  I'm saying that when it comes to businesses and making money, many people are risk averse, so Mac Pro customers in this area may appreciate seeing a bit more bedding in time with the M-series before they make the move.  I have no idea how prevalent that thought is, but I'm sure it exists.

    Note that Apple refreshed the PowerMac G5 in later 2005, after the Intel transition had started, and it was the final Mac to be updated, probably precisely to give a similar kind of assurance that the platform was good.
    Sure it is. You didn't initiate it, but you supported it. Again, Apple designing their own chips is neither immature nor an experiment in the way the OP stated and the way you have backed it up.
    Soliwilliamlondonwatto_cobra
  • Reply 48 of 71
    XedXed Posts: 2,519member
    avon b7 said:
    Xed said:
    Makes sense. Professionals don’t like experiments and like to wait for a well established matured technology before they shift horses. 
    You think Apple designing and using their chips is an “experiment”? We’ll that’s one fucking  long abd established experiment.

    https://en.m.wikipedia.org/wiki/Apple_silicon
    He is saying that professionals prefer to sit back and wait for hardware/software to mature before moving to a new platform.
    As the saying goes, no shit, Sherlock.
    Soliwilliamlondonwatto_cobra
  • Reply 49 of 71
    canukstormcanukstorm Posts: 2,695member
    Xed said:
    crowley said:
    Xed said:
    crowley said:
    Xed said:
    Makes sense. Professionals don’t like experiments and like to wait for a well established matured technology before they shift horses. 
    You think Apple designing and using their chips is an “experiment”? We’ll that’s one fucking  long abd established experiment.

    https://en.m.wikipedia.org/wiki/Apple_silicon
    Professionals tend to view their workstations a bit more sensitively than your average consumer electronics.
    Because someone wants Intel in their workstation it means that Apple designing their own chips for well over a decade is just an immature experiment. Got it!
    Apple hasn't shipped a single workstation with their own silicon.  They've only been shipping Macs with their own CPUs for 6 months.  I'm sure it'll probably be fine, but it'll definitely be an experiment.  The trash can Mac Pro was also an experiment, because that form factor hadn't been done before, so I imagine there's a few people still feeling a little sore from that.
    By that measure than any change becomes an experiement. Are you saying that professionals never buy new Macs? Are you saying that no one who uses their Mac for work bought a trashcan Mac Pro? How about professionals who focus on video and photos? Did no one buy the Pro Display XDR because it was an "experiment," as you put it, despite being superior in its abilities?

    The simple fact is that Apple is transitioning from Intel to ARM, and when they have all the parts ready and in place to make the transition for a Mac type they're going to fucking do it despite your foolish, anti-Apple, "professionals want real processors", derp derp argument.
    Don't be a dipshit. It isn't a exactly a secret that the trashcan Mac Pro was a failure in design and in sales for its intended market.  So much so that Apple barely even bothered to update that machine and they had to make a public apology to Pro users and let them know that a redesigned Mac Pro was in the works.  You and a handful of Apple fanatics are the only ones in denial.
    darkvaderwilliamlondonelijahgnadriel
  • Reply 50 of 71
    canukstormcanukstorm Posts: 2,695member
    avon b7 said:
    Xed said:
    Makes sense. Professionals don’t like experiments and like to wait for a well established matured technology before they shift horses. 
    You think Apple designing and using their chips is an “experiment”? We’ll that’s one fucking  long abd established experiment.

    https://en.m.wikipedia.org/wiki/Apple_silicon
    I don't think you should be reading 'experiment' in a literal sense in this case.

    His other comments qualify what he means very clearly.

    He is saying that professionals prefer to sit back and wait for hardware/software to mature before moving to a new platform. In fact, you could also read 'experiment' as referring to pro users themselves. 

    I'm sure many do wait and get by on very old yet stable setups.

    Being able to purchase a modern intel based system will surely appeal to many professionals for the reasons he indicated. 

    It is yet to be seen if an M series pro line will even be a commercial success for professional users. In an unprecedented move Apple had to apologise for what happened with the cyclinder line and pre-announce that a new design was in the works. 

    Apple hasn't been exactly up front with its pro plans - ever. It has pulled rugs out from under feet, backtracked and left products to die or stagnate at almost every turn. Radically altered pricing, file formats etc. 

    'Commitment' has always been an issue for Apple. It's quite easy to drop or vastly alter product direction in the consumer electronics market and Apple is constantly pushing consumers to upgrade. Even resulting in 'stealth' upgrades for major systems updates. Unless you read the system update dialog carefully, choosing to 'not update now' could result in the system being 'automatically' updated in the middle of the night. You only have to look at the way that dialog is presented to the user, to see the intent behind it. Also, there is no simple click option to permanently put off an upgrade. It has never been clear about what product features are dropped either.

    That's a lot of reasons for pro users to wait for platform stability and maturity before switching. 
    100% on the money.
    darkvaderwilliamlondonelijahgnadriel
  • Reply 51 of 71
    XedXed Posts: 2,519member
    Xed said:
    crowley said:
    Xed said:
    crowley said:
    Xed said:
    Makes sense. Professionals don’t like experiments and like to wait for a well established matured technology before they shift horses. 
    You think Apple designing and using their chips is an “experiment”? We’ll that’s one fucking  long abd established experiment.

    https://en.m.wikipedia.org/wiki/Apple_silicon
    Professionals tend to view their workstations a bit more sensitively than your average consumer electronics.
    Because someone wants Intel in their workstation it means that Apple designing their own chips for well over a decade is just an immature experiment. Got it!
    Apple hasn't shipped a single workstation with their own silicon.  They've only been shipping Macs with their own CPUs for 6 months.  I'm sure it'll probably be fine, but it'll definitely be an experiment.  The trash can Mac Pro was also an experiment, because that form factor hadn't been done before, so I imagine there's a few people still feeling a little sore from that.
    By that measure than any change becomes an experiement. Are you saying that professionals never buy new Macs? Are you saying that no one who uses their Mac for work bought a trashcan Mac Pro? How about professionals who focus on video and photos? Did no one buy the Pro Display XDR because it was an "experiment," as you put it, despite being superior in its abilities?

    The simple fact is that Apple is transitioning from Intel to ARM, and when they have all the parts ready and in place to make the transition for a Mac type they're going to fucking do it despite your foolish, anti-Apple, "professionals want real processors", derp derp argument.
    Don't be a dipshit. It isn't a exactly a secret that the trashcan Mac Pro was a failure in design and in sales for its intended market.  So much so that Apple barely even bothered to update that machine and they had to make a public apology to Pro users and let them know that a redesigned Mac Pro was in the works.  You and a handful of Apple fanatics are the only ones in denial.
    The Mac Pro was a market failure long before the trashcan. It's why they tried a different form factor in the first place. Calling that an "experiment" and in the same breath calling Apple foolish for designing their own chips is about the most assinine thing I've read on this forum.

    The current Mac Pro is better, but it's still not great. I doubt that it ever can be. The overhead is too high for the low volume of sales. The era for giant desktops is long over, regardless of what Golden Age you want to pretend will come back around.

    You call it failure—and it is for Apple—but in terms for yielding a profit for many professionals over the Mac Pro they were using it was a clearly a success. You may think the cost high but others found that it paid for itself rather quickly. I know several people that buy the latest Mac Pro because of this and they will surely buy a Apple Silicon Mac that renders faster than the Mac Pro they use today if it increases their revenue and profit stream. Why would you leave money on the table?

    This was always going to be the last Mac to switch because it's the least impacted by their user base and offers Apple the least back in value while being the most amount of effort. This shouldn't be a surprise. Once the transition comes I suspect we'll see the same form factor with Intel and Apple Silicon as options in a (nearly) otherwise identical form factor. Then the market will decide how much the slower and costlier Intel Mac is worth when there's an option. I wouldn't be surprised if we see both options in the 16" MBP when it's finally released.

    The Intel Mac Pro is a dinosaur on its way to extinction, even though it will likely get another round or two of Intel chips. You and a handful of Apple fanatics are the only ones in denial.
    edited July 2021 Solialbatrossflyerwilliamlondonwatto_cobra
  • Reply 52 of 71
    HPC includes anyone running MATLAB on a workstation, probably configured with 128G or more of RAM.

    Couldn't Apple just produce and sell a replacement processor module for the MacPro?
    watto_cobra
  • Reply 53 of 71
    crowleycrowley Posts: 10,453member
    Xed said:
    crowley said:
    Xed said:
    crowley said:
    Xed said:
    crowley said:
    Xed said:
    Makes sense. Professionals don’t like experiments and like to wait for a well established matured technology before they shift horses. 
    You think Apple designing and using their chips is an “experiment”? We’ll that’s one fucking  long abd established experiment.

    https://en.m.wikipedia.org/wiki/Apple_silicon
    Professionals tend to view their workstations a bit more sensitively than your average consumer electronics.
    Because someone wants Intel in their workstation it means that Apple designing their own chips for well over a decade is just an immature experiment. Got it!
    Apple hasn't shipped a single workstation with their own silicon.  They've only been shipping Macs with their own CPUs for 6 months.  I'm sure it'll probably be fine, but it'll definitely be an experiment.  The trash can Mac Pro was also an experiment, because that form factor hadn't been done before, so I imagine there's a few people still feeling a little sore from that.
    By that measure than any change becomes an experiement. Are you saying that professionals never buy new Macs? Are you saying that no one who uses their Mac for work bought a trashcan Mac Pro? How about professionals who focus on video and photos? Did no one buy the Pro Display XDR because it was an "experiment," as you put it, despite being superior in its abilities?

    The simple fact is that Apple is transitioning from Intel to ARM, and when they have all the parts ready and in place to make the transition for a Mac type they're going to fucking do it despite your foolish, anti-Apple, "professionals want real processors", derp derp argument.
    That's not my "argument" at all, calm down.  Are you a first adopter?  I'm not, and many people don't want to be first adopters, because any big shift is liable to see problems that need to be worked out.  Some people also wait to update iOS and macOS until they hear if there are bug issues.  The M1 shift has gone pretty well so far, but if my livelihood depended on my machine then I think I'd be even more reticent to buy into a branch new system until I knew that it was established and stable.  Especially since Apple's history with the pro workstations has been spotty at best.

    So no, I'm not being anti-Apple, and nor am I saying "professionals want real processors".  I'm saying that when it comes to businesses and making money, many people are risk averse, so Mac Pro customers in this area may appreciate seeing a bit more bedding in time with the M-series before they make the move.  I have no idea how prevalent that thought is, but I'm sure it exists.

    Note that Apple refreshed the PowerMac G5 in later 2005, after the Intel transition had started, and it was the final Mac to be updated, probably precisely to give a similar kind of assurance that the platform was good.
    Sure it is. You didn't initiate it, but you supported it. Again, Apple designing their own chips is neither immature nor an experiment in the way the OP stated and the way you have backed it up.
    Well fine.  I can't have a conversation with you if you're just going to make shit up.  Have a nice day.
    tenthousandthingselijahgnadriel
  • Reply 54 of 71
    Is the M1 RAM ECC?  This is typically a "must have" in HPC workstations.
    darkvaderwilliamlondonwatto_cobra
  • Reply 55 of 71
    crowleycrowley Posts: 10,453member
    Marvin said:
    avon b7 said:
    One of my doubts here is if we will finally see a move to cloud 'computing' in these next few years and away from needing local power both in terms of pure compute performance and the energy required to achieve it.

    I've seen white papers pointing in that direction but I always remember the 'netbooting' scenarios that were the 'future' years ago and never really materialised. 
    For things like video processing and real-time feedback, cloud computing isn't practical. Video footage is TBs of data, uploading data to the cloud for processing is the biggest hurdle. For things like rendering, cloud computing works better as the processing load is much higher than the data size.
    crowley said:
    Marvin said:

    Intel and AMD vastly overprice their workstation and server chips because it's such a low volume market and high cost investment. Apple can make these chips at a fraction of the price and they don't need to overprice them because selling chips isn't a line of business they are in, their profit comes from the sale of the whole unit. Apple Silicon allows Apple to undercut Intel/AMD in price by as much as $12k.
    Seems like a whole load of assumption going on there.  How can Apple make low-volume, high-cost investment chips at a fraction of the price?  You say it as if it's a given, but there's no certainty there.

    And if their profit comes from the sale of the whole of the unit then how can they undercut Intel/AMD by as much as $12k?  That's only $1k less than the price of a Mac Pro equipped with the best processor Apple will BTO!  That top of the line Mac Pro comes with an Intel 3275M which is around $3000 retail, so your numbers don't make sense to me at all.  

    Apple might be able to achieve some economies of scale by reusing chips that they've developed for other machines, but that will mean the Mac Pro has few advantages over those other machines apart from expansion slots (at a hefty premium).  I doubt they will do that.  And if they have to invest in a higher grade of performance chips for the Mac Pro then they may be able to do better and/or cheaper than Intel and AMD, but that's yet to be seen.
    Apple can make more cost-effective high-end chips because they are only building them for their own use and only have to produce a single unit. Intel/AMD make dozens of chip models and each one has to have a separate design and manufacturing process. They have to charge customers thousands to make that process profitable.

    Apple only has to manufacture around 50k chips per year for high-end use. Even if it cost them $500 per chip to manufacture, that's $25m.

    The current high-end spec of the Mac Pro is $24k for both CPU and GPU. Apple can match that spec on Apple Silicon for over $12k less. There are PCs that are in this price range too:

    https://www.newegg.com/p/N82E16859152110
    https://www.amazon.com/AMD-Ryzen-Threadripper-PRO-3995WX/dp/B08V5HPXVY
    https://www.amazon.com/EVGA-24G-P5-3975-KR-GeForce-Cooling-Backplate/dp/B08J61SS5R/

    Apple can make something competitive with this performance-wise for a much lower price on 3nm chips. They have complete freedom to make whatever they want. They've said in the past, the Mac Pro is a passion project for them, it doesn't make a lot of money and now they don't have to pass any revenue on to Intel or AMD.

    Like I say though, the tower form factor hasn't made much sense for Apple to keep making it for at least a decade and it makes less sense now than ever with such efficient hardware. Apple has stated repeatedly their goal was always to make the hardware disappear, that's why the new iMacs are so compact. The Mac Pro has stuck around as an ugly wart on that goal for far longer than necessary and this was only due to the failings of Intel, Nvidia and AMD over the years. Now they can go it alone and build exactly the hardware they want.
    Again, this is just a total bunch of guesswork, seems to have some massive gaps, and seems weirdly hostile to the Mac Pro as a concept.  You're completely overlooking Apple's R&D costs, licensing costs, the fact that they can't profit on any chip volume sales and that 3nm is as of now an unknown quantity that Apple have not used in the CPU of any shipping product.   

    The assertion that "Apple can match that spec on Apple Silicon for over $12k less" (previously "as much as $12k" now "over $12k"?) is completely baseless as of now.  Apple have no Apple Silicon product that can match that spec, and the products that they do have are not so much cheaper to justify anywhere near that kind of optimism.
    darkvaderelijahgnadriel
  • Reply 56 of 71
    darkvaderdarkvader Posts: 1,146member
    This seems unlikely.

    It's very unlike Apple to admit that they made a terrible mistake.  Apple silicon was and is a terrible mistake.
    williamlondonelijahg
  • Reply 57 of 71
    robabarobaba Posts: 228member
    Wow-lots of FUD here.  Only a few minutes before my mid morning shift begins, but I just want to point out a few things.

    M1X is in production as we speak.
    the Mac Pro workstation will feature a chiplet with multiple M1X instances on it.
    These chipsets will be cost effective simply because of the volume of the M1X.
    the Mac Pro Workstation will not need to exist in its current huge chunk of milled aluminum tower format simply due to the fact that Apple silicon produces a fraction of the heat of x86.
    watto_cobra
  • Reply 58 of 71
    mpantonempantone Posts: 2,033member
    Xed said:

    The Mac Pro was a market failure long before the trashcan. It's why they tried a different form factor in the first place. Calling that an "experiment" and in the same breath calling Apple foolish for designing their own chips is about the most assinine thing I've read on this forum.

    The current Mac Pro is better, but it's still not great. I doubt that it ever can be. The overhead is too high for the low volume of sales. The era for giant desktops is long over, regardless of what Golden Age you want to pretend will come back around.
    And yet, the mid-tower case is typical form factor in the HEDT/workstation market.

    Remember that Apple returned to the standard form factor BASED ON THE DESIRES AND FEEDBACK OF MAC PRO USERS. They did not like the trashcan. Apple made a mistake making the Mac Pro design whimsical rather than practical. When the system is under a desk, in a server room, or mounted in a rack, it really doesn't matter that it's a mid-tower PC case.

    Externalizing the PCIe bus via Thunderbolt was also poorly received. It's helpful for a notebook computer or a SFF desktop like the Mac mini. For a larger desktop configuration, relying on Thunderbolt for expansion makes very little sense. As an owner of a Sonnet Breakaway eGFX chassis, I will state that this is a rather large box for just one graphics card. There are no connections for anything else, so an inefficient use of space.


    williamlondonh2pelijahgnadrielwatto_cobra
  • Reply 59 of 71
    mpantonempantone Posts: 2,033member
    robaba said:
    Wow-lots of FUD here.  Only a few minutes before my mid morning shift begins, but I just want to point out a few things.

    M1X is in production as we speak.
    the Mac Pro workstation will feature a chiplet with multiple M1X instances on it.
    These chipsets will be cost effective simply because of the volume of the M1X.
    the Mac Pro Workstation will not need to exist in its current huge chunk of milled aluminum tower format simply due to the fact that Apple silicon produces a fraction of the heat of x86.
    My guess is that Apple will reuse the current Mac Pro chassis for the first iteration of the Apple Silicon Mac Pro.

    I would not expect a redesigned Mac Pro chassis until the second iteration of the ASi Mac Pro. The current Mac Pro design has been on the market for a while and there are now third-party accessories, mounting hardware, etc. for this specific design. Moreover, this Mac Pro has a modular component design, Apple might be able to redesign the individual modules and still keep the main chassis the same (or very close). Remember that Apple prefers to get some mileage out of an industrial design.

    Perhaps more importantly, Mac Pro customers (I'm thinking production houses, etc.) can swap out Intel Mac Pros with ASi Mac Pros without reconfiguring cages, racks, whatever.
    williamlondonrobabawatto_cobra
  • Reply 60 of 71
    killroykillroy Posts: 271member
    Marvin said:

    Like I say though, the tower form factor hasn't made much sense for Apple to keep making it for at least a decade and it makes less sense now than ever with such efficient hardware. Apple has stated repeatedly their goal was always to make the hardware disappear, that's why the new iMacs are so compact. The Mac Pro has stuck around as an ugly wart on that goal for far longer than necessary and this was only due to the failings of Intel, Nvidia and AMD over the years. Now they can go it alone and build exactly the hardware they want.

    I would have to disagree, if it wasn't for the pro towers they would not have the Pro market where I work. The infrastructure where I work requires PCI slots. There's a lot of 32 gigabit fiber-optic networking for video editing and all the apps or Adobe and Avid. If Apple ever put out another product like the trashcan my employer may never approve an Apple  product again no matter what chip is in it.
    edited July 2021 williamlondonh2pelijahgnadrielwatto_cobra
Sign In or Register to comment.