MP 7,1 Is the new 7,1 Mac Pro a failure on arrival

[Abazigal]

But why go with Intel at all?

More money, less performance.

My guess is that macOS has been thoroughly optimised for intel processors, and software such as Final Cut Pro make use of features like intel quick sync for much of their performance advantages.

It’s also possible that Apple may have a contract with intel to purchase their chips, or they feel like the advantages AMD has is a generational phenomenon that they can simply ride out.

If and when AMD can demonstrate that they are able to sustain this lead, Apple might consider switching, but it’s not something they will rush right into, IMO.

 

[Flint Ironstag]

Out of interest, I went to the dell Website and built a comparable system.12 core 32gb Ram 256gb NVMe drive Mouse and Keyboard.

The Dell has a lower clock speed CPU and the video card is slower than the 580X

Dell = $5572
Mac Pro $6999

So, lets amend that to an approx $1427 Apple tax. Maybe as little as $1100 if the CPU and GPU were the same.

A couple of things of note. The Dell has 8 x Sata connectors. A PCI to SSD card for the MP from Sonnet runs at $99 for a two SSD interface.

Not sure if the dell supports TB3 etc.

Other benefits of the MP of the Dell:

More aesthetically pleasing
Better build quality
Higher quality supporting components (case, screws, covers, etc)
Easier Access to internals
Product box has been specifically designed to be used again, for people who ship their MP's around.
Larger Power supply than the Dell (200%)
Dead silent under load due to airflow design.

Let's not forget that if you're in a major market, Dell will be onsite the same day or next to repair or replace your money-making workstation. Included in the $5572.

 

[Adult80HD]

Let's not forget that if you're in a major market, Dell will be onsite the same day or next to repair or replace your money-making workstation. Included in the $5572.

Maybe....but my experience with Dell's support back my IT days was abysmal at best. Banished them from the workplace. HP was much better...and guess what, more expensive. And on the specialty stuff? Almost never really same day or next day, at least not to replace hardware. Maybe diagnose the problem, but replacement hardware was often much slower. They've got caveats to let them out on those things. Mass produced stuff, yes. That said...Apple could do much better on the hardware support side. Of course, it also rarely fails.

 

[defjam]

Other benefits of the MP of the Dell:

More aesthetically pleasing
Better build quality
Higher quality supporting components (case, screws, covers, etc)
Easier Access to internals
Product box has been specifically designed to be used again, for people who ship their MP's around.
Larger Power supply than the Dell (200%)
Dead silent under load due to airflow design.

Maybe the latter two have some value but the others are of questionable value, especially the one in favor of the box.

 

[R3k]

Maybe the latter two have some value but the others are of questionable value, especially the one in favor of the box.

The box is pretty awesome actually. Its pretty much a road case.

 

[defjam]

The box is pretty awesome actually. Its pretty much a road case.

It may very well be. I feel that it's not a selling point.

 

[JazzyGB1]

Maybe....but my experience with Dell's support back my IT days was abysmal at best. Banished them from the workplace. HP was much better...and guess what, more expensive. And on the specialty stuff? Almost never really same day or next day, at least not to replace hardware. Maybe diagnose the problem, but replacement hardware was often much slower. They've got caveats to let them out on those things. Mass produced stuff, yes. That said...Apple could do much better on the hardware support side. Of course, it also rarely fails.

Apple's cMP rarely failed either (I'm typing this on my now 11 year old model) but it didn't start at $6k either, so the price for quality argument is a complete red herring.
Apple's new Mac Pro isn't $6k because its more reliable, it's $6k because they chose it to be, which is a complete kick in the teeth to many prior Mac Tower owners (like myself) who'd waited nearly 8 years for it.
You can come up with many excuses for their decision, but they are just that...excuses.
Apple are taking a liberty.
I've probably a few years left in my Mac Pro but I've now started using Cubase & Pro Tools in preparation for jumping platforms and when I'm as competent at using them as I am with Logic (which I've used since eMagic days), my next major purchase will likely be a Tower PC.
Not because I'm leaving Apple, but because Apple left me.

 

[defjam]

Apple's new Mac Pro isn't $6k because its more reliable, it's $6k because they chose it to be, which is a complete kick in the teeth to many prior Mac Tower owners (like myself) who'd waited nearly 8 years for it.

This. Take the PCIe slot covers as an example. They're solid, high quality parts. But did they need to be of such quality that they won't bend even if you attempt to do so by hand?

 

[Adult80HD]

This. Take the PCIe slot covers as an example. They're solid, high quality parts. But did they need to be of such quality that they won't bend even if you attempt to do so by hand?

This is all true...but that's their prerogative. If it means you leave Apple, so be it. The MP market is so small I doubt that they truly care that much about it overall except as a brand statement--and the build quality and aesthetics of this are all about brand statement, as is the price. This is a brand strategy for a lot of companies, it's just what it is. It sucks for those who feel boxed out now, I get it, but I don't see Apple changing their strategy any time soon, as overall it's clearly working very, very well for them.

 

[defjam]

This is all true...but that's their prerogative. If it means you leave Apple, so be it. The MP market is so small I doubt that they truly care that much about it overall except as a brand statement--and the build quality and aesthetics of this are all about brand statement, as is the price. This is a brand strategy for a lot of companies, it's just what it is. It sucks for those who feel boxed out now, I get it, but I don't see Apple changing their strategy any time soon, as overall it's clearly working very, very well for them.

I agree, was merely pointing out one of the aspects that supports Jazzy's statement.

 

[Krevnik]

My guess is that macOS has been thoroughly optimised for intel processors, and software such as Final Cut Pro make use of features like intel quick sync for much of their performance advantages.

It’s also possible that Apple may have a contract with intel to purchase their chips, or they feel like the advantages AMD has is a generational phenomenon that they can simply ride out.

If and when AMD can demonstrate that they are able to sustain this lead, Apple might consider switching, but it’s not something they will rush right into, IMO.

I'd probably add that it isn't just Apple in this position, either. PC brands like Apple switching to AMD wholesale aren't just a matter of swapping things out. The supply chain needs to be able to handle it. And generally, an established market with older companies are less likely to stab out on things that introduce risk. They'll let the smaller companies prove things out and take the risks.

So yeah, I agree it's a case of:

• AMD needed to deliver on the promises on Zen 2. These chips were pretty rough around the edges right up through release day with both BIOS and Microcode updates needed (CPUs with a day one patch, almost funny), making the engineering parts hard to evaluate. I decided to build a Ryzen 3600 gaming PC around launch week, and that was not a great idea. I like the machine, but it was a bit of a mess as I was needing both the fixes for Destiny 2, but also some of the weird "let's boost while mostly idle" behavior these chips had at launch. I can totally see Apple, Dell and the like all taking a wait and see approach here.
• AMD needs to be ready to scale up. A couple large PC brands switching would let AMD greatly increase their sales, maybe even double it. That also means they need to be able to deliver those parts. They are wholly dependent on TSMC for fabrication, which also introduces more moving parts for AMD to deal with when expanding production.
• Apple/etc needs to update PCB designs and BOMs as AMD is scaling up. So it makes more sense to wait for the next refresh, rather than change course in the middle of a cycle.

The 2019 Mac Pro in particular, along with the release timeline for the ThreadRipper 3000 chips was just badly timed. Apple would have been taking on the risk of jumping over to CPUs that are relatively aggressive and ambitious, based on engineering samples that weren't really ready for integration into pre-built systems to be stress tested. And trying to launch shortly after the CPUs themselves did. It would have paid off if they did, but it may have made the Mac Pro even more late. I suspect in some ways Apple wanted to cut down on risks like that in the short term.

Now, if Apple were to get a good relationship with AMD where Apple could work closely with them on to get engineering samples faster and with fixes earlier. I could see them jumping over easily. Intel and Apple do share a business relationship that benefits Apple, which is part of the calculus for Apple. Having input on Thunderbolt was good for them, and I can't imagine CPUs like the BGA desktop packages for the 8500/8700 for the 2018 Mini would exist without Apple asking for them.

Going to AMD would impact that relationship, and so from Apple's perspective, there's a decent level of risk involved for them. I'd be surprised if they aren't at least trying to see if they can put an AMD APU or CPU into a Mac at some point in the next year or so.

At the very least, Apple has been making the OS less Intel-centric lately. One reason you can build a Ryzen Hackintosh if you want now.

EDIT: In general though, I do think Intel has a problem. The inertia with the big PC brands will help give them time to respond, but Intel has overly segmented the market in order to maximize profit: Ultrabook, Laptop, Performance Laptop, NUC, Desktop, HEDT, Workstation, Server. These all have product lines super specialized for that segment. But it tends to force people into particular buckets as well. Want ECC? No affordable HEDT CPU for you, pay up for a Workstation CPU. And AMD is taking advantage of how Intel has sliced up the market to maximize profits to wedge themselves into the market. Especially in the Desktop -> HEDT -> Workstation space.

 

[fuchsdh]

I'd probably add that it isn't just Apple in this position, either. PC brands like Apple switching to AMD wholesale aren't just a matter of swapping things out. The supply chain needs to be able to handle it. And generally, an established market with older companies are less likely to stab out on things that introduce risk. They'll let the smaller companies prove things out and take the risks.

So yeah, I agree it's a case of:

• AMD needed to deliver on the promises on Zen 2. These chips were pretty rough around the edges right up through release day with both BIOS and Microcode updates needed (CPUs with a day one patch, almost funny), making the engineering parts hard to evaluate. I decided to build a Ryzen 3600 gaming PC around launch week, and that was not a great idea. I like the machine, but it was a bit of a mess as I was needing both the fixes for Destiny 2, but also some of the weird "let's boost while mostly idle" behavior these chips had at launch. I can totally see Apple, Dell and the like all taking a wait and see approach here.
• AMD needs to be ready to scale up. A couple large PC brands switching would let AMD greatly increase their sales, maybe even double it. That also means they need to be able to deliver those parts. They are wholly dependent on TSMC for fabrication, which also introduces more moving parts for AMD to deal with when expanding production.
• Apple/etc needs to update PCB designs and BOMs as AMD is scaling up. So it makes more sense to wait for the next refresh, rather than change course in the middle of a cycle.

The 2019 Mac Pro in particular, along with the release timeline for the ThreadRipper 3000 chips was just badly timed. Apple would have been taking on the risk of jumping over to CPUs that are relatively aggressive and ambitious, based on engineering samples that weren't really ready for integration into pre-built systems to be stress tested. And trying to launch shortly after the CPUs themselves did. It would have paid off if they did, but it may have made the Mac Pro even more late. I suspect in some ways Apple wanted to cut down on risks like that in the short term.

Now, if Apple were to get a good relationship with AMD where Apple could work closely with them on to get engineering samples faster and with fixes earlier. I could see them jumping over easily. Intel and Apple do share a business relationship that benefits Apple, which is part of the calculus for Apple. Having input on Thunderbolt was good for them, and I can't imagine CPUs like the BGA desktop packages for the 8500/8700 for the 2018 Mini would exist without Apple asking for them.

Going to AMD would impact that relationship, and so from Apple's perspective, there's a decent level of risk involved for them. I'd be surprised if they aren't at least trying to see if they can put an AMD APU or CPU into a Mac at some point in the next year or so.

At the very least, Apple has been making the OS less Intel-centric lately. One reason you can build a Ryzen Hackintosh if you want now.

EDIT: In general though, I do think Intel has a problem. The inertia with the big PC brands will help give them time to respond, but Intel has overly segmented the market in order to maximize profit: Ultrabook, Laptop, Performance Laptop, NUC, Desktop, HEDT, Workstation, Server. These all have product lines super specialized for that segment. But it tends to force people into particular buckets as well. Want ECC? No affordable HEDT CPU for you, pay up for a Workstation CPU. And AMD is taking advantage of how Intel has sliced up the market to maximize profits to wedge themselves into the market. Especially in the Desktop -> HEDT -> Workstation space.

Yep. Biggest thing to take away is that even if Apple never switches to AMD, Intel is going to have to respond, and that's good for consumers.

The Xeon -W line is a good example of that in action—the new X1XX CPUs that will (hopefully) end up in an iMac Pro refresh are substantially cheaper than their predecessors.

 

[Schismz]

This. Take the PCIe slot covers as an example. They're solid, high quality parts. But did they need to be of such quality that they won't bend even if you attempt to do so by hand?

So, they should just release a Mac Pro without the Heavy Metal Sculpture built out of Unobtanium, and offer a model with a cardboard box or plastic cover and feet for $1500 less. Presto, problem solved! The slot covers are really cool; everything inside the machine is highly-engineered, luxury level build quality. Price was clearly not something they were trying to constrain and lower, they went all out including the almost total lack of cables inside. I mean it's really nice; they probably could've knocked $1500 off the price by making the same machine less premium. (But then... who's going to buy an iMac Pro? If the Mac Pro is their "most powerful computer ever built" and the flagship, it needs to cost more than everything else or it messes up all their other SKUs)

I'm using conjecture, but does Apple plan/hope to sell 25,000 units, 50K units is a win? I have no idea, but I strongly suspect the price factors in all the support costs for a very niche product very few within the Apple ecosystem even know exists. With Coronavirus helping cause problems, I have this feeling they're selling a lot less of everything. Okay it's made in Texas in the US, from 100% parts that come from China. How many units actually made it out the door between the 4-6 weeks it was available within reasonable timeframes.

Mac Pro always appears in the EPP store immediately; something like a new iPhone or MacBook can takes months, because there are so many customers willing to pay full retail who want it, so supply is constrained even without the current shortages. With Mac Pro supply hasn't been constrained from EPP in 14 years... It's always been a niche market for Apple, and I think with this last iteration it's become ultra-niche, concept car/flagship territory where everybody has to pay for the minuscule volume sold. I mean, there is obviously a lack of parts, you can't even buy any of the cards in the Apple store if you didn't get them upon ordering; there is nothing in stock, and all arriving Real Soon Now.

 

[ZombiePhysicist]

. Okay it's made in Texas in the US, from 100% parts that come from China. How many units actually made it out the door between the 4-6 weeks it was available within reasonable timeframes.

This is wrong. To be allowed the “assembled in USA” status allows US FTC allows around at most 20% foreign content:

Assembled in USA Claims
A product that includes foreign components may be called "Assembled in USA" without qualification when its principal assembly takes place in the U.S. and the assembly is substantial. For the "assembly" claim to be valid, the product’s last "substantial transformation" also should have occurred in the U.S. That’s why a "screwdriver" assembly in the U.S. of foreign components into a final product at the end of the manufacturing process doesn’t usually qualify for the "Assembled in USA" claim.

Example: A lawn mower, composed of all domestic parts except for the cable sheathing, flywheel, wheel rims and air filter (15 to 20 percent foreign content) is assembled in the U.S. An "Assembled in USA" claim is appropriate.

Example: All the major components of a computer, including the motherboard and hard drive, are imported. The computer’s components then are put together in a simple "screwdriver" operation in the U.S., are not substantially transformed under the Customs Standard, and must be marked with a foreign country of origin. An "Assembled in U.S." claim without further qualification is deceptive.

Complying with the Made in USA Standard

Introduction Basic Information About Made In USA Claims The Standard For Unqualified Made In USA Claims Qualified Claims Comparative Claims The FTC and The Customs Service Other Statutes What To Do About Violations Introduction The Federal Trade Commission (FTC) is charged with preventing

www.ftc.gov

 

[nbritton]

So is the Mac Pro a failure on arrival? A case of too little too late? Or do you think it's the right Mac for our time?

Apple priced themselves out. I have a Dell R920 with four Intel E7-8895 v2 processors (60 cores, 120 threads) and 1024 GB of ram, it also has 120 TB of raw storage, and three Radeon RX 580. I bought everything on eBay for like $10k total and it will decimate Apple's highest end Mac Pro... my Geekbench 4 score is just over 100k, and if you run four Geekbench processes at the same time the aggregate score sums to like 250k. The point is, why do I need to fork over $50k to Apple when I can build out a substantially better system for just $10k? At this point, I'm thinking of the Mac in terms of it just being a thin client and all the heavy lifting is done on the backend with Linux. I just finished a proof of concept where I ran macOS in a virtual machine on my R920 with GPU and USB3 passthrough. It worked, the only technical challenge left to solve is how to cable three displays that are 10 meters away from the box.

 

[ZombiePhysicist]

Apple priced themselves out. I have a Dell R920 with four Intel E7-8895 v2 processors (60 cores, 120 threads) and 1024 GB of ram, it also has 120 TB of raw storage, and three Radeon RX 580. I bought everything on eBay for like $10k total and it will decimate Apple's highest end Mac Pro... my Geekbench 4 score is just over 100k, and if you run four Geekbench processes at the same time the aggregate score sums to like 250k. The point is, why do I need to fork over $50k to Apple when I can build out a substantially better system for just $10k? At this point, I'm thinking of the Mac in terms of it just being a thin client and all the heavy lifting is done on the backend with Linux. I just finished a proof of concept where I ran macOS in a virtual machine on my R920 with GPU and USB3 passthrough. It worked, the only technical challenge left to solve is how to cable three displays that are 10 meters away from the box.

[/sarcasmON] You must not be a "real" pro. You sound like a "Mac Tinker". [/sarcasmOff]

Yea, those are some crazy numbers.

Here is another thing. Cat8 lets you do 40Gbps networking today. Fiber goes up to 400Gbps. Apple is sooooooo woefully behind the curve on networking options, it's kind of stunning. Many of these options are dirt cheap too. Cat8 cabling and NIC cards are dirt cheap. I just dropped 100Gbps fiber around the house, the cabling is cheap. Apple not having a multi CPU chip option is also disappointing.

I wouldn't think it's fair to compare them to all the highest most extreme cases, but the value proposition is profoundly out of wack at this point.

 

[jasonmvp]

Apple is sooooooo woefully behind the curve on networking options, it's kind of stunning. Many of these options are dirt cheap too. Cat8 cabling and NIC cards are dirt cheap.

Hyperbole much? Sheesh. Apple isn't woefully behind in any of the networking options at all. Are you supposing they include 25G NICs? 40G NICs? 50? 100? How many other workstations out there have anything more than 10G to them at the moment? And are those workstations that do, the exception? (Answer: yes).

There's absolutely nothing wrong with 2x10GigE copper interfaces on the Mac Pro. Zilch. Nada. Nothing. It doesn't put Apple "woefully behind" anyone of note.

 

[ZombiePhysicist]

Hyperbole much? Sheesh. Apple isn't woefully behind in any of the networking options at all. Are you supposing they include 25G NICs? 40G NICs? 50? 100? How many other workstations out there have anything more than 10G to them at the moment? And are those workstations that do, the exception? (Answer: yes).

There's absolutely nothing wrong with 2x10GigE copper interfaces on the Mac Pro. Zilch. Nada. Nothing. It doesn't put Apple "woefully behind" anyone of note.

Yes, how many WiFi laptops were out when they released the iBook. How many USB ports were on PCs when they released the original iMac. Apple has always been bringing advances to the market early.

Your argument is basically "well everyone sucks more". Ok, we're all set in the land of low expectations. Revel in the tepid waters.

 

[jasonmvp]

Yes, how many WiFi laptops were out when they released the iBook. How many USB ports were on PCs when they released the original iMac. Apple has always been bringing advances to the market early.

Your argument is basically "well everyone sucks more". Ok, we're all set in the land of low expectations. Revel in the tepid waters.

No, my argument is more reality based. Find a copper-based 25G or 40G interface and controller chip that doesn't overheat itself into oblivion while it's running. Here's a hint: they don't exist. The 10G chip in the Mac is already a fire breather and needs its own heat sink. And you want faster?

Oh, wait, I know: go optics instead. That'll fix everything, right? Sure, and then drive the end users' costs up massively. And finally, what home or office infrastructure are these 25/40/... equipped Mac Pros going to connect to? Have you priced out faster switches lately? I'm sure you realize that those switches are ALL data center centric, right? They cost a cubic butt-ton PER PORT, and make so much noise you can't hear yourself think.

Pretty clear you don't do networking for a living. As it turns out: that's precisely what I do.

But that's OK, the Mac Pro provides you x16 PCI-E slots that can handle 100G! So run right out and snag a 100G card for your Mac Pr.... oh... wait... you don't have one yet, do you? Too busy bitching about its content and pricing.

Never mind. Carry on.

 

[ZombiePhysicist]

No, my argument is more reality based. Find a copper-based 25G or 40G interface and controller chip that doesn't overheat itself into oblivion while it's running. Here's a hint: they don't exist. The 10G chip in the Mac is already a fire breather and needs its own heat sink. And you want faster?

Oh, wait, I know: go optics instead. That'll fix everything, right? Sure, and then drive the end users' costs up massively. And finally, what home or office infrastructure are these 25/40/... equipped Mac Pros going to connect to? Have you priced out faster switches lately? I'm sure you realize that those switches are ALL data center centric, right? They cost a cubic butt-ton PER PORT, and make so much noise you can't hear yourself think.

Pretty clear you don't do networking for a living. As it turns out: that's precisely what I do.

But that's OK, the Mac Pro provides you x16 PCI-E slots that can handle 100G! So run right out and snag a 100G card for your Mac Pr.... oh... wait... you don't have one yet, do you? Too busy bitching about its content and pricing.

Never mind. Carry on.

Yes yes, producing a Nic controller that doesn't overheat, I dont know, by using a fab not using car lanes for lithography would be really an insurmountable thing for apple to ask for. As for fiber, it doesn't have to cost anything. The transceivers are the expensive part. Having the port isn't that costly.

But you're so progressive on these things, if only they had you on the iMac team back in the day, I'm sure we'd all still be enjoying our ADB keyboards.

 

[jasonmvp]

Yes yes, producing a Nic controller that doesn't overheat, I dont know, by using a fab not using car lanes for lithography would be really an insurmountable thing for apple to ask for. As for fiber, it doesn't have to cost anything. The transceivers are the expensive part. Having the port isn't that costly.

But you're so progressive on these things, if only they had you on the iMac team back in the day, I'm sure we'd all still be enjoying our ADB keyboards.

Wow. You don't understand the topic of the debate even remotely, so you lash out. Good job.

Apple doesn't produce the NIC controller used in the Mac. It's an Aquantia chip, one that's used on several PC motherboards, et al. It works, and works quite well. It just gets hot. Any 10GBASE-T controller chip does so, as it turns out. 10GBASE-T is blazingly hot. Given that, who would Apple turn to to create an on-board 25G or 40G or .. controller chip that can handle two ports and doesn't melt in place? Oh, and you haven't even specified the speeds these interfaces should run, either. Just that "Apple is behind."

Multi-mode fiber isn't expensive. The optics are. And the switch ports they go into ALSO are. You conveniently left that part out of your counter-point.

Anything else you'd like me to take apart in your position here? Or are you content to keep spewing hyperbole about a product you really don't know much about?

 

[ZombiePhysicist]

Wow. You don't understand the topic of the debate even remotely, so you lash out. Good job.

Apple doesn't produce the NIC controller used in the Mac. It's an Aquantia chip, one that's used on several PC motherboards, et al. It works, and works quite well. It just gets hot. Any 10GBASE-T controller chip does so, as it turns out. 10GBASE-T is blazingly hot. Given that, who would Apple turn to to create an on-board 25G or 40G or .. controller chip that can handle two ports and doesn't melt in place? Oh, and you haven't even specified the speeds these interfaces should run, either. Just that "Apple is behind."

Multi-mode fiber isn't expensive. The optics are. And the switch ports they go into ALSO are. You conveniently left that part out of your counter-point.

Anything else you'd like me to take apart in your position here? Or are you content to keep spewing hyperbole about a product you really don't know much about?

REALLY! You have such insight. Next you'll tell me they dont make their own processors. Please tell us more captain obvious. Apple works with many vendors to get what they want. They have that power. They can as for a lower nanometer chip that has less heat.

And I didn't leave anything out. I mentioned the transceivers. The switch ports they go into are not that bad at all.

Go fish.

 

[jasonmvp]

Apple works with many vendors to get what they want. They have that power. They can as for a lower nanometer chip that has less heat.

Which vendor would make that chip for them? Name them, please. I'd love to know who can do that, as it's not something even the big data center switch guys have solved. They "solve" it by having fans blow at 100% constantly, to keep the entire switch cool. And the main generator of said heat are the controller chips!

So, help the entire industry out here! Just think of how efficient all these rack-mounted 10GBASE-T switches could become if they just had your <brand X> controller chips in them!

Oh, wait a sec, never mind. There's a reason they don't exist.

And I didn't leave anything out. I mentioned the transceivers. The switch ports they go into are not that bad at all.

Define "not that bad at all" please. Which switches can a 25G, 40G, 100G, etc optic go into? Help a guy out here, because clearly I've been looking in the wrong areas for switches. And let me remind you that you still haven't told me what speed interfaces you'd propose for the Mac, either.

Here's some guidance for you: You don't understand what you're talking about when it comes to networking. You "heard" something or "read" something somewhere and you mistakenly believe you have all the answers on this topic. As it turns out: you don't have any of them. That's called "ignorance" which isn't a bad thing. It merely means, "not knowing something" and can be fixed with education. What is a bad thing is when someone who is educated on the topic and does this sort of thing for a living (and has for over a quarter of a century) tries to explain why you're mistaken, you continue to argue.

That's called "stupidity", or the act of protecting one's ignorance by not learning.

 

[ZombiePhysicist]

Which vendor would make that chip for them? Name them, please. I'd love to know who can do that, as it's not something even the big data center switch guys have solved. They "solve" it by having fans blow at 100% constantly, to keep the entire switch cool. And the main generator of said heat are the controller chips!

So, help the entire industry out here! Just think of how efficient all these rack-mounted 10GBASE-T switches could become if they just had your <brand X> controller chips in them!

Oh, wait a sec, never mind. There's a reason they don't exist.



Define "not that bad at all" please. Which switches can a 25G, 40G, 100G, etc optic go into? Help a guy out here, because clearly I've been looking in the wrong areas for switches. And let me remind you that you still haven't told me what speed interfaces you'd propose for the Mac, either.

Here's some guidance for you: You don't understand what you're talking about when it comes to networking. You "heard" something or "read" something somewhere and you mistakenly believe you have all the answers on this topic. As it turns out: you don't have any of them. That's called "ignorance" which isn't a bad thing. It merely means, "not knowing something" and can be fixed with education. What is a bad thing is when someone who is educated on the topic and does this sort of thing for a living (and has for over a quarter of a century) tries to explain why you're mistaken, you continue to argue.

That's called "stupidity", or the act of protecting one's ignorance by not learning.

First of all youre obnoxious as all out. You make demands like a petulant child. Watch your tone man. It's beyond repellant. Furthermore, doctor heal thyself.

As for what vendor. Emulex for example uses some Broadcom and custom Asic work. Apple could bring those to TMSC or some other fab with it's weight and get better fab chipsets.

As for prices, dont move the goalposts. I'm not talking about switches. I'm talking about supplying a port on a machine. You can get 40GB NICs for like 175-225 (1 or 2 port). Not terrible.

EMULEX OCE14401-NX Refurbished

Buy OCE14401-NX Emulex Oneconnect 40Gigabit Ethernet Card 40GB Enet 1Port Qsfp+ Pcie3.0 X8 40GB/S Nic No Optics DAC. Refurbished. Buy Online. Free Ground Shipping.

www.allhdd.com

MELLANOX MCX313A-BCCT CONNECTX-3 PRO SINGLE-PORT ADAPTER,PCI EXPRESS X8. BULK. IN STOCK.

Supplying computer parts on Demand. Trusted Source for all your Hardware needs. We ensure Quality Products with Warranty. Free ground shipping- AA2ZSupply.com

www.aa2zsupply.com

Obviously something like a 100/400gbps switch is new and crazy expensive:

FS N8500-32C 32-Port 100Gb QSFP28 L3 Data Center Managed Ethernet Switch, Bare-Metal Hardware - FS

Buy FS N8500-32C 32 Port 100Gb QSFP28 managed Spine/Core Ethernet network switch with Bare-Metal Hardware at best price. FS Fiber Optic 100 Gigebit Switches are Layer 2 / Layer 3 aggregation and core layer Switches.

www.fs.com

FS N9500-32D 32-Port 400Gb QSFP-DD L3 Tomahawk Data Center Managed Ethernet Switch, Bare-Metal Hardware - FS

FS N9500-32D 32 Port 400G Switch caters for industry leading 400G rate and future development trend. The 400Gb Switch with rich choice of port speed and density including 25G, 40G, 50G, 100G, 200G and 400G enabling consistent network architectures ensures exceptional reliability, serviceability...

www.fs.com

Not that it works like this as you'd still need controller/chipset, but that comes to around 300 a port which would be pretty crazy pricey for now.

 

[jasonmvp]

First of all youre obnoxious as all out. You make demands like a petulant child. Watch your tone man. It's beyond repellant. Furthermore, doctor heal thyself.

Watch my tone? Sit down, kid. You don't get to tell me to "watch my tone".

As for prices, dont move the goalposts. I'm not talking about switches. I'm talking about supplying a port on a machine. You can get 40GB NICs for like 175-225 (1 or 2 port). Not terrible.

But you HAVE to talk about switches. Because if you put a faster Ethernet port in the Mac, what's it supposed to connect to? Basically, you're maneuvering the buyer of the Mac into:

• paying for the ports on the Mac
• paying for the optics or DACs that connect to those ports
• PAYING FOR SOMETHING THAT CONNECTS ON THE OTHER END!

10GBASE-T is currently the most flexible option for higher speed connections, and they easily autoneg down to GigE. Your solution would not provide that level of flexibility and would force end users to have very expensive switches in their offices/studios/etc. Switches that make a lot of fricken noise!

Keep trying.