Switching is a terrible hardware restriction leading to shared bandwidth.
Apple Silicon Mac Pro uses PCIe switches to support all slots
71–80 of 85 posts
Re: Apple Silicon Mac Pro uses PCIe switches to support all slots
#72Earlier quoted context omitted.
The new Mac Pro can be rack mounted. It has PCIe gen 4 for audio and video I/O, networking, and storage cards.
I'd be sceptical about "can be rack mounted" in the context of Apple hardware. Remember those? https://www.sonnettech.com/product/legacyproducts/rackmacpro... https://www.myelectronics.nl/us/mac-studio-empower-station-r... At least the new case is available with rails, but cooling and management are still crap.
I don't think apple ever officially claimed the trashcan mac pro could be rack mounted, those are kits provided by 3rd parties.
But the 2019 mac pro did have an official rack mount variant of the case.
Re: Apple Silicon Mac Pro uses PCIe switches to support all slots
#73Earlier quoted context omitted.
Why wouldn't they? If a Sun Blade 100 is a workstation and I can do the same tasks faster on a Mac mini or Mac Pro why wouldn't they be workstations. Don't get too hung up on labels they can distract from what is front of you. If Apple Silicon actually is an entire motherboard shrunk to die size the shrunken die motherboard will be faster. Place that chip onto a larger motherboard and you get real efficiencies that a…
Well, workstation implies a lot. And the price. It's not just a marketing term. It implies more PCIe lanes than a normal desktop (which in this case is not being realised) and better graphics capability for intensive tasks. The latter is up for debate. I'd imagine a 4090 equipped HEDT still kicks its ass in most workstation-type workloads. The unified memory is great for LLM style tasks of course and it's one of the…
I think you are making up your own yardstick and claiming it’s universally accepted.
One instance is that PCI lanes didn't exist when the term was conjured up therefore can’t be part of inherent definition.
The mac pro has 64 lanes acroos 6 slots. PCI-e external expansion is another 8 ports of expansion limited to 40 GB/s. Unfortunately the concept of lanes doesn't seem to exist for external PCI-e ports so its hard to judge if those would count as more lanes than a normal desktop.
The 20 year old workstation is to demonstrate that the slab form factor has been called a workstation from the beginning.
Re: Apple Silicon Mac Pro uses PCIe switches to support all slots
#74Switching is a terrible hardware restriction leading to shared bandwidth.
Unless you want to do device to device transfers, in which case you can move it inside the switch. You could for example transfer texture data from an SSD straight to GPU memory, a la [NVidia's GPUDirect](https://developer.nvidia.com/blog/gpudirect-storage/)
Re: Apple Silicon Mac Pro uses PCIe switches to support all slots
#75This could be bad or great depending on the use case.
- Bad: bottlenecked access to the CPU and its memory
- Good: Bypass the CPU on device to device transfer, in the style of NVidia's GPUDirect (https://developer.nvidia.com/blog/gpudirect-storage/)
Re: Apple Silicon Mac Pro uses PCIe switches to support all slots
#76Earlier quoted context omitted.
Well, workstation implies a lot. And the price. It's not just a marketing term. It implies more PCIe lanes than a normal desktop (which in this case is not being realised) and better graphics capability for intensive tasks. The latter is up for debate. I'd imagine a 4090 equipped HEDT still kicks its ass in most workstation-type workloads. The unified memory is great for LLM style tasks of course and it's one of the…
I think you are making up your own yardstick and claiming it’s universally accepted. One instance is that PCI lanes didn't exist when the term was conjured up therefore can’t be part of inherent definition. The mac pro has 64 lanes acroos 6 slots. PCI-e external expansion is another 8 ports of expansion limited to 40 GB/s. Unfortunately the concept of lanes doesn't seem to exist for external PCI-e ports so its hard t…
It's because the term itself has changed. Not every PC is a workstation. These days we call it that if it has at least HEDT or server hardware. More PCIe lanes than desktops comes with that. Even Apple's previous Mac Pros have been matching that definition.
Re: Apple Silicon Mac Pro uses PCIe switches to support all slots
#77Earlier quoted context omitted.
I think you are making up your own yardstick and claiming it’s universally accepted. One instance is that PCI lanes didn't exist when the term was conjured up therefore can’t be part of inherent definition. The mac pro has 64 lanes acroos 6 slots. PCI-e external expansion is another 8 ports of expansion limited to 40 GB/s. Unfortunately the concept of lanes doesn't seem to exist for external PCI-e ports so its hard t…
It's because the term itself has changed. Not every PC is a workstation. These days we call it that if it has at least HEDT or server hardware. More PCIe lanes than desktops comes with that. Even Apple's previous Mac Pros have been matching that definition.
Which standard body is that we?
It's has always been a subjective term and I believe it has more to do with output than consumption of resources during a compute cycle. There is no reason to try to dispute some claim to the term.
Re: Apple Silicon Mac Pro uses PCIe switches to support all slots
#78Earlier quoted context omitted.
The defining characteristic of PCIe isn't faster bus timings. It is that PCIe isn't even a "bus" in the traditional sense. With a bus, you have a number of shared physical wires for addresses and data, and an arbitration system that assigns those shared wires to one device at a time. The available bandwidth is just the bandwidth of the bus, divided into the time you can reserve it for yourself. PCIe is a tree-shaped…
That sounds like a good description but do you think the term serial should be worked in there?
On some level yes, on some level no. The physical interface for each point-to-point link is serial, but one device can have several of those links and use them (send/receive packets on them) in parallel.
Re: Apple Silicon Mac Pro uses PCIe switches to support all slots
#79Earlier quoted context omitted.
Fault-tolerant systems with five nines uptime requirements.
How many people would run such fault-tolerant systems on Mac workstations?
None, you also need a fault-tolerant OS like Stratus VOS, but the question was not limited to Macs. I doubt they will sell many Mac Pros, the Studio’s rendered it irrelevant.
Re: Apple Silicon Mac Pro uses PCIe switches to support all slots
#80Earlier quoted context omitted.
That sounds like a good description but do you think the term serial should be worked in there?
On some level yes, on some level no. The physical interface for each point-to-point link is serial, but one device can have several of those links and use them (send/receive packets on them) in parallel.
Does the lane count correspond to the quantity of parallel links? would the switches be single transistors?