I'm just tired of mysterious crashes that happen on all my Intel-based Macs (with NO apps installed on the Mac.) Yes, I've reinstalled the OS many times. Yes I checked the logs. Yes, I took my Macs to the Apple Genius Bars for diagnostic troubleshooting and they found nothing wrong. So I'm eager to try Apple Silicon Macs.
No one's going to comment on this absurd post?
I think I’d be an Apple hater if I owned multiple Macs, used them for absolutely nothing since I installed no application and yet they still crash all the time! To add insult to injury no one can help with this “problem.” And yet I’m eager to buy the latest Macs! Perhaps they can not crash while doing nothing (very quickly)!
I am guessing the reason egpu is not supported is related to drivers. Apple is not shipping AMD drivers for Apple Silicon and doesn’t use a default firmware that speaks VGA etc. with proper drivers added they would probably work. Just a guess.
I'm guessing it has more to do with limited PCI throughput, which was done to keep these systems from cannibalizing sales of higher end models. Because they are so damned performant, this could be an obvious problem for Apple.
I've been unable on my M1 Mac, using Terminal, to launch additional iterations of an application so far. Has anyone else tried this? I wonder if this is a security level increase that can be circumvented if required?
"Now the RAM is what Apple calls unified memory ..."
Unified memory was invented and named by Nvidia - back in 2013 - and is a widely known and used technology. So their options for calling it something else were a bit constrained.
Not sure why Nvidia would claim to have invented or named a technology that's been around since the last century, but I first saw Unified Memory Architecture when SGI introduced it in the O2 workstation back in 1996.
Perhaps they coined the term “Unified Memory Architecture” then, but microcomputers (including every single Apple 2 model in 70’s on, and earlier Macs, and many other microcomputers from the era) shared the same physical RAM and memory-mapped address space shared by video and whatever else.
It was primarily done to simplify system design and make it cheaper, but the 6502 would have also lost performance doing it any other way because it didn’t have special I/O-oriented instructions anyway.
It did cause constraints on system speed because RAM needed to be fast enough to always support video output with no delays while also not slowing down memory access by the main CPU: some machines of the era had performance e notably faster if video output was disabled as a result. The Apple 2 series was also designed this way to ensure accessing video memory as required to output video would result in DRAM refresh as a desired side-effect: Steve Wozniak took great pride in optimizing for lowest-cost designs. The Apple disk controller hardware had only a few cheap chips on it and the main CPU did all the real grunt work and was pretty much fully consumed by reading in sectors off the disk on the Apple drives, whereas the Commodore 64 drive had its own 6502 variant that took care of that, but STILL managed to get far worse performance, and it was all for nought: if a computer with that little memory needs something off a disk, you aren’t likely doing anything with the CPU while waiting because it’s out of data/code to work with!
Note: these were the machines I had access to and sufficient knowledge of: I’d not be surprised if earlier machines also had the same general design.
I've been unable on my M1 Mac, using Terminal, to launch additional iterations of an application so far. Has anyone else tried this? I wonder if this is a security level increase that can be circumvented if required?
Waiting for my M1 Mac Mini to arrive (likely around Christmas due to BTO configuration) but something to be aware of is not all applications are meant/designed to have more than a single instance of it running, so there is that, without having any information on which application you’re trying to launch.
An interesting example to consider is Pages: you can see using top or Activity Monitor that you can launch several instances of it, but only one instance appears to show a GUI.
I've been unable on my M1 Mac, using Terminal, to launch additional iterations of an application so far. Has anyone else tried this? I wonder if this is a security level increase that can be circumvented if required?
Waiting for my M1 Mac Mini to arrive (likely around Christmas due to BTO configuration) but something to be aware of is not all applications are meant/designed to have more than a single instance of it running, so there is that, without having any information on which application you’re trying to launch.
An interesting example to consider is Pages: you can see using top or Activity Monitor that you can launch several instances of it, but only one instance appears to show a GUI.
Yay, it works using a shell script on an ARM native app and an Intel app under Rosetta. I'd previously just dragged an alias out of the app's Package Contents/Contents/MacOS. That no longer works with Big Sur on Intel or ARM.
Here is the script to use in Script Editor for anyone interested. Compile, save, and run.
do shell script "open -n /Applications/YourApp.app"
I've not done a search but having read the review I was thinking are there no Mac apps which can use external RAM as there are for Windows? RAM doesn't need to be the fastest for hungry apps, it just needs to be there to prevent a calamity. Likewise for storage, one of those thunderbolt Ports could be used for additional RAM or data storage?
I've not done a search but having read the review I was thinking are there no Mac apps which can use external RAM as there are for Windows? RAM doesn't need to be the fastest for hungry apps, it just needs to be there to prevent a calamity. Likewise for storage, one of those thunderbolt Ports could be used for additional RAM or data storage?
Don’t know what you mean by external RAM in Windows. Seems like a bit of confusion on what you mean?
macOS uses a virtual memory architecture, basically like every other modern OS shipping these da ys. So, when a process actually needs memory, it will be allocated RAM by the OS. It will compress memory in RAM used from other apps that aren’t in use. If more memory is needed than what is available with RAM, the least used, lowest priority processes will have their memory written onto to storage (demand paging) so that more RAM can be allocated to the most used highest priority processes.
For storage, yes, absolutely. The M1 Macs support 40 Gbit/s Thunderbolt transfers and you should be able to see 3 GByte/s read and write speeds from a variety of SSDs and RAIDs.
Comments
I think I’d be an Apple hater if I owned multiple Macs, used them for absolutely nothing since I installed no application and yet they still crash all the time! To add insult to injury no one can help with this “problem.” And yet I’m eager to buy the latest Macs! Perhaps they can not crash while doing nothing (very quickly)!
I'm guessing it has more to do with limited PCI throughput, which was done to keep these systems from cannibalizing sales of higher end models. Because they are so damned performant, this could be an obvious problem for Apple.
It was primarily done to simplify system design and make it cheaper, but the 6502 would have also lost performance doing it any other way because it didn’t have special I/O-oriented instructions anyway.
It did cause constraints on system speed because RAM needed to be fast enough to always support video output with no delays while also not slowing down memory access by the main CPU: some machines of the era had performance e notably faster if video output was disabled as a result. The Apple 2 series was also designed this way to ensure accessing video memory as required to output video would result in DRAM refresh as a desired side-effect: Steve Wozniak took great pride in optimizing for lowest-cost designs. The Apple disk controller hardware had only a few cheap chips on it and the main CPU did all the real grunt work and was pretty much fully consumed by reading in sectors off the disk on the Apple drives, whereas the Commodore 64 drive had its own 6502 variant that took care of that, but STILL managed to get far worse performance, and it was all for nought: if a computer with that little memory needs something off a disk, you aren’t likely doing anything with the CPU while waiting because it’s out of data/code to work with!
Note: these were the machines I had access to and sufficient knowledge of: I’d not be surprised if earlier machines also had the same general design.
An interesting example to consider is Pages: you can see using top or Activity Monitor that you can launch several instances of it, but only one instance appears to show a GUI.
Here is the script to use in Script Editor for anyone interested. Compile, save, and run.
do shell script "open -n /Applications/YourApp.app"
ys. So, when a process actually needs memory, it will be allocated RAM by the OS. It will compress memory in RAM used from other apps that aren’t in use. If more memory is needed than what is available with RAM, the least used, lowest priority processes will have their memory written onto to storage (demand paging) so that more RAM can be allocated to the most used highest priority processes.