- Question: Q: What is application memory?
- All replies
- Mac Pro (2019) memory specifications
- Memory requirements
- Model specifications
- With M1 Macs, memory just isn’t what it used to be
- Part of the package
- Benefits of being unified
- Do you need it?
- What’s next?
- What is Unified Memory and how does it work on Apple Silicon?
- What is RAM and how is the M1 different here?
- What is Unified Memory and how does it work?
Question: Q: What is application memory?
Posted on Mar 31, 2018 2:38 PM
All replies
Loading page content
Page content loaded
It is the memory required by an app to run. How much RAM do you have? How much free drive space do you have? Try rebooting the computer. You can run activity monitor to see what is using memory.
Mar 31, 2018 4:15 PM
You get this message for 2 reasons
A) Your boot disk has run out of space, and macOS can no longer create /var/vm/swapfile(s). macOS will create swap files when the applications you are concurrently running need more memory than your Mac has RAM to satisfy. Swapfiles take up disk space, so if your disk is full, you can get this message.
B) You have lots of free disk space (let’s say 50 gigabytes or more of free disk space), but the set of applications you are concurrently running what a HUGE amount of virtual memory address space and macOS has to create so many page tables entries to keep track of all the virtual memory for all the apps currently running, that macOS would use so much RAM it would not have enough available for applications.
Each page of virtual memory address space requires a page table entry. It does not matter if that page is being used to hold program code, application frameworks, shared libraries, user data, memory mapped files, or other virtual memory objects. It also does not matter if the contents of that page are currently in RAM, in a swapfile, or in a regular file on disk. They all require that macOS maintain page tables with tracking information for every one of those pages.
Generally, when ‘B’ occurs, there are one or more applications that have what is called a virtual memory leak. They keep asking the operating system for more virtual address space, use it, then forget to give the address space back, and proceed to ask for a new range of virtual addresses. Wash, Rinse, Repeat until macOS can no longer manage all the virtual memory page tables needed to keep track of all the virtual memory all the currently running applications are using.
Quitting an application will free up all the virtual memory it is currently using, including any it lost track of due to leaks. macOS did not lose it and knows how to give it back when the App quits. It is the app that has lost track.
Источник
Mac Pro (2019) memory specifications
Learn about the types of memory (RAM) used in Mac Pro (2019).
Mac Pro has a 6-channel memory controller that supports 12 memory slots you can use to install up to 1.5TB of memory using 2933MHz DDR4 ECC compliant memory DIMMs. To ensure compatibility, Apple recommends that you use Apple-approved memory. You can purchase Apple-approved memory online from the Apple Store or by visiting an Apple Retail Store or Apple Authorized Reseller.
Memory requirements
When you upgrade the memory in your Mac Pro, it must meet these requirements:
- 2933MHz DDR4 full-length DIMM. Mixing memory speeds is not recommended.
- Error-correcting code (ECC)
- 8GB, 16GB, 32GB, 64GB, or 128GB DIMMs
- Registered (R-DIMM) or Load-Reduced DIMM (LR-DIMM). Do not mix R-DIMMs and LR-DIMMs.
- 288-pin
- Use the same size memory modules across all slots to maximize performance.
- DIMMs with heatsinks are not supported and may damage the DIMM mechanism.
The Mac Pro supports both R-DIMMs and LR-DIMMs, but you can’t mix these two memory types. Mixing R-DIMMs and LR-DIMMs will result in a memory error flash of the status indicator light when you turn on your computer.
Model specifications
Mac Pro model | Max RAM | Speed | Type |
---|---|---|---|
8-core | 768GB | 2666MHz DDR4 ECC | LR-DIMM or R-DIMM |
12-core | 768GB | 2933MHz DDR4 ECC | LR-DIMM or R-DIMM |
16-core | 768GB | 2933MHz DDR4 ECC | LR-DIMM or R-DIMM |
24-core | 1.5TB | 2933MHz DDR4 ECC | LR-DIMM or R-DIMM |
28-core | 1.5TB | 2933MHz DDR4 ECC | LR-DIMM or R-DIMM |
Configured from Apple, Mac Pro models include the following DIMM types:
- R-DIMMs for configurations up to 192GB (6x32GB).
- LR-DIMMs for configurations of 384GB (6x64GB) or more.
All memory included with your Mac Pro is 2933MHz. Mac Pro models with 8 core processors operate memory at 2666MHz.
Источник
With M1 Macs, memory just isn’t what it used to be
The first Macs powered by Apple-designed processors are finally here. And from the outside, they’re almost dead ringers for the Intel-based Macs they’re replacing.
But on the inside, they’re not like other computers. Apple has brought its approach to system design, learned through years of iteration on the iPhone and iPad, to the Mac for the first time.
Those of us who are used to thinking of personal computers in certain terms are going to need to adjust to this new reality. It’s a world in which Apple sells three different Mac models without even disclosing the clock speed of the processor inside. (It doesn’t do it for the iPhone or iPad, after all.)
But perhaps the item on the spec sheet that will require the biggest diversion from the old way of thinking is system memory. It’s a feature that’s already frequently misunderstood (and frequently confused with storage size), and now Macs with Apple silicon are using it in an entirely different way.
The old way of thinking of RAM is dead. Welcome to the world of the Unified Memory Architecture.
Part of the package
Like Intel chips with integrated graphics, the M1 chip includes a graphics processor, and system memory is shared by both processor cores and graphics cores. (And also, in the M1’s case, the cores that make up the Neural Engine.) But in shifting its terminology to describe a unified memory architecture, Apple’s trying to point out that the M1’s approach is a bit different.
The biggest difference is that in the M1, the memory is a part of the M1 architecture itself. There’s no memory slot or slots on the motherboard of an M1 Mac, nor is there an area where a memory chip has been permanently soldered on. Instead, the memory is integrated into the same package that contains the M1 itself.
What this means is that when you buy an M1-based Mac and choose a memory configuration, that’s it. There have been many other Macs with soldered-on memory that couldn’t be upgraded, but this is a little different, since the memory is basically part of the M1 package itself.
The first round of Apple silicon Macs, which include the Mac mini, can have a maximum of 16GB of memory. That could be by design.
Looking at the first round of M1 Macs, it seems that the M1 is only capable of using 8GB or 16GB of memory. That may not actually be a hard limit—perhaps Apple is holding back in order to limit these low-end systems. But it’s more likely that we won’t see Macs running Apple silicon with more than 16GB until Apple provides a higher-end variant of the M1.
Benefits of being unified
But Apple isn’t integrating memory into its systems-on-a-chip out of spite. It’s doing it because it’s an approach that can lead to some dramatic speed benefits.
The M1 processor’s memory is a single pool that’s accessible by any portion of the processor. If the system needs more memory for graphics, it can allocate that. If it needs more memory for the Neural Engine, likewise. Even better, because all the aspects of the processor can access all of the system memory, there’s no performance hit when the graphics cores need to access something that was previously being accessed by a processor core. On other systems, the data has to be copied from one portion of memory to another—but on the M1, it’s just instantly accessible.
These new Macs are, in their way, kind of alien. The tradition in personal computers was that everything was modular, an outgrowth of the early PC era. Even though the Mac never participated in the build-a-PC ethos, the parts Apple used to assemble Macs came from that industry. Compare that with the smartphone, where Apple has continued to integrate more portions of the system into its single processor package in order to increase efficiency. These new Macs are far more like smartphones than like traditional PCs.
Do you need it?
One of the biggest criticisms I’ve seen about this first round of M1 Macs has been that they just don’t offer enough memory, maxing out at 16GB. Keeping in mind that these are the lowest-powered Mac models, it’s likely that future models will offer more RAM options.
But it’s also worth considering just how squishy the need for more memory can be when you poke at it. Sure, a lot of people feel they need it—but do they, really?
Yes, when a Mac runs out of physical memory, it will page the contents of memory to disk—and even super-fast SSDs are slower than main memory! Though the speed differences are a lot less than back when we used slow spinning disk drives.
An Apple silicon Mac Pro with only 16GB of memory? It won’t happen. You’ll see Apple silicon with more RAM and processing power in the near future.
What would cause your Mac to run out of physical memory? If you leave an awful lot of apps open at once, or if your browser has hundreds of tabs open, or if you’re using an app that loads a very large file (like, say, a Photoshop file) into memory. If you’re someone who does this a lot, you probably want more memory…. but then again, if you’re someone who does this a lot, you might not want to buy one an M1 Mac right now. The mid-range and high-end models that will undoubtedly offer more RAM options and more processor power are undoubtedly coming next year.
But if you combine the efficiency of the unified memory architecture with the speed of SSD storage, and consider most everyday use cases, I’m pretty sure that most regular users could get by with 8GB of unified memory—or, if you want to be absolutely sure, upgrade that to 16GB. (I did.)
What’s next?
I can’t imagine 2021 passing without Apple rolling out a new set of Macs with more powerful processors and more memory options. The high-end MacBook Pros and the iMacs, at the very least, could use updates that provide some options beyond the basic M1.
In the long run, is it possible that Apple would build systems with external graphics processors with their own dedicated memory? It seems inevitable, at least at the high end—what’s a Mac Pro for if you can’t stick a ridiculous graphics card in it?
But Apple is also very likely to just keep scaling up memory options as it scales up its processors, adding more memory as it adds cores—and those chips are the ones likely to be offered in most Mac models.
The unified memory architecture in the M1 is one of the reasons these Macs are so amazingly fast—but all Mac users are going to have to relinquish some of our assumptions about how our computers work, and how they’re configured. And if you really can’t bear buying any Mac with only 16GB of RAM, don’t get mad—be patient. More Apple silicon Macs are on the way.
Источник
What is Unified Memory and how does it work on Apple Silicon?
With the release of Apple’s new M1-powered iPad Pro and 24-inch iMac, there’s renewed interest in the amazing efficiency of the M1 chip. The launch of the M1 chip brought Apple’s first use of unified memory architecture (UMA) on Apple silicon. This new approach to memory enables Apple to squeeze out higher performance from less total RAM. So how exactly does unified memory on Apple Silicon really work? Let’s take a look, starting with a few basics about memory in general and how things are new with the M1 Mac design.
What is RAM and how is the M1 different here?
RAM stands for “Random Access Memory”. This is the main component of the system memory on any computer. System memory provides a temporary repository for data used by your computer at a given instant. Data stored in system memory can include files you’re currently viewing, as well as files needed by macOS. Traditionally, RAM exists physically as a long stick that fits into a slot on your motherboard. The M1 is actually a revolution in this way as well.
Apple designed the M1 as a system on a chip (SoC), with the RAM included as part of this package. While integrating RAM with the SoC is common in smartphones such as the iPhone 12 series, this is a relatively new idea for desktop and laptop computers. Adding RAM to the SoC design enables faster access to memory, improving efficiency.
In addition to physically adding the RAM to the SoC, Apple has changed the fundamental way the system uses memory. This is where unified memory on Apple silicon comes into play.
What is Unified Memory and how does it work?
Unified memory is about minimizing the redundancy of data copied between different sections of memory used by the CPU, GPU, etc. Copying is slow and wastes memory capacity. With a traditional memory implementation, part of your RAM is reserved for the GPU. If your laptop is advertised with 16GB of RAM, and 2GB is allocated to the GPU, you only have 14GB available for system tasks. Apple solves this problem with UMA, making memory allocation more fluid and increasing performance.
Gaming provides the best example to understand the benefits of unified memory. When you play a game on your Mac, the CPU first receives all the instructions for the game and then pushes the data that the GPU needs to the graphics card. The graphics card then takes all that data and works on it within its own processor (the GPU) and built-in RAM.
If you have a processor with integrated graphics, the GPU still maintains its own chunk of memory, as does the processor. The CPU and GPU work on the same data independently and then passes the results back and forth between their memory repositories. If you drop the requirement to move data back and forth, it’s easy to see how keeping everything in the same storage area could improve performance. The unified memory approach truly revolutionizes performance by allowing all components access to the same memory at the same place.
Apple truly achieved greatness with the M1 SoC. In addition to integrating RAM physically, the new unified memory architecture allows more efficient use of available memory. Using this new memory implementation, the new M1 iMacs can do just about anything, including running Windows 10. Placing all memory in a single pool means that any component can ramp up usage when needed, seamlessly allocating resources where needed.
Источник