If the memory is byte addressable, shouldn't it be
2^64 / (2^10 * 2^10 * 2^10) = 2^34 GB
?
A general Google search says 8GB or 16GB max.
How is that true? Also, how does it depend on the Operating System?
Super User is a question and answer site for computer enthusiasts and power users. It only takes a minute to sign up.
Sign up to join this communityIf the memory is byte addressable, shouldn't it be
2^64 / (2^10 * 2^10 * 2^10) = 2^34 GB
?
A general Google search says 8GB or 16GB max.
How is that true? Also, how does it depend on the Operating System?
There are multiple interleaving factors.
First of all, you are currently unable to assemble a system that has 264 bytes (16 exibytes) of physical RAM.
Second, just because an architecture uses 64-bit pointers, doesn't mean that all the bits of those pointers are actually used. Notably, current x86-64 CPUs (aka AMD64 and Intel's current 64-bit chips) actually use 48-bit address lines (AMD64) and 42-bit address lines (Intel) (see http://en.wikipedia.org/wiki/X86_64#Virtual_address_space_details ), theoretically allowing 256 terabytes of physical RAM.
Second, motherboards have their own limits on how much RAM they can support, both physically and logically. Physically, there are only going to be so many slots available for RAM. As for the "logical" limits, I don't fully understand why this is still the case for x86-64 (the memory controller having long since been migrated into the CPU itself), but there it is. Presumably corners are being cut on address lines to save a few bucks in design and manufacturing.
Third, an operating system may have internal limitations as to how much RAM it can efficiently support. In part, this is actually to prevent needing overly large data structures to keep track of usage for memory that isn't really there. Last I checked, Linux allows 128TB of virtual address space per process on x86-64, and can theoretically support 64TB of physical RAM.
Fourth, some operating systems (e.g. Windows) will artificially limit how much RAM can be used as a tactic to make users upgrade to more expensive versions if they want more RAM (Windows 7 Starter is limited to 2GB, Home Basic to 8, Home Premium to 16, and Professional and above are 192GB, and Windows Server releases have far higher limits).
If you use MS Windows, the maximum memory is constrained by the operating system.
For Windows 7, the maximums are:
Windows Server 2008 ranges from 8GB for entry level to 2TB for the top of the line.
See http://msdn.microsoft.com/en-us/library/aa366778%28VS.85%29.aspx for a full-ish list.
These limits are primarily for marketing reasons. By restricting the cheaper versions they make the high-end versions (comparatively) more attractive.
I think that recent Linux builds work with around 32TB of RAM, although not all hardware can support that.
A 64-bit machine should be able to address up to 264 addressable units (in architectures designed over the last few decades, addressable units are invariably bytes, a.k.a. octets). If you define a "gigabyte" as 230 addressable units, then yes, 234 gigabytes would be another way to express the same count.
What Google search produced those ridiculous tiny results? Maybe the page you found was talking about the amounts of RAM you can buy in reasonably cheap boxes of a certain model -- you can most definitely have much more RAM (all addressable, of course) even in boxes that are far from being very costly.
Operating systems most definitely can and do restrict the amount of addressable memory (for example, they can do so by limiting the size of page tables, perhaps to some configurable maximum that can be set by recompiling or otherwise reconfiguring the OS -- it's hard to be more specific without focusing on some specific group of OSs).
You can have as much RAM as you'd like but it all depends on the operating system and the application you're running - as someone else stated.
If you're running an hungry CPU intensive application that's also intensive on other hardware like your hard disk then you best seriously consider a second CPU, in a completely separate machine. For example, running an database application and email server both on separate machines.
It doesn't matter how much memory you have - in today's market the whole system will grind to a halt. Single CPUs simply can't handle the multitasking the way you'd expect.
For a single machine with loads of memory it's better but not that brilliant if you have multiple applications running on the same system - regardless with the amount of memory installed. This is the very reason why big companies use not only 64bit architecture to address more memory but they also load balance the whole lot with two or more servers.
If you're really serious about performance then definitely go 64 bit but, also consider load balancing with a second server.
I've been down that road - configuring a single machine with maximum RAM. Yet when running multiple applications it still runs like a dog. That's because the CPU itself just can't handle it.
My advice is to save your bucks on memory and purchase a second server and then load balance the two. It's a lot simpler plus - once you get into the grips of it - it's easy to expand when you need more processing power and speed.
Also, a 64bit architecture machine can cope with 32bit software but with a 32 bit machine you're limited to only 32 bit or less applications - including the operating system.
Since the whole market is going 64bit in terms of both hardware and software I would advise to go 64bit for the long term but, forget the idea of squashing all onto one single machine with bucket loads of memory because you'll be disappointed.
If you are uncertain what "load balancing" is then google it.
I think it can take more than an exabyte "EB".
Look at this 32-bit processor can take 232 B = 4 GB, but at most it take 3 GB "because they are 1 GB for reversed addresses"
That means 64-bit processor can take 264 B = 16 EB, but I can't how addresses that locked for reversed used. so we can say 64-bit processor take more than 1 EB. and this specification does not depend of what the OS the PC used or what its model.
Taking an Intel or AMD CPU as an example, there are several limitations to both physical memory and virtual memory. Other people have talked about hardware pin-out and motherboard limitations, but there are also internal limitations.
A terminal page table entry in the MMU contains 52 significant bits of physical address and can theoretically map up to 4096 TB of physical RAM.
The MMU implements a 4-level page table whose top level consists of 512 × 512 GB entries. So virtual memory is limited to 256 TB.
Insofar as I know, all 64-bit operating systems implement a DMAP (a direct-map), which is super-convenient. However, because virtual memory is limited to 256 TB, the size of the DMAP is also limited. Typically 128 TB is reserved for user mode, and 128 TB for the kernel. Of that 128 TB for the kernel, up to half may be reserved for the DMAP. Due to the usefulness of the DMAP, operating systems will usually limit physical memory to what the DMAP can handle.
So most (all?) 64-bit operating systems for Intel/AMD the physical memory supported is limited to somewhere around 64 TB, leaving 64 TB for general kernel virtual memory and 128 TB for user virtual memory.
--
Internal to the CPU, the CPU caches also have a limitation on the number of physical address bits they store, since each additional address bit eats more transistors to the cache tag. DRAM controller pin limitations are not necessarily the final word, since memory can also be addressed through DMI links.
Different CPU architectures (e.g., ARM, PowerPC, etc.) will have different limitations.