JasonB (no@spam.com) on 5/14/07 wrote:
---------------------------
>It's not the kernel having the problem, it's the application not being able to
>use the full 2GB of physical (and virtual) address space that should be available
>to it, and one of the big causes of that is virtual address space fragmentation.
>
>The point is that it starts to be an issue way before the theoretical 4GB
>limitation of 32 bit computers. Anything that attempts to use much more than half
>of a 2GB system's memory needs to take care to avoid having problems, and a memory
>footprint of "more than 1GB" is a lot less unusual for a modern desktop app than "4GB".

I agree with you 100%. If you have a single application requiring about 2GB of virtual space, it clearly calls for 64-bit. It actually has nothing to with RAM amount. I think you are perfectly justified to beat you customers on the head until they use a 64-bit OS.

I am thinking about the general case of 32-bit Windows having to deal with > 1 GB RAM present. I explained in my previous post that it doesn't seem that Windows is constrained in the same way that Linux is - there shouldn't be additional overhead in handling 2-3GB RAM, compared to 1GB RAM. Linus'es comment, while undoubtedly true, is perhaps not completely general as far as Windows is concerned.

(Of course this is at the expense of complexity - for an ex-Windows kernel developer the Linux kernel seems like a revelation)

Consider the case when the user is running many applications but none of them require more than a GB of _virtual space_, let alone RAM. 2 GB of RAM should serve this case nicely.

Are we technically justified in requiring a 64-bit OS for that? Linus says yes - I tend to agree that the comparative simplicity of the Linux design is worth it. Still, it is interesting to discuss the alternatives. It is a pity that Dave Cuttler doesn't post here regularly :-)

regards,
Tzvetan