<div dir="ltr">thanks, Michael. <br>The boxens have 32G RAM each, thus should be all within reach on RHEL 5.1/x86_64. <br><br><div class="gmail_quote">2008/7/28 Michael B. Trausch <span dir="ltr"><<a href="mailto:mike@trausch.us">mike@trausch.us</a>></span><br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div class="Ih2E3d">On Mon, 2008-07-28 at 15:46 -0600, JK wrote:<br>
> Do the 64-bit archs support bigger pages than 4MiB? Even on x86-32<br>
> with PAE, a fully-decked-out (64G) server was using a huge chunk of<br>
> kernelspace just for pagetables and the physical page bookkeeping.<br>
> I'd expect that would definitely have to change when squaring the<br>
> address space.<br>
<br>
</div>On the IA-32, there are 4 KiB and 4 MiB page sizes [1]; on the x86-64,<br>
in Long Mode, you can have page size that are 4 KiB, 2 MiB, or 1 GiB<br>
[2]. Not sure why 4 MiB seems to be missing from the x86-64<br>
architecture, but it is supported in Legacy Mode and probably<br>
Compatibility Mode, I should hope, or 32-bit software that uses a 4 MiB<br>
page size wouldn't run. :)<br>
<br>
--- Mike<br>
<br>
[1] <a href="http://www.cse.unsw.edu.au/%7Ecs9244/06/seminars/06-nfd.pdf" target="_blank">http://www.cse.unsw.edu.au/~cs9244/06/seminars/06-nfd.pdf</a><br>
[2] <a href="http://en.wikipedia.org/wiki/X86-64#Virtual_address_space_details" target="_blank">http://en.wikipedia.org/wiki/X86-64#Virtual_address_space_details</a><br>
<div><div></div><div class="Wj3C7c"><br>
--<br>
My sigfile ran away and is on hiatus.<br>
</div></div><br>_______________________________________________<br>
Ale mailing list<br>
<a href="mailto:Ale@ale.org">Ale@ale.org</a><br>
<a href="http://mail.ale.org/mailman/listinfo/ale" target="_blank">http://mail.ale.org/mailman/listinfo/ale</a><br>
<br></blockquote></div><br></div>