Roger Heflin wrote:
Howard Wilkinson wrote:
I am looking for a definitive answer to the question of where the PAE
kernels become useful. I have seen various articles that mention
needing PAE kernels if you have more then 4GB of physical memory in a
32-bit processor environment. I have also seen statements that say
you need them if you have 4GB or more of memory. Now which is right?
Also, even if you need a PAE kernel because the last few bytes are
not addressable when you have exactly 4GB is this useful or is the
trade off of larger page tables and pages going to eat any benefit of
being able to address these few bytes and if so when does the PAE
kernel become useful?
Howard.
It depends on the bios, you would have to try with and without the PAE
kernel and see if the amount of usable ram changes.
Some bioses won't remap any memory below 4GB (that is covered by
something else) to over 4GB, if your bios does not remap anything
above 4GB when you only have 4GB (or less) then PAE won't buy you
anything. And since often moving the covered memory from below 4GB,
sometimes means moving some non-covered memory and therefore lowering
the memory usable for an OS that does not support PAE-ie that other
OS, often the bios *WON'T* move the covered memory at all because it
would lower the usable memory below 4GB for that other OS.
This is almost always true on the Desktop class MB's, and it is
sometimes true on the higher end stuff also.
Roger
Hmmm, I was afraid that this would be the answer! So the "definitive"
answer is there is not one :-(
Given that some of the BIOS comes from ad-on cards and this is going to
affect each machine differently, then can the resulting map be different
with different cards, or even with the same cards but in different
slots, or even as in the bad old days of Windows 3.1 where the different
value of the MAC address on a network card meant a different memory layout.
The specific case that I have is 20 Intel SRMK4 machines all with 4GB of
physical memory. Currently all running Fedora 32-bit kernels. I have yet
to test the PAE kernels because I do not have any suitable benchmarks
but I will see what happens.
Thanks to all for the responses.
Howard.
NOTE: my idea of useful is that the benefits outweigh the costs and the
system gains overall. So a very complex calculation all round :-)
--
fedora-list mailing list
fedora-list@xxxxxxxxxx
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
