Re: PAGE 1k

Re: PAGE 1k

[User &]

Hi

Thanks people for the all informations.

Merry christmas.

Breno

RE: PAGE = 1k

[Gilad Ben-Yossef]

On Tue, 2002-12-24 at 00:12, Gilad Ben-Yossef wrote:
> If so, you might want to check out how the Millipede system from the
> Technion works - it basically creates several "views" of a page using
> segmentation and other MMU features which creates artifical locality of
> memory structures on the same page in units smaller then a hardware
> page. 

Bah... forgot the link:
http://www.cs.technion.ac.il/Labs/dsl/projects/millipede/default.htm

and you might have to dig through to find the actual place where they
discuss this particular technique. AFAIK they've tried many different
one and this ISN'T the one they're currently implementing...

-- 
Gilad Ben-Yossef <gilad@benyossef.com>
http://benyossef.com

 "Geeks rock bands cool name #8192: RAID against the machine"

RE: PAGE = 1k

[Gilad Ben-Yossef]

On Mon, 2002-12-23 at 18:52, User & wrote:


> yes i know ,it´s depend of hardware, but maybe we can do something as mmu
> with software or something like this , i don´t know.

No, but you can fake it ;-)

I'm taking a wild guess here that you're interested in this subject
because of distributed shared memory. That is, you're interested in
smaller page sizes because you want to be aware of changes done to
memory structures in sizes smaller then a page. 

If so, you might want to check out how the Millipede system from the
Technion works - it basically creates several "views" of a page using
segmentation and other MMU features which creates artifical locality of
memory structures on the same page in units smaller then a hardware
page. 

Hope this helps,
Gilad.

-- 
Gilad Ben-Yossef <gilad@benyossef.com>
http://benyossef.com

 "Geeks rock bands cool name #8192: RAID against the machine"

RE: PAGE = 1k

[Rik van Riel]

On Mon, 23 Dec 2002, User & wrote:

> yes i know ,it´s depend of hardware, but maybe we can do something as mmu
> with software or something like this , i don´t know.

No, the software cannot patch the hardware.

The MMU does address translations in 4 kB blocks and there's
nothing you can do about that.

Rik
-- 
Bravely reimplemented by the knights who say "NIH".
http://www.surriel.com/		http://guru.conectiva.com/
Current spamtrap:  <a href=mailto:"october@surriel.com">october@surriel.com</a>

RE: PAGE = 1k

[User &]

Sean
yes i know ,it´s depend of hardware, but maybe we can do something as mmu
with software or something like this , i don´t know.



Breno

Re: PAGE = 1k

[Sean Neakums]

commence  User & quotation:

> it´s possible to use pages of 1kb on x86 ?

To my knowledge, the page size used by Linux is dictated by the
hardware, which means that on x86, for the most part, Linux uses 4k
pages.

-- 
 /                          |
[|] Sean Neakums            |  Questions are a burden to others;
[|] <sneakums@zork.net>     |      answers a prison for oneself.
 \                          |

PAGE = 1k

[User &]

Hi people

it´s possible to use pages of 1kb on x86 ?

thanks
Breno