[openbeos] Re: memory beyond 1 gig
- From: "Jean-Baptiste Queru" <jb@xxxxxxxxx>
- To: <openbeos@xxxxxxxxxxxxx>
- Date: Sat, 8 Sep 2001 10:22:48 -0700
> Maybe you are not aware of to whom you speak.
> JBQ was an employee of Be.
> An excellent engineer and a person who has a very advanced
> knowledge of kernel engineering and issues.
Well, that's actually very funny. The person who *really* has very
advanced knowledge of kernel engineering is Manuel himself. One
of the reasons why I learnt quite some kernel engineering is that
his office was next to mine in Menlo Park and that he would always
have answers to the twisted questions I could have about kernel
engineering...
> It is a huge bug. I am learning what caused this right now, as I look at
the
> kernel issues. And it is a bug. No question. Now - Be knew about it and
probably
> didn't care because up until the last year or so memory was not cheap. A
> gig was beyond the realm of most people.
I have to agree with Manuel here, though. It's not a bug. It's the
unfortunate
consequence of a design decision. The code we're ralking about originates
back to a time when servers were running with 16MB of RAM. The design
was good enough to scale up to the point where it is efficient until 128 to
256MB
of RAM, as still works with 1GB of RAM. Still, I still say that the design
decision
was the right one. To put things back in perspective, it is akin to writing
code now
which will work efficiently on machines with 64GB of RAM, and which will
start
having issues with 256GB of RAM, sometime around 2010. Sounds quite far off?
> That is not true. Very simply, Be chose a physical memory model that has
limitations.
No. Be did not choose the physical memory model. In order (considering the
machines Be's OS has been running on), those had been chosen by AT&T,
Motorola, and Intel. Be chose a kernel architecture policy. And that policy
has its strong positive points, and its strong negative points.
> >This full paragraph makes me think that you don't fully understand VM
> >management.
>
> Trust me - you are wrong here.
When it comes to BeOS kernel architecture (or to kernel architecture in
general), telling Manuel that he is wrong will make anyone who knows
him burst out in a really big laugh.
--jbq
Other related posts:
