[lkml]   [1997]   [Mar]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: Now, why didn't we think of this before?

Let's implement system calls via RPC while we are at it...

A few things:

1) threaded kernel is being worked on, we should be scaling nicely
on SMP as a result some time in the near future

2) re-entrant kernel, in the cases where it even makes sense to do
so we will get it for free via #1, if you want a pre-emptive kernel
that is another story all together

Memory may be cheap, however on chip cache is not. This is one of the
biggest arguments against persistant resizable storage, and why real
systems continue to be monolithic and written in C/Assembler.

People who want a Linux kernel in Java VM, go ahead be my guest and
implement such a thing. However I will never encourage anyone to use
it who happens to care about performance. I have enough trouble
getting every last cycle out of the kernel using C and Assembler.
(try to get 1200 web operations per second on a Java VM based kernel
using current processors, not gonna happen)

Eventually perhaps systems will be fast enough that a Java VM written
kernel would have even passable performance. It will be quite a
statement of how we can waste CPU cycles, the VLSI guys are
essentially wasting their time at such a point because the software
people are in a state of delirium. If we have left it any time
recently, this would put us right back into the software crisis
again that everyone moans about.

Yow! 11.26 MB/s remote host TCP bandwidth & ////
199 usec remote TCP latency over 100Mb/s ////
ethernet. Beat that! ////
-----------------------------------------////__________ o
David S. Miller, /_____________/ / // /_/ ><

 \ /
  Last update: 2005-03-22 13:39    [W:0.078 / U:14.868 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site