Loading ...
Sorry, an error occurred while loading the content.

Re: [nslu2-linux] Re: SlugOS/BE 2.5 Beta vs. Debian/NSLU2 4.0r1 Peformance

Expand Messages
  • Rod Whitby
    This is the difference between a generic desktop distribution (like Debian), and a targetted embedded distribution (like SlugOS). To get the same
    Message 1 of 5 , Dec 12, 2007
    • 0 Attachment
      This is the difference between a "generic" desktop distribution (like
      Debian), and a "targetted" embedded distribution (like SlugOS).

      To get the same performance, you would need to recompile the Debian
      kernel, libraries and all applications.

      ... or just use SlugOS and Optware ;-)

      -- Rod

      smitna68 wrote:
      > thanks a lot for your answer, now I understand it a little bit more.
      > But this fact makes me not really "happy"...
      > Is there a way to change this behavior (ex. patching the kernel)?
      >
      > Daniel
      >
      > --- In nslu2-linux@yahoogroups.com, Mike Westerhof <mwester@...> wrote:
      >> On Tue 11/12/07 12:03 AM , "smitna68" daniel@... sent:
      >>> --- In n
      >>> slu2-linux@yahoogroups.com, "smitna68" wrote:>
      >>>> Hello,
      >>>>
      >>>> I discovered a significant difference between these
      >>> "versions":>
      >>>> NSLU2, 2x "De-underclocked" device
      >>>>
      >>>> 1. SlugOS/BE 2.5 Beta formerly OpenSlug, Kernel
      >>> 2.6.12.2> 2. Debian/NSLU2 4.0r1 (installation
      >>> descripted> http://www.cyrius.com/debian/nslu2/unpack.html),
      >>> Kernel 2.6.18-5-ixp4xx>
      >>>> Mounts:
      >>>> 1.
      >>>> /dev/sda1 / ext2 defaults 1
      >>> 1> /dev/sda2 swap swap pri=42 0
      >>> 0> proc /proc proc defaults
      >>> 0 0> tmpfs /media/ram tmpfs defaults
      >>> 0 0> usbfs /proc/bus/usb usbfs defaults
      >>> 0 0>
      >>>> 2.
      >>>> proc /proc proc defaults
      >>> 0 0> /dev/sda1 / ext3
      >>> defaults,errors=remount-ro 0 > 1
      >>>> /dev/sda2 none swap sw
      >>> 0 0> /dev/sda1 /media/usb0 auto rw,user,noauto
      >>> 0 0> /dev/sda2 /media/usb1 auto rw,user,noauto
      >>> 0 0>
      >>>> Example 1, Perl-Script, Command-Line:
      >>>>
      >>>> #!/usr/bin/perl
      >>>>
      >>>> # declare array
      >>>> my @data;
      >>>>
      >>>> # start timer
      >>>> $start = time();
      >>>>
      >>>> # perform a math operation 200000 times
      >>>> for ($x=0; $x
      >>>> {
      >>>> $data[$x] = $x/($x+2);
      >>>> }
      >>>>
      >>>> # end timer
      >>>> $end = time();
      >>>>
      >>>> # report
      >>>> print "Time taken was ", ($end - $start), "
      >>> seconds";>
      >>>> Results:
      >>>> SlugOS/BE 2.5 Beta:
      >>>> Time taken was 3 seconds
      >>>>
      >>>> Debian/NSLU2 4.0r1:
      >>>> Time taken was 11 seconds
      >>>>
      >>>>
      >>>> Example 2, PHP-Script (PHP 5.2.5, native
      >>> self-compiled), Web-Server> (lighttpd 1.4.18, native self-compiled):
      >>>> lighttpd/1.4.18 with PHP fast-cgi
      >>>> Script:
      >>>>
      >>>> function microtime_float() {
      >>>> list($usec, $sec) = explode(" ",
      >>> microtime());> return ((float)$usec + (float)$sec);
      >>>> }
      >>>>
      >>>> $time_start = microtime(true);
      >>>> echo "\ntime (start):
      >>> ".$time_start."\n";> for ($i=0; $i
      >>>> $k = sin(abs(100-$i)*(rand(1, 200)));
      >>>> echo $k." ";
      >>>> }
      >>>> $time_end = microtime(true);
      >>>> $time = $time_end-$time_start;
      >>>> echo "\ntime: ".$time."\n";
      >>>> ?>
      >>>>
      >>>> Results:
      >>>>
      >>>> SlugOS/BE 2.5 Beta:
      >>>> 1.838
      >>>>
      >>>> Debian/NSLU2 4.0r1:
      >>>> 6.155
      >>>>
      >>>> I can't find the reason for this. There are no
      >>> "performance-eaten"> background-tasks, CPU-Load on both devices at
      > 3-4%
      >>> before starting the> tests, memory-load at 11-13 MB, also both
      >>> devices.>
      >>>> Could someone explain this behavior?
      >> SlugOS is built with software floating point, meaning that no
      > floating point instructions are issued. Debian relies on a different
      > mechanism, one that attempts to use hardware floating point if it is
      > present. The former is the ideal case for the NSLU2 since it has not
      > floating point hardware, so it's significantly faster. However, the
      > Debian approach is more generic, and since Debian runs on many more
      > ARM devices than just the NSLU2, it makes sense that it should support
      > HW floating point if such exists on the underlying hardware.
      >>
      >>>> Thanks in advance!
      >>>>
      >>> Sorry,
      >>>
      >>> nobody out there with an idea?
      >>> I think this is an important point to discuss before switching from
      >>> SlugOS to the Debian firmware...
      >>>
      >> Mike (mwester)
      >>
      >
      >
      >
      >
      >
      > Yahoo! Groups Links
      >
      >
      >
      >
    Your message has been successfully submitted and would be delivered to recipients shortly.