Comments on Montavista's "Open Source Real-Time Linux Project"?

[Marius Groeger]

On Mon, 11 Oct 2004, Paolo Mantegazza wrote:

> "We are reluctant to publish any performance numbers until we have 
> completed the mutex-spinlock partitioning and provisioned support 
> for RW locks. At that point, we expect the worst case preemption 
> latencies to be in the hundreds of microseconds on a typical 
> workstation."
>
> my answer is: I'm not interested in it. In fact they will miss a 
> very few 10s of us anyhow, i.e. what I need.
>
> So, at the moment, I know it is not a substitute of what I'm using 
> for only relatively fast control systems. Then I've got to go on 
> keeping and improving my own tool, leaving any discussioon on the 
> technical content of the proposed solution to those that need real 
> time in the range of unspecified "hundreds of microseconds".

I would like to add that any generic discussion about such numbers is 
largely a matter of PR (regardless of commercial software or OSS). The 
honest answer when asked for worst case numbers is "I will need to 
measure it ony your system including your entire kernel and 
application software environment. I may also have to turn off all 
caches and TLBs, or install cache and TLB eaters in order to catch the 
worst case." Unfortunately, this inevitably leads to very unpopular 
numbers...

Regards,
Marius

-- 
Marius Groeger <mgroeger at sysgo.com>
SYSGO AG                      Embedded and Real-Time Software
Voice: +49 6136 9948 0                  FAX: +49 6136 9948 10
www.sysgo.com | www.elinos.com | www.osek.de | www.imerva.com