Subject: Re: [linux-audio-dev] Re: new preemptive kernel-patch from Montavista available
From: Nigel Gamble (nigel_AT_nrg.org)
Date: Thu Nov 23 2000 - 23:02:01 EET
On Thu, 23 Nov 2000 yodaiken_AT_fsmlabs.com wrote:
> It's amazing how two people can look at the same thing and see something completely
> different. I look at REAL/IX and IRIX and see them as definitive proof that this
> technique is a total disaster. Both operating systems bloated into impossible to
> maintain, buggy, slow, disasters that needed specialized and very expensive hardware
> to provide unimpressive RT.
I don't know where your evidence is for these claims? Have you ever had
to maintain either REAL/IX or IRIX yourself? I have, and you are
incorrect.
> MODCOMP's marketing claims 100microseconds worst case from interrupt to driver activation (not process activation) and "typical" "in the 50 microsecond range on a 133 MHz Pentium based system"
> (not clear whether this is a off-the-shelf or MODCOMP hardware). Ordinary Linux
> give "typical" under 10 microseconds. Why is Linux performance so much better than the
> _marketing numbers_ for REAL/IX ?
When I worked on REAL/IX, the numbers were 100us max interrupt
latency (which is the number you quote), 100 us max in interrupt
handler, and 100us max process preemption latency, giving a total
maximum latency from external interrupt to real-time process of
300us. This was on a standard Motorola VME-based 25MHz 68030
processor - no special hardware. You are only quoting times
to interrupt handler. How does this help me to run XMMS, or any
other standard Linux application that tries to make use of Posix
real-time scheduler system calls, without glitches?
You are comparing apples to oranges here.
Anyway, I refuse to get into a marketing discussion, I'll just keep
putting out patches that keep improving interactivity for real
Linux applications.
Nigel Gamble nigel_AT_nrg.org
Mountain View, CA, USA. http://www.nrg.org/
This archive was generated by hypermail 2b28 : Fri Nov 24 2000 - 00:01:41 EET