I've been measuring IRQ latencies with RTL2.0 and would like to canvas
your thoughts.
A couple of years ago, I did the same thing with RTL0.6 and got some very
impressive numbers. Here's the setup: A 200MHz Pentium M/B with an
ISA-DIO24 (8255) card. I interrupt the card and the handler pulses an
output pin. I measure (on a scope) the timing between these two pulses.
With RTL0.6, I got about 2 microseconds between the two pulses.
I've done the same with RTL2.0 and get 4 microseconds. I repeated the
measurement with a PCI-DIO24 card and get the same number ~ 4us with a
jitter of about 1us. Has anyone done similar measurements? Are these
numbers reasonable or am I off?
Btw, it's the same motherboard I used with RTL 0.6. Seems to indicate that
RTL 2.0 has slowed down. Also, I understand from Victor that one can mess
with the BIOS config to try and speed things up. I haven't a clue as to
what to mess with. I'd appreciate any pointers you might have.
I haven't tried task (thread) dispatching with RTL2.0, but I did this with
RTL 0.6 and it was about 12 microseconds but with a huge variance, sometimes
up to 30 microseconds. What can I expect with RTL 2.0?
Thanks! Ken
-- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
---
For more information on Real-Time Linux see:
http://www.rtlinux.org/rtlinux/
