The output of pidin fam is very difficult. Numbers in each columns often
contain 9 digits and that seems to screw the output.
- Mario
BTW, why do PIDs in RTP get so big?
One minute after booting i’m seeing PIDs like 500000. I know I haven’t run
half a million processes yet.
–
Bill Caroselli – 1(530) 510-7292
Q-TPS Consulting
QTPS@EarthLink.net
“Mario Charest” <mcharest@clipzinformatic.com> wrote in message
news:9titb4$ah1$1@inn.qnx.com…
The output of pidin fam is very difficult. Numbers in each columns often
contain 9 digits and that seems to screw the output.
- Mario
Bill Caroselli <qtps@earthlink.net> wrote:
BTW, why do PIDs in RTP get so big?
One minute after booting i’m seeing PIDs like 500000. I know I haven’t run
half a million processes yet.
The bottom 12 bits of the PIDs are re-used while the top bits are used
as a count. That is why you see such large values.
chris
–
cdm@qnx.com > “The faster I go, the behinder I get.”
Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
OK. Kind of reminiscent of QNX2 days.
But 12 bits limits us to only 4096 concurrent processes. That wasn’t very
forward thinking, was it?
–
Bill Caroselli – 1(530) 510-7292
Q-TPS Consulting
QTPS@EarthLink.net
“Chris McKillop” <cdm@qnx.com> wrote in message
news:9tnf1b$t1m$3@nntp.qnx.com…
Bill Caroselli <> qtps@earthlink.net> > wrote:
BTW, why do PIDs in RTP get so big?
One minute after booting i’m seeing PIDs like 500000. I know I haven’t
run
half a million processes yet.
The bottom 12 bits of the PIDs are re-used while the top bits are used
as a count. That is why you see such large values.chris
–
cdm@qnx.com > “The faster I go, the behinder I get.”
Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
Bill Caroselli <qtps@earthlink.net> wrote:
OK. Kind of reminiscent of QNX2 days.
But 12 bits limits us to only 4096 concurrent processes. That wasn’t very
forward thinking, was it?
Well, I would hope nothing in “user space” relies on that being the case.
It isn’t a big deal to change that in the future if there is ever a demand
for more processes to be running concurrently (at least that is my
understanding from talking with people). This isn’t windows remember. 
chris
–
cdm@qnx.com > “The faster I go, the behinder I get.”
Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
“Chris McKillop” <cdm@qnx.com> wrote in message
news:9tnf1b$t1m$3@nntp.qnx.com…
Bill Caroselli <> qtps@earthlink.net> > wrote:
BTW, why do PIDs in RTP get so big?
One minute after booting i’m seeing PIDs like 500000. I know I haven’t
run
half a million processes yet.
The bottom 12 bits of the PIDs are re-used while the top bits are used
as a count. That is why you see such large values.
Out of curiously, tops bits are used as a count to count what, how many
times
that 12 bits PID have been used?
chris
–
cdm@qnx.com > “The faster I go, the behinder I get.”
Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
Why not just use all bits as an int and increment for each new process. It
wouldn’t wrap around until 4 billion processes have been loaded.
–
Bill Caroselli – 1(530) 510-7292
Q-TPS Consulting
QTPS@EarthLink.net
“Mario Charest” <mcharest@clipzinformatic.com> wrote in message
news:9tte7b$e7e$1@inn.qnx.com…
“Chris McKillop” <> cdm@qnx.com> > wrote in message
news:9tnf1b$t1m$> 3@nntp.qnx.com> …
Bill Caroselli <> qtps@earthlink.net> > wrote:BTW, why do PIDs in RTP get so big?
One minute after booting i’m seeing PIDs like 500000. I know I
haven’t
run
half a million processes yet.
The bottom 12 bits of the PIDs are re-used while the top bits are used
as a count. That is why you see such large values.Out of curiously, tops bits are used as a count to count what, how many
times
that 12 bits PID have been used?
chris–
cdm@qnx.com > “The faster I go, the behinder I get.”
Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
\
4096 concurrent processes? Sure, back in the old fork()/exec() days…
don’t we write everything threaded now? One or two processes should be
enough.
Kris
Bill Caroselli <qtps@earthlink.net> wrote:
OK. Kind of reminiscent of QNX2 days.
But 12 bits limits us to only 4096 concurrent processes. That wasn’t very
forward thinking, was it?–
Bill Caroselli – 1(530) 510-7292
Q-TPS Consulting
QTPS@EarthLink.net
“Chris McKillop” <> cdm@qnx.com> > wrote in message
news:9tnf1b$t1m$> 3@nntp.qnx.com> …
Bill Caroselli <> qtps@earthlink.net> > wrote:BTW, why do PIDs in RTP get so big?
One minute after booting i’m seeing PIDs like 500000. I know I haven’t
run
half a million processes yet.
The bottom 12 bits of the PIDs are re-used while the top bits are used
as a count. That is why you see such large values.chris
–
cdm@qnx.com > “The faster I go, the behinder I get.”
Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
–
Kris Warkentin
kewarken@qnx.com
(613)591-0836 x9368
“Computer science is no more about computers than astronomy is about telescopes”
–E.W.Dijkstra
Previously, Kris Eric Warkentin wrote in qdn.public.qnxrtp.applications:
4096 concurrent processes? Sure, back in the old fork()/exec() days…
don’t we write everything threaded now? One or two processes should be
enough. >
You MUST be joking! Not everything is advantaged by the use of threads. We don’t all use QNX as an embedded platform. In our case we use it for desktop apps. Isn’t QNX6 supposed to be massively scalable?
Kris
Bill Caroselli <> qtps@earthlink.net> > wrote:
OK. Kind of reminiscent of QNX2 days.But 12 bits limits us to only 4096 concurrent processes. That wasn’t very
forward thinking, was it?–
Bill Caroselli – 1(530) 510-7292
Q-TPS Consulting
QTPS@EarthLink.net
“Chris McKillop” <> cdm@qnx.com> > wrote in message
news:9tnf1b$t1m$> 3@nntp.qnx.com> …
Bill Caroselli <> qtps@earthlink.net> > wrote:BTW, why do PIDs in RTP get so big?
One minute after booting i’m seeing PIDs like 500000. I know I haven’t
run
half a million processes yet.
The bottom 12 bits of the PIDs are re-used while the top bits are used
as a count. That is why you see such large values.chris
–
cdm@qnx.com > “The faster I go, the behinder I get.”
Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
\Kris Warkentin
kewarken@qnx.com
(613)591-0836 x9368
“Computer science is no more about computers than astronomy is about telescopes”
–E.W.Dijkstra
Ian Cannon wrote:
Previously, Kris Eric Warkentin wrote in qdn.public.qnxrtp.applications:
4096 concurrent processes? Sure, back in the old fork()/exec() days…
don’t we write everything threaded now? One or two processes should be
enough. >
You MUST be joking! Not everything is advantaged by the use of threadsWe don’t all use QNX as an embedded platform. In our case we use it
for desktop apps. Isn’t QNX6 supposed to be massively scalable?
Joking isn’t quite the word, I think sarcasm is a better characterization.
btw: there is absolutely no reason why co-operating processes aren’t the
model of choice on embedded systems as well as desktop systems, after
all the vm price is being paid even for the most basic QNX kernel.
Ian Cannon <ianc@tecom.com.au> wrote:
Previously, Kris Eric Warkentin wrote in qdn.public.qnxrtp.applications:
4096 concurrent processes? Sure, back in the old fork()/exec() days…
don’t we write everything threaded now? One or two processes should be
enough. >You MUST be joking! Not everything is advantaged by the use of threads. We don’t
all use QNX as an embedded platform. In our case we use it for desktop apps. Is
n’t QNX6 supposed to be massively scalable?
Yes, but scalablity is more then just the number of processes. Even very
busy websites probably only have processes numbering in the 100’s. My very
busy RTP desktop has a total of 56 processes.
Looking at this from another perspective. Assuming a box with 4G of RAM
then when you hit 4096 processes you are talking about a grand total of
1M of physical memory per process in the system. On a more reasonable
sized machine, say 256M of RAM, it is only 64K per-process.
And as I said before, if 4096 was to become a real issue for our customers
it would simply be increased. But it is a very realistic number to be
using for present day systems. I tried doing a counting “while(1) fork()”
loop on a couple of different machines. I was able to get just under 3K
processes on Linux before I crashed the kernel. At other times it would
start giving me errors around 600-800 processes (and not crash).
chris
–
cdm@qnx.com > “The faster I go, the behinder I get.”
Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<