Re: where is my FIFO for subdevice 1 in PCI-6111card

hallo,David Schleef ,


what I want can be simplified as: how can I change the read sampling speed in
this example "simultaneous start"?(such as 1000000Hz)

I can seemly effectively change the writing speed through changing "freq" ,
but not the reading speed.  By Dr.Rolf Mueller,  "freq"  in the same example
can change the writing and read speed simultaneously with same rate.


I think that it is not comedi's problem, because this example runs good by
Dr.Rolf Mueller's computer under RHlinux. The problem lies in my hardware or
software. But I have changed nearly all things . I concentrate in this
example, because many people know this example and the problem I have here
is same as that in my application( with high frequncy send out a waveform and
begin to listen to the responding  echo for about 0.05second about 50000
datas at the same time or pricisely some 0.0001 seconds later).

Before I have to test this card by Dr.Rolf Mueller's computer, have you some
advices to me?


Wang Maosen

attached:
The "simultaneous start" examples works by me as following(only changed
"freq"):
_____________________________________________________________________
freq=1000:
.....................................write 4096     (I trust this write not,
perhaps only 2048 data have been writen for 2048 AO buffer)
.....................................write 4096
.....................................write 4096
.....................................write 4096
read 4096
read 60
 read 4040
(........repeating )
____________________________________________________________________
freq=100000:
write 4096     (perhaps only 2048 data have been writen for 2048 AO buffer)
write 4096
(more than 500 times write, than come 3 times read)
 read 4096
 read 60
 read 4040
(........repeating bis pipe broken, then only read)
_____________________________________________________________________



David Schleef wrote:

> On Fri, Feb 07, 2003 at 01:15:25AM +0100, Wang Maosen wrote:
> > hallo, all
> >
> >
> > Last week I have compared my comdei_test results with Dr. Rolf Mueller's
> > and found that there are no  applicable FIFO for subdevice 1 in my
> > PCI-6111 card. According to the Manual, there must be 2 FIFOs in
> > PCI-611x card(AI buffer8192 and AO buffer 2048), his 6110 card has FIFO
> > for subdevice 1 in test result, but my 6111's FIFO for AO is "not
> > applicable"(see below 1st attached file). I have at same time also tried
> > the example "simultaneous start" from David Schleef in his computer, it
> > runs right by his computer, but not by my computer, the synchronization
> > from this example lost contoll totally in my computer(see below 2rd
> > attached file).
>
> "Not applicable" means that the test can't be used, not that it
> doesn't work.
>
> > write 4096
> > write 4084
> > ...
> > (more than 500 times write, than come 3 times read)
> > ...
> > write 4096
> > write 4096
> > read 4096
> > read 60
> > read 4040
> > write 4084
> > write 4096
> > ...(repeat bis pipe broken, then only read)
>
> It looks like it's working fine.  How are you expecting it to work?
> (Well, except for the eventual broken pipe.  That's a problem.)
>
> dave...