Re: Working ... but not quite

["William Scalione" <wscalione-at-nc.rr.com>]

Also for SHMEM KEY here is the answer I got from paul the other day
By the way, thanks Paul

Hi Bill

SHMEM_KEY is nothing more than an identifier for a block of memory - If
several different programs need to read or write common data, some way of
accessing a shared block is often easiest. Because this shared memory is
allocated dynamically, it is impossible to pass the base address around, so
a
kernel function is called with SHMEM_KEY and a real address is returned. A
bit like a mobile phone number - All you need is the number, you don't care
where the phone is located (and how many are all zeros ?).

Shared memory blocks can be allocated within a loop incrementing SHMEM_KEY
each time for every block. When the memory is released, the key can be
decremented - When it reaches zero, all shared memory has been freed up.


Regards, Paul.



On Tuesday 10 Sep 2002 1:55 am, William Scalione wrote:
> So what exactly is SHMEM_KEY, and if running a TNG version, are you saying
> that it can be set to any integer except zero?




>
>
> List
>
> Let's get everything we know about the ini file to Robin.  Go to his site
> and see what needs to be filled in or expanded.
>
>
> Robin
>
> Thanks for the kind offer to help with documentation.  Yes I think that
> it will be useful.  I'm sure that the faster the pc the faster it can
> spit out pulses but I don't think that there is a 1 to 1 correlation.
> I've been content with the thought that the relationship was ordinal.
> Perhaps we should do some testing.
>
> Ray
>
>
> On Monday 16 September 2002 06:28 am, you wrote:
> > On Monday 16 September 2002 01:40, Ray Henry wrote:
> > > Robin
> > >
> > > You will not get the same pulse rate from Linux and from Dos using
> > > the same box.  Four KHz sounds good for a 150 machine.  Some report
> > > that they can get about 15 KHz from 450+ machines with steppermod.
> > > You will get somewhat less with freqmod.
> >
> > oh .. OK  .. better upgrade the box then .. base machines here are
> > 1.7ghz anyway now, so assuming it scales linearly then it should be OK.
> >
> > > It's on the "to-do" list.
> >
> > OK .. well .. I'll make a start on it for my own benefit, from what I
> > can work out. I've parsed the ini file and will volunteer to collate
> > any information and format it into html and ps/pdf .. I'll work out
> > what I can from the source .. if a couple of people who *do* know whats
> > what could have a quick look at http://www.quacky.co.uk/~robin/ini.txt
> > and fill in a few gaps (I'll merge any info thats submitted and keep it
> > updated) it should be possible to document most of it fairly quickly.
> >
> > I'd like to document a) 'standard' value from emc.ini, b) the range of
> > acceptable values and c) some notes on the effects and constraints etc.
> > (ie what it does)
> >
> > Should it ever become nearly complete it can be merged back into CVS if
> > you think it useful.
> >
> > > Ray
> > >
> > > On Sunday 15 September 2002 06:56 pm, Robin wrote:
> > > > Right ... more questions ..
> > > >
> > > > I'm getting 'axis follow' errors when I make rapid traverses (well
> > > > .. slow traverses .. it quits a long time before it gets anywhere
> > > > near 'rapid') .. what sort of maximum pulse rate should I be able
> > > > to reliably get out of steppermod.o and a Pentium 150 ... I need
> > > > around 20 to 30khz to get best performance out of the motors ...
> > > > I'm not getting much more than 4khz at the moment ... any
> > > > suggestions?
> > > >
> > > >
> > > > and ... is there a clear and concise explanation of the various
> > > > terms in the ini file ... ?? I'm struggling to identify all the
> > > > lines .. the handbook identifies some of the lines but some are
> > > > undocumented .. am I missing something and theres a big lump of
> > > > documentation apart from the emc handbook somewhere ... ?? or is
> > > > this still on the to-do list .. ??
> > > >
> > > > many thanks.
>