TI had a "c54x dsplib" a few years back. Try to find the
user's
guide on their website. The users guide doc # that I have (from
1998) is SPRA480.
If you are willing to spend some money, we also have some 54x DSP
libraries that have been throughly tested and are reliable. Check
out our website. Most of the functions listed in the 55x library
description from our website are also available in our 54x library.
-Shawn Steenhagen
Software Specialist
Applied Signal Processing
www.appliedsignalprocessing.com
--- In , Felix Chez <felixchez@y...> wrote: > Hi,
>
> Anyone knows where can I get a list of C C5416 Library
> APIs?
>
> Felix C.
>
> __________________________________
>
Reply by Felix Chez●December 20, 20032003-12-20
Hi,
Anyone knows where can I get a list of C C5416 Library
APIs?
Felix C.
__________________________________
Reply by Jeff Brower●December 18, 20032003-12-18
Felix-
> Any idea where can I find the C routine? or do
you
> know who have done any assembly code routine to
> address this issues?
I think all you have to do is write and compile some C code like:
long a, b, c;
a = b*c;
and then look in the .map file to figure out which rts??.lib file was linked.
I'm
going to guess the name of the routine is _lmul or _mul32 or something
similar;
that's what to look for in the .map file. Once you find that, you should
be
able to
locate the source for it (look for the same lib file name but with .c or
*.src
extension); the actual source that does the work should be in asm.
-Jeff
> --- Jeff Brower <> wrote:
> > Felix-
> >
> > > How do I do a 20-bit and 20-Bit MAC, That is,
> > >
> > > MAC *AR3+0%, *AR2+0%, A
> > >
> > > Assuming AR3 is pointing to Coefficient Circular
> > > Buffer and AR2 is pointing to Input Sample
> > Circular
> > > Buffer. Now the question is what happen when the
> > two
> > > is more than 16-bit.
> >
> > C54xx memory is assumed to be 16-bit wide, so the
> > above instruction will read only
> > 16-bit operands and perform a 16x16 multiply. To do
> > long multiplies, up to 32x32,
> > you will need to make -- gasp -- a C function call
> > to one of TI's software libraries,
> > which will take a whole heck of a lot more MIPS than
> > the above instruction.
> >
> > -Jeff
> >
> > _____________________________________
> > Note: If you do a simple "reply" with your email
> > client, only the author of this message will receive
> > your answer. You need to do a "reply all" if you
> > want your answer to be distributed to the entire
> > group.
> >
> > _____________________________________
> > About this discussion group:
> >
> > To Join: Send an email to
> >
> >
> > To Post: Send an email to
> >
> > To Leave: Send an email to
> >
> >
> > Archives: http://www.yahoogroups.com/group/c54x
> >
> > Other Groups: http://www.dsprelated.com
> >
> >
> >
> >
> >
>
> __________________________________
>
> _____________________________________
> Note: If you do a simple "reply" with your email client, only the
author of this message will receive your answer. You need to do a "reply
all" if you want
your answer to be distributed to the entire group. >
> _____________________________________
> About this discussion group:
>
> To Join: Send an email to
>
> To Post: Send an email to
>
> To Leave: Send an email to
>
> Archives: http://www.yahoogroups.com/group/c54x
>
> Other Groups: http://www.dsprelated.com
>
Reply by Felix Chez●December 18, 20032003-12-18
Jeff-
Any idea where can I find the C routine? or do you
know who have done any assembly code routine to
address this issues?
Felix C.
--- Jeff Brower <> wrote: > Felix-
>
> > How do I do a 20-bit and 20-Bit MAC, That is,
> >
> > MAC *AR3+0%, *AR2+0%, A
> >
> > Assuming AR3 is pointing to Coefficient Circular
> > Buffer and AR2 is pointing to Input Sample
> Circular
> > Buffer. Now the question is what happen when the
> two
> > is more than 16-bit.
>
> C54xx memory is assumed to be 16-bit wide, so the
> above instruction will read only
> 16-bit operands and perform a 16x16 multiply. To do
> long multiplies, up to 32x32,
> you will need to make -- gasp -- a C function call
> to one of TI's software libraries,
> which will take a whole heck of a lot more MIPS than
> the above instruction.
>
> -Jeff
>
> _____________________________________
> Note: If you do a simple "reply" with your email
> client, only the author of this message will receive
> your answer. You need to do a "reply all" if you
> want your answer to be distributed to the entire
> group.
>
> _____________________________________
> About this discussion group:
>
> To Join: Send an email to
> To Post: Send an email to
>
> To Leave: Send an email to
> Archives: http://www.yahoogroups.com/group/c54x
>
> Other Groups: http://www.dsprelated.com
__________________________________
Reply by Jeff Brower●December 18, 20032003-12-18
Felix-
> How do I do a 20-bit and 20-Bit MAC, That is,
>
> MAC *AR3+0%, *AR2+0%, A
>
> Assuming AR3 is pointing to Coefficient Circular
> Buffer and AR2 is pointing to Input Sample Circular
> Buffer. Now the question is what happen when the two
> is more than 16-bit.
C54xx memory is assumed to be 16-bit wide, so the above instruction will read
only
16-bit operands and perform a 16x16 multiply. To do long multiplies, up to
32x32,
you will need to make -- gasp -- a C function call to one of TI's
software
libraries,
which will take a whole heck of a lot more MIPS than the above instruction.
-Jeff
Reply by Felix Chez●December 17, 20032003-12-17
How do I do a 20-bit and 20-Bit MAC, That is,
MAC *AR3+0%, *AR2+0%, A
Assuming AR3 is pointing to Coefficient Circular
Buffer and AR2 is pointing to Input Sample Circular
Buffer. Now the question is what happen when the two
is more than 16-bit.