Hi
I would like to know if there only way to simulate long long int 64 bit .
As long long int 64 bit is not supported in VDSP++.
For example when using Bit wise operation on 32 int data type is there any
way to use 2 , 32 bit int .
is there any alternate and fast way for these sort of Bit operation and use
2 32 bit fixed point to simulate 64 bit fixed point .
Thanking you
With best regards and wishes
Vatsa
64 bit fixed operation SHARC 213xx series using 2 32 bit int data type
Started by ●October 31, 2008
Reply by ●October 31, 20082008-10-31
Sure - use the carry bit!
Look up multi-precision integer methods on the web. Every processor does
the same thing - you use several registers and combine the results by
either multiple operations for a multiply or by adding the carry bit on an
overflow for addition/subtraction. I'll bet ADI has something on their
web site for other processors if not that one.
Patience, persistence, truth,
Dr. mike
On Sat, 1 Nov 2008, wallman16 wrote:
> Hi
> I would like to know if there only way to simulate long long int 64 bit .
> As long long int 64 bit is not supported in VDSP++.
> For example when using Bit wise operation on 32 int data type is there any
> way to use 2 , 32 bit int .
> is there any alternate and fast way for these sort of Bit operation and use
> 2 32 bit fixed point to simulate 64 bit fixed point .
>
> Thanking you
> With best regards and wishes
> Vatsa
>
Look up multi-precision integer methods on the web. Every processor does
the same thing - you use several registers and combine the results by
either multiple operations for a multiply or by adding the carry bit on an
overflow for addition/subtraction. I'll bet ADI has something on their
web site for other processors if not that one.
Patience, persistence, truth,
Dr. mike
On Sat, 1 Nov 2008, wallman16 wrote:
> Hi
> I would like to know if there only way to simulate long long int 64 bit .
> As long long int 64 bit is not supported in VDSP++.
> For example when using Bit wise operation on 32 int data type is there any
> way to use 2 , 32 bit int .
> is there any alternate and fast way for these sort of Bit operation and use
> 2 32 bit fixed point to simulate 64 bit fixed point .
>
> Thanking you
> With best regards and wishes
> Vatsa
>
Reply by ●November 3, 20082008-11-03
Hello Mike ,
Thanks for the information on the Multi Precision integer method .
I searched for this thing over the Internet with regards to SHARC processor
and could not find anything .
Instead i have used 32 bit register and as i am using bit wise shift i have
written a sample code in asm
copy shift copy back is this fine as ADI wont support 64 bit .
Thank you for the information with regards to this issue
With best regards and wishes
Srivatsa M
On Sat, Nov 1, 2008 at 3:25 AM, Mike Rosing wrote:
> Sure - use the carry bit!
>
> Look up multi-precision integer methods on the web. Every processor does
> the same thing - you use several registers and combine the results by
> either multiple operations for a multiply or by adding the carry bit on an
> overflow for addition/subtraction. I'll bet ADI has something on their
> web site for other processors if not that one.
>
> Patience, persistence, truth,
> Dr. mike
> On Sat, 1 Nov 2008, wallman16 wrote:
>
> > Hi
> > I would like to know if there only way to simulate long long int 64 bit .
> > As long long int 64 bit is not supported in VDSP++.
> > For example when using Bit wise operation on 32 int data type is there
> any
> > way to use 2 , 32 bit int .
> > is there any alternate and fast way for these sort of Bit operation and
> use
> > 2 32 bit fixed point to simulate 64 bit fixed point .
> >
> > Thanking you
> > With best regards and wishes
> > Vatsa
>
--
Thanking you
With best regards and wishes
Vatsa
Thanks for the information on the Multi Precision integer method .
I searched for this thing over the Internet with regards to SHARC processor
and could not find anything .
Instead i have used 32 bit register and as i am using bit wise shift i have
written a sample code in asm
copy shift copy back is this fine as ADI wont support 64 bit .
Thank you for the information with regards to this issue
With best regards and wishes
Srivatsa M
On Sat, Nov 1, 2008 at 3:25 AM, Mike Rosing wrote:
> Sure - use the carry bit!
>
> Look up multi-precision integer methods on the web. Every processor does
> the same thing - you use several registers and combine the results by
> either multiple operations for a multiply or by adding the carry bit on an
> overflow for addition/subtraction. I'll bet ADI has something on their
> web site for other processors if not that one.
>
> Patience, persistence, truth,
> Dr. mike
> On Sat, 1 Nov 2008, wallman16 wrote:
>
> > Hi
> > I would like to know if there only way to simulate long long int 64 bit .
> > As long long int 64 bit is not supported in VDSP++.
> > For example when using Bit wise operation on 32 int data type is there
> any
> > way to use 2 , 32 bit int .
> > is there any alternate and fast way for these sort of Bit operation and
> use
> > 2 32 bit fixed point to simulate 64 bit fixed point .
> >
> > Thanking you
> > With best regards and wishes
> > Vatsa
>
--
Thanking you
With best regards and wishes
Vatsa
