hi,
I am try to using ieee floting point operations in the BF533.But i
didn't findthe instruction support for that..If it is possible 1.31
ieee format in the BF533 .Pls help me
friendly
V.M.Kamalesh
Kumar
Can i use ieee floating point operation in BF533
Started by ●October 13, 2005
Reply by ●October 13, 20052005-10-13
On Thu, 13 Oct 2005, V.M.Kamalesh Kumar ^_^ wrote:
> hi,
>
> I am try to using ieee floting point operations in the
>BF533.But i didn't findthe instruction support for that..If it is
>possible 1.31 ieee format in the BF533 .Pls help me
Blackfin is not a float processor. I strongly suggest downloading the
manual from ADI's web site. They explain the frational format in lots of
detail.
Patience, persistence, truth,
Dr. mike
> hi,
>
> I am try to using ieee floting point operations in the
>BF533.But i didn't findthe instruction support for that..If it is
>possible 1.31 ieee format in the BF533 .Pls help me
Blackfin is not a float processor. I strongly suggest downloading the
manual from ADI's web site. They explain the frational format in lots of
detail.
Patience, persistence, truth,
Dr. mike
Reply by ●October 14, 20052005-10-14
On Thursday 13 October 2005 07:46, you wrote:
> hi,
>
> I am try to using ieee floting point operations in the
> BF533.But i didn't findthe instruction support for that..If it is possible
> 1.31 ieee format in the BF533 .Pls help me > friendly > V.M.Kamalesh Kumar
Did you check the C compiler manual which comes with VisualDSP++ ?
You can use 'float' and 'double' types if you work with the C/C++ compiler,
since these ANSI types are fully supported.
Check the manual and search in the index for 'IEEE floating-point support'.
Note, that there are two different floating point libraries, one is IEEE
conformant, the other is not quite, but is rather fast.
However, because Blackfin cannot handle such types on processor level,
implementation of all floating point handling is in software and is very slow!
compared to well-written fractional based assembly code.
You should probably not use floats where time or memory footprint is a matter.
If both is not critical at all, you can even use the compiler library
functions, if you work in assembly language,
The manual explains how this can be done.
Besides: IEEE describes number types in standard 754.
In that list there is a 32bit float and a 64bit float type.
Both are supported by the compiler.
1.31 is not a named float type. However, there's a 1.31 fractional type!
If you mean that, it should be available as a generic type on the Blackfin.
As far as I know, memory is usually 1.15, while arithmetic result registers
can hold 1.31 types.
Again, check the manual.
Bernhard
----
Diese E-Mail enthaelt vertrauliche und/oder rechtlich geschuetzte
Informationen. Wenn Sie nicht der richtige Adressat sind oder diese
E-Mail irrtuemlich erhalten haben, informieren Sie bitte sofort den
Absender und vernichten Sie diese Mail. Das unerlaubte Kopieren sowie
die unbefugte Weitergabe dieser Mail ist nicht gestattet.
Ueber das Internet versandte E-Mails koennen leicht unter fremden Namen
erstellt oder manipuliert werden. Aus diesem Grunde bitten wir um
Verstaendnis, dass wir zu Ihrem und unserem Schutz die rechtliche
Verbindlichkeit der vorstehenden Erklaerungen und Aeusserungen
ausschliessen.
This E-mail may contain confidential and/or privileged information. If
you are not the intended recipient or have received this E-mail in
error, please notify the sender immediately and destroy this E-mail. Any
unauthorized copying, disclosure or distribution of the material in this
E-mail is strictly forbidden.
E-mails via Internet can easily be prepared or manipulated by third
persons. For this reason we trust you will understand that, for your own
and our protection, we rule out the legal validity of the foregoing
statements and comments.
> hi,
>
> I am try to using ieee floting point operations in the
> BF533.But i didn't findthe instruction support for that..If it is possible
> 1.31 ieee format in the BF533 .Pls help me > friendly > V.M.Kamalesh Kumar
Did you check the C compiler manual which comes with VisualDSP++ ?
You can use 'float' and 'double' types if you work with the C/C++ compiler,
since these ANSI types are fully supported.
Check the manual and search in the index for 'IEEE floating-point support'.
Note, that there are two different floating point libraries, one is IEEE
conformant, the other is not quite, but is rather fast.
However, because Blackfin cannot handle such types on processor level,
implementation of all floating point handling is in software and is very slow!
compared to well-written fractional based assembly code.
You should probably not use floats where time or memory footprint is a matter.
If both is not critical at all, you can even use the compiler library
functions, if you work in assembly language,
The manual explains how this can be done.
Besides: IEEE describes number types in standard 754.
In that list there is a 32bit float and a 64bit float type.
Both are supported by the compiler.
1.31 is not a named float type. However, there's a 1.31 fractional type!
If you mean that, it should be available as a generic type on the Blackfin.
As far as I know, memory is usually 1.15, while arithmetic result registers
can hold 1.31 types.
Again, check the manual.
Bernhard
----
Diese E-Mail enthaelt vertrauliche und/oder rechtlich geschuetzte
Informationen. Wenn Sie nicht der richtige Adressat sind oder diese
E-Mail irrtuemlich erhalten haben, informieren Sie bitte sofort den
Absender und vernichten Sie diese Mail. Das unerlaubte Kopieren sowie
die unbefugte Weitergabe dieser Mail ist nicht gestattet.
Ueber das Internet versandte E-Mails koennen leicht unter fremden Namen
erstellt oder manipuliert werden. Aus diesem Grunde bitten wir um
Verstaendnis, dass wir zu Ihrem und unserem Schutz die rechtliche
Verbindlichkeit der vorstehenden Erklaerungen und Aeusserungen
ausschliessen.
This E-mail may contain confidential and/or privileged information. If
you are not the intended recipient or have received this E-mail in
error, please notify the sender immediately and destroy this E-mail. Any
unauthorized copying, disclosure or distribution of the material in this
E-mail is strictly forbidden.
E-mails via Internet can easily be prepared or manipulated by third
persons. For this reason we trust you will understand that, for your own
and our protection, we rule out the legal validity of the foregoing
statements and comments.
