DSPLIB questions

Kevin43482

Other Parts Discussed in Thread: OMAPL138, CCSTUDIO, TMS320F2812

OMAPL138 EVM

I used CCS to develop my algorithm code, which contains double precision floating point operations. I used a few library functions in C67x DSPLIB. Later I wanted to incorporate my algorithm code into a sample dsplink code - message. I tried to build the combined code in Linux, but C67x DSPLib has no linux version. I am quite new to OMAP and linux. Here are my questions:

1. Can I build the DSP code in CCS instead? Can GPP load it and execute it without problem? my guess is yes.

2. I assume you can rebuild C67x DSPLib in linux since the source code is provided. Could you tell me how to build a library?

3 C67x DSPLIB has DP floating point functions, but they disappear from C674x DSPLIB. Is it because C674x makes no improvement on DP floating point operations? Do DP operations in C67x DSPLIB outperform C operators? for example, divdp vs "/".

Thanks

over 15 years ago

0 Mariana over 15 years ago

TI__Mastermind 24340 points

Kevin said:
I used CCS to develop my algorithm code, which contains double precision floating point operations. I used a few library functions in C67x DSPLIB. Later I wanted to incorporate my algorithm code into a sample dsplink code - message. I tried to build the combined code in Linux, but C67x DSPLib has no linux version

Have you seen the page below?

http://processors.wiki.ti.com/index.php/Example_DSPLIB/DSPLink_Application_on_OMAP-L1x

Kevin said:
1. Can I build the DSP code in CCS instead? Can GPP load it and execute it without problem? my guess is yes.

It is possible but not simple. I would suggest looking at the page above before.

For the simple Hello DSPproject, there is a page that explains how to "transform" the Linux makefiles into a CCS project:

http://processors.wiki.ti.com/index.php/CCS_Project_project_for_OMAP-L137_HelloDSP_DSPLINK_example

Kevin said:
2. I assume you can rebuild C67x DSPLib in Linux since the source code is provided. Could you tell me how to build a library?

Again, see page: http://processors.wiki.ti.com/index.php/Example_DSPLIB/DSPLink_Application_on_OMAP-L1x

Kevin said:
3 C67x DSPLIB has DP floating point functions, but they disappear from C674x DSPLIB. Is it because C674x makes no improvement on DP floating point operations?

Please note that the C674x DSPLIB is a floating point library.

Just make sure that you are using the correct mv option:

http://processors.wiki.ti.com/index.php/-mv_option_to_use_with_the_C674x

Kevin said:
Do DP operations in C67x DSPLIB outperform C operators? for example, divdp vs "/".

Usually yes.

0 Ian Guest49803 over 15 years ago in reply to Mariana

Prodigy 180 points

Hi Mariana

I have the same question about double precision support in the C674x DSPLIB. It seems to be missing. When will it appear, or can the C67x DSPLIB modules be used?

Ian

0 Kevin43482 over 15 years ago in reply to Mariana

Prodigy 200 points

Thanks, Mariana. I was not aware of the first link you provided. That's really helpful.

@Ian.. Yes, you can use dsplib67x on C674x.

0 Kevin43482 over 15 years ago in reply to Mariana

Prodigy 200 points

On the mv option, why does the dsplib/dsplink example use -mv6400+? If I want to change it in DSPLink sample codes, where should I look for it in those multi-level make files in linux?

0 Deepali Uppal over 15 years ago in reply to Kevin43482

TI__Expert 4930 points

Kevin,

DSPLink using mv6400+ for floating point devices is a bug.

In order to fix it, you need to update variable STD_CC_FLAGS in file

$DSPLINK/make/DspBios/c64xxp_5.xx_linux.mk if you are building on a Linux host

$DSPLINK/make/DspBios/c64xxp_5.xx_windows.mk if you are building on a Linux host

Deepali

0 Oscar Baumann over 13 years ago in reply to Deepali Uppal

Intellectual 835 points

Hi, I want to know something, I am usien CCStudio v3.1 with tms320f2812 dsp, and i want to know if i can use dsplib, becuase i didn´t find a lot of information. thanks.

Processors

Processors forum

DSPLIB questions