• Locked Locked
  • Replies 3 replies
  • Subscribers 97 subscribers
  • Views 1339 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Regarding TI : Is OMAP3530 use Fixed point or Floating Point

UWM_res
Intellectual 350 points
Other Parts Discussed in Thread: OMAP3530

As i searched the net, i came across that OMAP3530 is used floating point, but at some places i did read that it is Fixed point. No where in the manuals or documenst, does OMAP3530 mentions it's uses Fixed point or floating point

 

Could some one let me know, which method does OMAP3530 use and link to the documents.

 

OMAP3 and OMAP4 chips include an ARM Cortex-A8 or A9 (ARMv7) and frequently a fixed point C64x+ DSP  http://en.wikipedia.org/wiki/Texas_Instruments_OMAP

Supports 2D graphics acceleration using PowerVR SGX OpenGL ES 2.0 compliant graphics accelerator. NEON SIMD Engine with fixed and floating point operations  http://e-consystems.com/blog/omap-35x-camera/


  • TI__Guru 64600 points

    The DSP (Digital Signal Processor) inside of the OMAP3530 is based on the C64x DSP architecture which is a fixed point processor.

    The Cortex-A8 CPU inside the OMAP3530 incorporates the ARM Neon SIMD vector floating point (VFP) engine.

     

  • in reply to BrandonAzbell
    Intellectual 350 points

    Brandon:

    Do you know, if DSP applications use the fixed or floating Point processor, if Fixed point , Is it a 16 bit or 32 bits?

     I assume ARM Neon SIMD vector floating point (VFP) engine is used for image and videos, correct me if i am wrong...

    where are these information listed, in TI's website or Wiki..?

  • in reply to UWM_res
    TI__Guru 64600 points

    These are very open ended questions that have multiple answers.  In short, there are a number of DSP applications that are more easily implemented with a native floating point processor.  There are applications that fixed point is appropriate.  It truly depends.

    There is numerous pieces of information on TI's website regarding the C64x architecture and I would also suggest exploring our Wiki site for processors.

    For video and imaging, again, you can do these on the ARM.  There is a tradeoff in terms of performance versus everything else you want the ARM to do which typically includes a high level operating system like Linux or WinCE.  In addition, the ARM generally is the one handling I/O, etc.  Therefore, TI has a software framework called CodecEngine which enables imaging and video processing on the DSP inside the OMAP3530, freeing up bandwidth on the ARM for managing data into and out of the DSP, as well as provide high performance on other processing, I/O, application level processing, etc.