Databases
StackOverflow2013
Postsdbo
POWould a C6000 DSP be outperformed by a Cortex A9 for FP
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POWould a C6000 DSP be outperformed by a Cortex A9 for FP
    primarykey
    data
    text
    <p>I'm using an OMAP L138 processor at the moment which does not have a hardware FPU. We will be processing spectral data using algorithms that are FP intensive thus the ARM side won't be adequate. I'm not the algorithm person but one is "Dynamic Time Warping" (I don't know what it means, no). The initial performance numbers are:</p> <blockquote> <p>Core i7 Laptop@ 2.9GHz: 1 second <br> Raspberry Pi ARM1176 @ 700MHz: 12 seconds <br> OMAP L138 ARM926 @ 300MHz: 193 seconds</p> </blockquote> <p>Worse, the Pi is about 30% of the price of the board I'm using!</p> <p>I do have a TI C674x which is the other processor in the OMAP L138. The question is would I be best served by spending many weeks trying to:</p> <ul> <li>learn the DSPLINK, interop libraries and toolchain not to mention forking out for the large cost of Code Composer or</li> <li>throwing the L138 out and moving to a Dual Cortex A9 like the Pandaboard, possibly suffering power penalties in the process.</li> </ul> <p>(When I look at FPU performance on the A8, it isn't an improvement over the Rasp Pi but Cortex A9 seems to be).</p> <p>I understand the answer is "it depends". Others here <a href="https://stackoverflow.com/a/15518347">have said</a> that "you unlock an incredible fast DSP that can easily outperform the Cortex-A8 if assigned the right job" but for a defined job set would I be better off skipping to the A9, even if I had to buy an external DSP later? </p>
    singulars
    1. PO
      singulars
      1. PTAnswer
    1. This table or related slice is empty.
    1. PTQuestion
    1. USCommunity
    1. UScarveone
    plurals
    1. PL
      singulars
      1. LTLinked
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PO
      singulars
      1. PTAnswer
    2. PO
      singulars
      1. PTAnswer
    1. VO
      singulars
      1. POWould a C6000 DSP be outperformed by a Cortex A9 for FP
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. POWould a C6000 DSP be outperformed by a Cortex A9 for FP
      1. USsansuiso
      1. VTFavorite
    1. COThis is a hard question to answer without knowing more about the application: Is it commercial? (if so you can't use a Pandaboard for that), what are your IO requirements to feed your algorithms (again, eval boards often don't have fast IO, an so on.
      singulars
      1. POWould a C6000 DSP be outperformed by a Cortex A9 for FP
      1. USmarko
    2. COIt is commercial and IO doesn't appear to be an issue. In essence it comes down to raw FP performance. Is is just as fast to do, say, single or double precision add/subtract on the FP DSP as it would be on an ARM Cortex A9. From the other post the answer appears to be "Yes, if optimised correctly". The problem then becomes: "Is it worth the cost/learning curve/time to use the DSP even if it saves power"? I'm inclined to say "No." (but then I have to change my current development platform.)
      singulars
      1. POWould a C6000 DSP be outperformed by a Cortex A9 for FP
      1. UScarveone
    3. COMy comment about commercial use is because the Pandaboard isn't available in production quantities, nor is robust enough (and I suspect neither is Raspberry Pi). In terms of raw performance, NEON can in theory perform two single precision FP MACs a cycle per core (but there's a huge pipeline which makes branches very inefficient). Hard to believe the DSP is going to do much more than this.
      singulars
      1. POWould a C6000 DSP be outperformed by a Cortex A9 for FP
      1. USmarko
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload