The KiwiSDR 2 online store is open for orders! Please visit kiwisdr.nz

BBG vs BBAI Water fall rate ~33% slower? [apples/oranges comparison]

I just did comparison of a V1.664 waterfall, and a 1.665 waterfall side by side, both at 'fast' rate setting, both at 2 averages. I have noticed that it takes v1.665 waterfall 43 seconds to fill the display whereas v1.664 take just 31 seconds.

Spectrum display appears to move in time with each other - very hard to see any difference there.

Comments

  • How are you doing the comparison?

    Are you using two Kiwis side-by-side, one running 664 and one running 665? Are you connecting to the one running 665 via a proxy URL?

    Is there something else different about them? Does one have autorun connections enabled?

  • Also, how many users are connected to each ?

    The waterfall will slow down when there are multiple users.

    Regards,

    Martin

  • Side by side comparison using web browser, using two monitors on the same pc, one running 664 and one running 665

    The one running 664 is a KIWI1 (It is a BBAI running debian 9.13)

    The one running 665 is the KIWI2 (BBG running debian 11.9)

    Both have 7 active WSPRDAEMON connections.

    No other differences.

  • Are you JOKING? BBG vs BBAI???

    I'm going to start fining or banning people for posting this kind of stuff..

  • Both Running at 1GHz should there be that much difference?

    If so Ill go ahead and upgrade one of the BBAI to 665 and analyze again.

  • One costs $45, the other $150. Want to guess why? 🙄

  • I was thinking because you can run it at 1.5 GHz.

    However, I have now upgraded a BBAI to 665 and compared to the other BBAI running 664, and they do in fact run at the same rate.

    The difference is the BBG.


    My apologies I would never have though that the waterfall performance was improved that much by going with the superior board.


    You can delete this thread sorry for the noise.

  • jksjks
    edited March 26

    It's not even the clock rate.

    The BBG/BBB uses a TI AM3359 single-core ARM A8.

    The BBAI uses a TI AM5729, dual-core ARM A15.

    Night and day.

    And the BBAI-64 uses a TI TDA4VM dual-core ARM A72.

    Because the waterfall is the lowest priority process in the entire system (due to realtime constraints and the load on the Beagle from the FFT the waterfall requires) it is heavily influenced by the Beagle processing power.

  • edited March 26

    I would never have guessed that it made such a difference.

    When looking at the admin page, monitoring the CPU and eCPU utilization values displayed,

    there isn't a huge change when you open up a browser session with spectrum and waterfall.

    Does it auto scale utilization not to exceed threshold utilization of 'x'?

    those utilization numbers displayed seem to reflect 40% and 70 % headroom respectively.

    As a note, on the admin page, I do now see that the BBAI runs at 23fps vs the BBG at 15fps so that would seem to reflect my observation in the "speed"



    I edited the post title to reflect that this is not a difference in software version. Correctly stated it is a hardware difference,

  • I'm sorry, I have to get back to work..

Sign In or Register to comment.