[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Discuss-gnuradio] PLL Problem
From: |
Matt Ettus |
Subject: |
Re: [Discuss-gnuradio] PLL Problem |
Date: |
Tue, 27 Dec 2005 11:20:35 -0800 |
User-agent: |
Mozilla Thunderbird 1.0.7-1.1.fc4 (X11/20050929) |
David Scaperoth wrote:
> I have been testing the capabilities of the pll blocks to hopefully be
> used for a coherent receiver.
>
> I have been using the gr.pll_ref_out(...) to lock a low IF signal.
>
> What I have discovered is that the pll seems to be overdriving the
> processor (2.4GHz P4) when its used at high sampling rates (200kS/sec or
> above). Is this something you have encountered??
It was used without problem for wideband FM reception on a Pentium M
laptop. I don't recall the sample rate, but I think it must have been
640 ks/s.
Are you using a P4 or a P4 celeron? The celeron versions perform very
poorly because of the tiny caches.
> The destination block also seems to affect the processing burden. When
> using the audio card sink as a destination block (24kS/sec) the pll
> seems to work fine for a sinusoidal tone, but when I replace the audio
> card sink with null sink (same sampling rate) or the oscope gui, the
> processor goes crazy. I may be setting this up wrong, but just the
> same I wanted to see if you have encountered this before.
Are the samples coming from a USRP? If the chain has nothing pacing it,
like a USRP or an audio card, then you will use 100% cpu no matter how
simple your processing.
> I need this to work at high sampling rates like 2MS/sec so that I can
> detect larger bandwidth signals, but right now I'm no sure.
No work has been done with optimizing that PLL block, so I'm sure you
could speed it up quite a bit.
Matt