Hi Tim and Ray,

I should of searched the 5K side of the KB when looking for processor
recommendations for the 1K (actually I didn't even know the KB existed
till I got my SDR1K.)  Would have saved some grief.  As there is a lively
trade in the 1K with new users all the time likely to make the same
mistake, we ought to link the 1K's to this page 
http://kb.flex-radio.com/article.aspx?id=10410 .  

The on the air results using a Celeron don't do the 1K any credit.  The
best I've gotten is a R4!  In fact, to save giving flexradio a bad rep,
I'm dropping back to the 706 till the new processor arrives in a couple
of days.

Perhaps we ought to issue a big red Celeron no! no! rather than saying a
3Gig processor will work without qualification.  Very deceptive.  We
oughta talk to marketing...

I'm following Ray J's advice.  My mobo documentation says most versions
will support dual core and FSB is fast enough, but I'm out of time and $
for more dinkin' around.  Gotta get on the air.  Will just do a 3.2ghz
Pentium D  for now, and wait till I'm told a quad core is the cat's
pajamas!  

Thanks for the answers, and now on to those other two links you gave me,
Tim.

Thanks and thanks again,

Rob ;o)~
AB7CF




 
On Sun, 23 Sep 2007 19:42:32 -0400 "Tim Ellison" <[EMAIL PROTECTED]>
writes:
> My vote is for the Core 2 Duo.  One with at least 4 MB L2 cache.  At
> some point in the not so distant future, I believe that software 
> changes
> will make better use of dual core processors.  Bob has already made
> tweaks to libfftw3f-3.dll to make better use of dual core machines.
> 
> Look at the Intel E6600 chip.
> 
> 
> -Tim
> 
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED] On Behalf Of Rob Dennison
> Sent: Sunday, September 23, 2007 7:32 PM
> To: [email protected]
> Subject: Re: [Flexradio] Up and Running
> 
> Hi to Tim, Ken and Ed,
> 
> Many thanks to all of you for responding.  Clearly the Celeron is on 
> its
> way to the junk box.  
> 
> Back before I bought a box especially for the SDR1K to be, I cruised 
> the
> Website looking for advice on what processor to buy, the site only 
> said
> greater than 3Ghz.  If the tasks are well suited to dual processing, 
>  a
> 1.6Ghz dual core maybe just makes 3Ghz total.  However, even if 
> PowerSDR
> is locked into one core, the sig processing only gets 1.6Ghz max.
> 
> Reading the three emails I see one neutral on replacement type, one 
> in
> favor of a dual core and one in favor of a Pentium P4.  So, given a
> socket 775, what's "best" Celeron replacement?   I'm thinking best 
> price
> vs. performance?
> 
> Thanks in advance,
> 
> Rob ;o)~
> AB7CF
> 
> Ps. Man I remember when it was great fun creating flight control 
> systems
> for missiles using 10Mhz processors and DARPA spent a bucket 
> creating a
> 25Mhz processor...
> 
> 
> 
> 
> 
> 
> 
> On Sun, 23 Sep 2007 14:02:17 -0500 "Ed Toal" <[EMAIL PROTECTED]> writes:
> > 
> > Hi Rob,
> > 
> > I started using my 2.4ghz Celeron machine to get my SDR1K with an
> > FA-66
> > running and had similar results.  In my case the CPU usage was 
> between
> 
> > 40 and 60%.  When I moved the radio to my station machine, a dual 
> core
> 
> > Athalon at 2.1ghz the CPU usage dropped to 10 to 20% and the audio 
> 
> > results were much better.
> > 
> > So then I dropped a 3.0ghz P4 in place of Celeron and put the 
> radio 
> > back on that machine.  The CPU usage was much better and no audio 
> 
> > problems.
> > As I remember, the CPU usage was in the 30% range.  The CPU was 
> about 
> > $80 at Newegg.com.
> > 
> > That proved to me that it was more than advertised speed, the P4 
> is a 
> > tad quicker but doing things much more efficiently.
> > 
> > No more CELERONS in this household!
> > 
> > The FA-66 here runs at 192khz with buffer size set at minimum and 
> WDM 
> > and ASIO set for smaller.
> > 
> > Yes, you will see a spectrum display bulge below the tuned 
> frequency, 
> > but it won't bother you.  As well, the whole display makes a rise 
> and 
> > fall out from the center.  Again, this won't affect anything.  It 
> 
> > seems like this could be compensated for by software, but I 
> suspect it
> 
> > really isn't worth the time and effort to do so.
> > 
> > Ed N9MW
> > 
> > -----Original Message-----
> > From: [EMAIL PROTECTED] 
> > [mailto:[EMAIL PROTECTED] On Behalf Of Rob 
> Dennison
> > Sent: Sunday, September 23, 2007 11:50 AM
> > To: [email protected]
> > Subject: [Flexradio] Up and Running
> > 
> > Hi All,
> > 
> > Have my SDR1K up and running.   First contact was CE3/VE7SV.  
> Heavy
> > lifting is over.  With a plan jane installation using an FA-66 (no 
> 
> > special RFI or common mode treatments) everything is working very 
> 
> > well.
> > Love the radio...  better than dared I imagine.
> > 
> > So now I'm into fine tuning with a few questions:  
> > 
> > 1.  I'm using an XP Pro box with a 3.46 Ghz Celeron D socket 775 
> and
> > 1
> > Gig of Ram.  Running PowerSDR 1.10.2 and DXLabs only.  
> > 
> > With a 192khz sampling rate I'm experiencing dropouts (even when 
> > running stand alone.)  Running PowerSDR v1.10.2 stand alone and 
> using 
> > a 192khz sampling rate CPU usage averages about 35%. Dropouts 
> occur 
> > often and can be forced by dragging the powerSDR window. When 
> running 
> > with DXLabs they are frequent (enough to interfere with CW 
> reception 
> > at about 22 wpm and to interfere with CWX auto CW transmission in 
> the 
> > same speed
> > range.)
> > CPU
> > utilization still runs about 35% with peaks above 75 %.  Dragging 
> any 
> > window will cause serious drop outs.
> > 
> >  Using 96Khz sampling, CPU utilization drops to ~25% with very few 
> if 
> > any dropouts stand alone.  There are dropouts with DXLabs and 
> > utilization occasionally bouncing up to 31% with occasional 
> excursions
> 
> > into the 70% range.  Dragging a window moves utilization right up 
> to 
> > 100%.
> > 
> > I'm currently using the 96Khz sampling rate.  What to do?  Buy a 
> dual 
> > core?  Pour on the memory?  Both?
> > 
> > 2.  I calibrated everything which was certainly necessary.  
> > 
> > Am a little uncertain about the Level CAL results.  I used a HP 
> 8601A 
> > set to -50dbm.  When I entered -50 in the Level (dbm) control the 
> 
> > results were clearly not right (a noise level at -55dbm.)  When I 
> 
> > entered -100 in
> > the Level (dbm) control the results seem "reasonable?"   With the
> > generator connected the display shows -50 dbm right smack on.   
> With 
> > the
> > antenna connected the noise level is about -110dbm and S5.   (The 
> 
> > noise
> > level is very high in my neighborhood, sigh...)  With a dummy load 
> 
> > connected right to the box the noise level is about -130dbm and 
> S2.
> >  
> > 
> > Is this about right? 
> > 
> > 3.  Mostly out of curiosity, when using the panadapter, there's 
> always
> 
> > big amplitude bulge about 10Khz below the center frequency?
> > 
> > I assume it's some sort of processing artifact?
> > 
> > Other then these relatively small issues everything seems going 
> fine.
> > Congratulations on producing a marvelous radio...
> > 
> > Rob
> > AB7CF
> > 
> >  
> > 
> > _______________________________________________
> > FlexRadio mailing list
> > [email protected]
> > 
> http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz
> > Archive Link: 
> > http://www.mail-archive.com/flexradio%40flex-radio.biz/
> > FlexRadio Knowledge Base: http://kb.flex-radio.com/ FlexRadio 
> > Homepage: http://www.flex-radio.com/
> > 
> > 
> > 
> > 
>  
> 
> _______________________________________________
> FlexRadio mailing list
> [email protected]
> http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz
> Archive Link: 
> http://www.mail-archive.com/flexradio%40flex-radio.biz/
> FlexRadio Knowledge Base: http://kb.flex-radio.com/ FlexRadio 
> Homepage:
> http://www.flex-radio.com/
> 
> 
> 
 

_______________________________________________
FlexRadio mailing list
[email protected]
http://mail.flex-radio.biz/mailman/listinfo/flexradio_flex-radio.biz
Archive Link: http://www.mail-archive.com/flexradio%40flex-radio.biz/
FlexRadio Knowledge Base: http://kb.flex-radio.com/
FlexRadio Homepage: http://www.flex-radio.com/

Reply via email to