[hpsdr] [Discuss-gnuradio] Intel Atom is NICE.

Robert McGwier rwmcgwier at gmail.com
Mon Dec 29 18:17:51 PST 2008


The new I7 intel processors have a really amazing amount of horsepower
but a tremendous amount of onboard stuff is designed to work
intimately with the peripherals to bring on high speed graphics.  Now
if your argument is that all of this is going to be moved into the
GPP,  I might agree that we will see some of this.  If you are arguing
that the demand for an almost exponential increase in capabilities in
graphics hardware for the purposes of virtual reality and rendering
live and without memorex is not going to be happening, we are in
profound disagreement.

John, can you give me (and the others here) a pointer to the list of
GPIB, etc. devices your fantastic stuff currently supports?  I went
from seeing it in the early days with a handful of things supported to
seeing friends of mine (like W2GPS) running their labs with your code.


Bob


On Mon, Dec 29, 2008 at 8:31 PM, John Miles <jmiles at pop.net> wrote:
> Larrabee, I'm thinking, will be the real SDR platform of choice.  A Larrabee
> box with USB 3.0 is going to put a staggering amount of DSP power into
> peoples' hands.  It won't even make sense to mess with FPGAs at that point,
> I hope.
>
> GPUs are unquestionably an interim hack; I don't think they'll live to see
> the next decade.  They'll go the way of the Weitek.
>
> -- john, KE5FX
>
>> -----Original Message-----
>> From: hpsdr-bounces at lists.hpsdr.org
>> [mailto:hpsdr-bounces at lists.hpsdr.org]On Behalf Of Bob McGwier
>> Sent: Monday, December 29, 2008 4:40 PM
>> To: 'Newman, Timothy'; 'Marcus D. Leech'
>> Cc: hpsdr at lists.hpsdr.org; FlexRadio at flex-radio.biz;
>> qs1r at yahoogroups.com; tim.newman.vt at gmail.com; discuss-gnuradio at gnu.org
>> Subject: Re: [hpsdr] [Discuss-gnuradio] Intel Atom is NICE.
>>
>>
>> ***** High Performance Software Defined Radio Discussion List *****
>>
>> And GPU's are going to become commodity priced quickly and
>> possibly even move into the GPP and replace older ways of doing
>> floating point.  With Nvidia CUDA, you can write code for your
>> GPP, call GPU with intrinsics to get pretty quick payback while a
>> better longer term strategy is worked on.
>>
>> The future of really hard to program heterogeneous/not symmetric
>> multiple core processors,  irrespective of how great the
>> bandwidth is,  I don't think is looking all that rosy.  It simply
>> cannot take months and months to get speed to make the processor
>> pay or the cost per flop, when ALL COSTS are amortized (expensive
>> people, etc.) begins to look bad.
>>
>> Bob
>>
>>
>>
>
>

 1230603471.0


More information about the Hpsdr mailing list