Nice! That is what it should be over a 10 Gbe line.
You can raise it with spp, but it will cause fragmentation in a 10 GBe line.
So fft NOC block is limited because it is general for all applications.
The fosphor I use, uses it's own OpenCL FFT and cl/gl shared graphics.
It loops to create that "live" spectrum fill, but also to refill its FFT.
FFT is still 1024 complex samples, and it gets filled up each time in the
loop!
Any excess is padded with old input and FFTed again:)
But this is a dedicated FFT not a general block...
BR,
Nikos
On Wed, Jun 25, 2025 at 4:47 AM zhiwen_zhou@seu.edu.cn wrote:
Nice! That is what it should be over a 10 Gbe line.
You can raise it with spp, but it will cause fragmentation in a 10 GBe line.
So fft NOC block is limited because it is general for all applications.
The fosphor I use, uses it's own OpenCL FFT and cl/gl shared graphics.
It loops to create that "live" spectrum fill, but also to refill its FFT.
FFT is still 1024 complex samples, and it gets filled up each time in the
loop!
Any excess is padded with old input and FFTed again:)
But this is a dedicated FFT not a general block...
BR,
Nikos
On Wed, Jun 25, 2025 at 4:47 AM <zhiwen_zhou@seu.edu.cn> wrote:
> My MTU is 9000 and uhd_rx_streamer_max_num_samps returns 1996.
> _______________________________________________
> USRP-users mailing list -- usrp-users@lists.ettus.com
> To unsubscribe send an email to usrp-users-leave@lists.ettus.com
>
