Ofdm_benchmark_tx.py on E100 results in sequence error in burst

Hi everyone,

I updated the ofdm_benchmark_tx.py script to use UHD and I’m running it
on
my E100. If I run it at a high enough sample rate, I get a lot of “S”
characters printed to the screen. Some googling shows that this means
there’s packet loss within a burst. What causes this? Is there any way
to
avoid it?

Thanks,
Morgan

On 08/23/2011 05:15 PM, Morgan R. wrote:

Hi everyone,

I updated the ofdm_benchmark_tx.py script to use UHD and I’m running it on
my E100. If I run it at a high enough sample rate, I get a lot of “S”
characters printed to the screen. Some googling shows that this means
there’s packet loss within a burst. What causes this? Is there any way to
avoid it?

What is the exact sample rate? Can you post the updated flowgraph?

Have you tried using the NEON optimized fftw mentioned earlier?

Philip

On 08/23/2011 05:15 PM, Morgan R. wrote:

Hi everyone,

I updated the ofdm_benchmark_tx.py script to use UHD and I’m running it on
my E100. If I run it at a high enough sample rate, I get a lot of “S”
characters printed to the screen. Some googling shows that this means
there’s packet loss within a burst. What causes this? Is there any way to
avoid it?

Also, what kernel version are you running? Have you updated the boot
files and root file?

Philip

On Tue, Aug 23, 2011 at 5:58 PM, Philip B.
[email protected]wrote:

Also, what kernel version are you running? Have you updated the boot files
and root file?

Philip

The updated flowgraph is here:

I’m trying to run it with an fft-length of 1024 and 700 occupied tones.
I’m
using a sample rate of 4Msps.

My E100 has a kernel version of 2.6.38. I’m using fftw optimized for
NEON
from Vesperix, which was posted to this list a month or so ago. I
haven’t
updated any of the boot or root files for a couple of months, I don’t
think
I’m on the most recent version.

The usrp_e module has the following version information:
version: 0.2
srcversion: C95640F14A5715ACE28E715

Thanks,
Morgan