On 09/06/2019 09:53 PM, Lundberg, Daniel via USRP-users wrote:
Thank you all, I will look into DPDK (didn't know about it until now!)
and investigate what is available with those processors.
Marcus, all we need to do is generate samples from a set of pre-canned
files, record a loopback to file
On 09/06/2019 07:49 PM, Austin Adam wrote:
Hi Marcus,
We have been using the same adapter since even before we sent out the
device, and it always worked without fail. It makes sense what you are
saying, but because we had it working before, I feel like the issue is
deeper than that.
I cannot
Thank you all, I will look into DPDK (didn't know about it until now!) and
investigate what is available with those processors.
Marcus, all we need to do is generate samples from a set of pre-canned files,
record a loopback to file, and also Tx to another system, which sends samples
back that w
Thanks!
I saw a noc_block_rfnoc_adder.xml in gr_ettus which looks as though
it's all set to control such a selectable add-or-subtract block, but I
never found the block.
On Fri, 2019-09-06 at 15:26 -0700, Nick Foster wrote:
> Here's a modified add-only block. You'll have to make a matching .xml
>
Hi Marcus,
We have been using the same adapter since even before we sent out the
device, and it always worked without fail. It makes sense what you are
saying, but because we had it working before, I feel like the issue is
deeper than that.
I cannot ping the device in the current state.
If I don'
On 09/06/2019 07:22 PM, Austin Adam via USRP-users wrote:
Update:
I updated my host computer to UHD version 3.14.1.0 and then
afterwards, wrote that same version to the SD card. When I put the SD
card back in the USRP and ran the 'uhd_find_devices' command, it
actually showed up! Finally! But
Update:
I updated my host computer to UHD version 3.14.1.0 and then afterwards,
wrote that same version to the SD card. When I put the SD card back in the
USRP and ran the 'uhd_find_devices' command, it actually showed up!
Finally! But then, upon restarting the USRP by running "shutdown -h now"
via
On 09/06/2019 05:53 PM, Lundberg, Daniel via USRP-users wrote:
Does anyone have a known good hardware configuration to support full
(or at least close to full) 200 MS/s streaming from the N320?
Preferably with 1 Tx and 2 Rx channels.
As a data point, a recent i3 (4 cores) seems to be chokin
Hi Daniel,
As Marcus mentioned, an i3 is not ideal for streaming at such high rates.
For 200 MS/s of usable bandwidth, you'll need to stream at 250 MS/s per
channel.
A colleague has ran 2x2 @ 250 MS/s using an Intel Xeon E5-1620 v3 @
3.50GHz, and I've ran at those rates with an i9-9900x @ 4.4 G
Here's a modified add-only block. You'll have to make a matching .xml
descriptor and GRC block (if you're using gr-ettus).
Probably it would be a super useful thing to have an add/sub block, instead
of an addsub block. A register-controlled mux to select which operation you
want. I'll think about
Hi Daniel,
i3 doesn't sound like the processor family of choice here; a few more
cores can't hurt. Basically, assume one CPU core per stream for the
wire- to CPU-format conversion, plus a bit of demand for someone
handling OS/network card interrupts.
What're you doing with the samples afterwards?
Nick-
Could you share the tricks to remove one of the output ports? I don't
I'm having similar issues with my modified addsub block and don't have
enough room on the e310 fpga for extra fifos. It's not obvious from the
noc_block_addsub code, the use of splitstream and dummy variables is
very confus
Hi Nate,
I'm using UHD 3.14.0.1. I am not using DPDK.
Regarding the tuning, I think I was not clear in my email. I have no
trouble streaming to RAM disk using the standard Radio->DDC->host graph. I
mentioned that I was running 2x50MS/s, but I can go up to 2x200MS/s with
success. My issue is th
Does anyone have a known good hardware configuration to support full (or at
least close to full) 200 MS/s streaming from the N320? Preferably with 1 Tx
and 2 Rx channels.
As a data point, a recent i3 (4 cores) seems to be choking above 62.5 MS/s.
This is over dual SFP+ ports. At higher sampli
Hello Adnan,
I'm currently not aware of anyone doing that.
However, since one of the typical applications of beefier FPGAs is math
accelerators for linear algebra problems, it's more than likely someone
did in fact implement an SVD before, and you might just need to connect
it to a nocshell to ma
I have a two-channel signal conditioning block based on
noc_block_addsub between the radio and a DDC. The entire flowgraph is
radio(2 channels out)->addsub(2 channels in and out)-> two separate ZMQ
Push Sinks. I'm using the DDC as an integrator/decimator because a pair
of FIRs or moving average blo
Hello,
We are trying to perform singular vector decomposition. The idea is to work on
an RFNoC block that takes in summation of samples from the Radio source and
will perform SVD.
Is anybody working on something similar?
Currently, the RFNoC OFDM synchronizer block has timing constraint issues
I am assuming that when I run 'uhd_find_devices' it shows the current
version in the output, so if that is the case, then my host is on version
*UHD_3.14.0.HEAD-0-g6875d061*, and the USRP is on version
*UHD_3.14.0.0-0-g6875d061.
*Just based on the output here:
*root@ni-n3xx-3177E63:~# uhd_find_de
Here's a way in Python you can pull in SC16 and recover the FC32 from a
file. It'll scale the values from the INT16_MAX back to -1.0 to 1.0 too in
one shot.
import numpy
file="somefile.iq"
dat = numpy.fromfile(file, dtype=numpy.int16)
data = dat[0::2]/32767.0 + 1j*dat[1::2]/32767.0 # ::02 says e
Hi Rob,
What version of UHD are you using?
2x RX 50 MS/s streams should work without much issue with a fast enough
host, especially to a ram disk.
Are you using DPDK? DPDK support for X3xx was recently added to UHD and
will reduce the overhead on the host side, which can help quite a bit. Some
a
Hi David,
The N320 has a discrete component daughterboard and does not support
setting a bandwidth filter.
The RFIC based USRPs such as the B2xx, E31x, E320 (AD9361) and N300/N310
(AD9371) support the set_rx_bandwidth API call.
Regards,
Nate Temple
On Fri, Sep 6, 2019 at 4:55 AM Truan David via
Hi Rajesh - The block "OFDM Sync Short" is part of the GR out-of-tree (OOT)
module "gr-ieee802-11" ... as are many of the other blocks in the image you
provided. If that OOT is not installed already, it shouldn't be difficult
to do so. Hope this is useful! - MLD
On Fri, Sep 6, 2019 at 5:10 AM Dr.
Hi Joel - IIRC UHD takes and provides std::complex values that are
in (-1.0,+1.0), meaning that the minimum (most negative) value is
-1.0+epsilon and the maximum (most positive) value is 1.0-epsilon, where
"epsilon" is the smallest positive 32-bit float value (approximately
1.17549e-38).
If you're
Hi,
As part of an effort to improve capability to store incoming receive chain
samples to files on my SSD without errors ('O' or 'D'), I decided to wire
an X310 noc graph to include the DmaFIFO. My thought was that the DmaFIFO
could better tolerate varying rates of sample consumption at the OS.
Be
Hi Imad,
Thanks very much for your reply. Unfortunately that solution does not work.
Because if it reads 4 bytes (two int16_t in C/C++ parlance; world) which for argument sake holds 15, 16 (which are integers). These
are then cast as floats which makes them 15.0 and 16.0 and then when
viewed
Hi,
We received our N320 and started doing some basic RX tests on it and everything
seems ok. We were able to get an emitted chirp and display it using GQRX.
However, when calling the set_rx_bandwidth() UHD function and then reading the
actual RX bandwidth, it always read back 250MHz. Using rx_
Hi Joeal,
You may try something along the lines of:
import numpy as np
def read_from_file(file_name, count=-1):
dt = np.dtype(' samples = np.fromfile(file_name, dt,
count).astype(np.float32).view(np.complex64)
return samples
So this reads samples as two-byte integers (assuming
Hi Sebastian,
Try also setting the spp via the stream args with stream_args.args["spp"].
You can see an example in rfnoc_rx_to_file.cpp around line 403.
Jonathon
On Fri, Sep 6, 2019 at 3:54 PM Sebastian Bräuer
wrote:
> Hi Jonathon,
>
> thanks for the quick response. It is a UHD application and
28 matches
Mail list logo
