S
seckinoncu8070@gmail.com
Tue, Feb 1, 2022 8:33 AM
Unfortunately, that doesn’t work neither. Same error is occured again. I can understand why it dropped packets (cecause of PC or NIC) but I have no clue about why it stops transmitting.
Unfortunately, that doesn’t work neither. Same error is occured again. I can understand why it dropped packets (cecause of PC or NIC) but I have no clue about why it stops transmitting.
MD
Marcus D. Leech
Tue, Feb 1, 2022 3:59 PM
Unfortunately, that doesn’t work neither. Same error is occured again.
I can understand why it dropped packets (cecause of PC or NIC) but I
have no clue about why it stops transmitting.
Could you check the error counters on your network card before and after
you start a "run" where this happens?
How sensitive is it to sample rate for two channels? That is, if you
have two channels at 5Msps does this happen at all?
Are you running inside a VM, or on native hardware? What OS?
On 2022-02-01 03:33, seckinoncu8070@gmail.com wrote:
>
> Unfortunately, that doesn’t work neither. Same error is occured again.
> I can understand why it dropped packets (cecause of PC or NIC) but I
> have no clue about why it stops transmitting.
>
>
Could you check the error counters on your network card before and after
you start a "run" where this happens?
How sensitive is it to sample rate for two channels? That is, if you
have two channels at 5Msps does this happen at all?
Are you running inside a VM, or on native hardware? What OS?
RK
Rob Kossler
Tue, Feb 1, 2022 4:37 PM
Unfortunately, that doesn’t work neither. Same error is occured again.
I can understand why it dropped packets (cecause of PC or NIC) but I
have no clue about why it stops transmitting.
Could you check the error counters on your network card before and after
you start a "run" where this happens?
How sensitive is it to sample rate for two channels? That is, if you
have two channels at 5Msps does this happen at all?
Are you running inside a VM, or on native hardware? What OS?
While debugging this issue, my suggestion is to focus exclusively on the
UHD example "benchmark_rate" rather than operating from gnuradio. Once
things are working correctly in "benchmark_rate", then move to a different
UHD example "tx_samples_from_file" (perhaps initially with a null source
and then moving to a file source). Once that is working correctly, move
back into gnuradio world. This approach just makes debugging a bit clearer.
Rob
On Tue, Feb 1, 2022 at 10:59 AM Marcus D. Leech <patchvonbraun@gmail.com>
wrote:
> On 2022-02-01 03:33, seckinoncu8070@gmail.com wrote:
> >
> > Unfortunately, that doesn’t work neither. Same error is occured again.
> > I can understand why it dropped packets (cecause of PC or NIC) but I
> > have no clue about why it stops transmitting.
> >
> >
> Could you check the error counters on your network card before and after
> you start a "run" where this happens?
>
> How sensitive is it to sample rate for two channels? That is, if you
> have two channels at 5Msps does this happen at all?
>
> Are you running inside a VM, or on native hardware? What OS?
>
While debugging this issue, my suggestion is to focus exclusively on the
UHD example "benchmark_rate" rather than operating from gnuradio. Once
things are working correctly in "benchmark_rate", then move to a different
UHD example "tx_samples_from_file" (perhaps initially with a null source
and then moving to a file source). Once that is working correctly, move
back into gnuradio world. This approach just makes debugging a bit clearer.
Rob
>
> _______________________________________________
> USRP-users mailing list -- usrp-users@lists.ettus.com
> To unsubscribe send an email to usrp-users-leave@lists.ettus.com
>
S
seckinoncu8070@gmail.com
Wed, Feb 2, 2022 1:11 PM
Hello Marcus,
-
I attached the related pictures of error counters. There is no problem caused by NIC.
-
There is also problem with 5 Msps sample rate but it took longer time (almost 4 minutes) to give error. When I increase the sample rate it took shorter time to give error.
-
TX at 5 Msps gives error after 4 minutes
-
TX at 10 Msps gives error after 2.5 minutes
-
TX at 20 Msps gives error after 1.5 minutes
-
TX at 50 Msps gives error after 50 seconds
-
I am running on native hardware. Ubuntu 20.04 and UHD 4.1.
Hello Marcus,
* I attached the related pictures of error counters. There is no problem caused by NIC.
* There is also problem with 5 Msps sample rate but it took longer time (almost 4 minutes) to give error. When I increase the sample rate it took shorter time to give error.
* TX at 5 Msps gives error after 4 minutes
* TX at 10 Msps gives error after 2.5 minutes
* TX at 20 Msps gives error after 1.5 minutes
* TX at 50 Msps gives error after 50 seconds
* I am running on native hardware. Ubuntu 20.04 and UHD 4.1.
S
seckinoncu8070@gmail.com
Wed, Feb 2, 2022 1:24 PM
Hello Rob,
I got same error when operating on “benchmark_rate”. I attached the related pictures.
-
At 5 Msps, duration = 10 minutes —> No error
-
At 20 Msps, duration = 2 minutes —> No error
-
At 25 Msps, duration = 2 minutes —> Error detected
-
At 50Msps, duration = 2 minutes —> Error detected
-
At 100 Msps, duration = 2 minutes —> Error detected
In addition, when I got “U” errors while running “benchmark_rate”, it continues to transmit.
Hello Rob,
I got same error when operating on “benchmark_rate”. I attached the related pictures.
* At 5 Msps, duration = 10 minutes —> No error
* At 20 Msps, duration = 2 minutes —> No error
* At 25 Msps, duration = 2 minutes —> Error detected
* At 50Msps, duration = 2 minutes —> Error detected
* At 100 Msps, duration = 2 minutes —> Error detected
In addition, when I got “U” errors while running “benchmark_rate”, it continues to transmit.
RK
Rob Kossler
Wed, Feb 2, 2022 2:39 PM
Hello Rob,
I got same error when operating on “benchmark_rate”. I attached the
related pictures.
-
At 5 Msps, duration = 10 minutes —> No error
-
At 20 Msps, duration = 2 minutes —> No error
-
At 25 Msps, duration = 2 minutes —> Error detected
-
At 50Msps, duration = 2 minutes —> Error detected
-
At 100 Msps, duration = 2 minutes —> Error detected
In addition, when I got “U” errors while running “benchmark_rate”, it
continues to transmit.
USRP-users mailing list -- usrp-users@lists.ettus.com
To unsubscribe send an email to usrp-users-leave@lists.ettus.com
Ok. If benchmark_rate fails, this means that either your host is not
optimized or else it is underpowered, I think. Did you try the items at this
<https://kb.ettus.com/USRP_Host_Performance_Tuning_Tips_and_Tricks#Adjust_Network_Buffers>link?
Are you getting any error messages or warnings (other than those shown)
when you run benchmark_rate? How about rx_streaming - are you able to
consume Rx streaming at high rates without errors?
On Wed, Feb 2, 2022 at 8:25 AM <seckinoncu8070@gmail.com> wrote:
> Hello Rob,
>
> I got same error when operating on “benchmark_rate”. I attached the
> related pictures.
>
> -
>
> At 5 Msps, duration = 10 minutes —> No error
> -
>
> At 20 Msps, duration = 2 minutes —> No error
> -
>
> At 25 Msps, duration = 2 minutes —> Error detected
> -
>
> At 50Msps, duration = 2 minutes —> Error detected
> -
>
> At 100 Msps, duration = 2 minutes —> Error detected
>
> In addition, when I got “U” errors while running “benchmark_rate”, it
> continues to transmit.
> _______________________________________________
> USRP-users mailing list -- usrp-users@lists.ettus.com
> To unsubscribe send an email to usrp-users-leave@lists.ettus.com
>
MD
Marcus D. Leech
Wed, Feb 2, 2022 2:43 PM
On 2022-02-02 09:39, Rob Kossler wrote:
Or there's something wrong with the ethernet interface hardware, or the
drivers for it.
Other folks have (quite recently) done 2-channel transmit experiments
with N320/N321 without any issue. I don't have one of these myself, so
I cannot do
experiments.
But if "benchmark_rate" fails at low sample rates (and 5Msps X 2 is
pretty modest), then something is woefully wrong with your system.
On 2022-02-02 09:39, Rob Kossler wrote:
> Ok. If benchmark_rate fails, this means that either your host is not
> optimized or else it is underpowered, I think. Did you try the items
> at this
> <https://kb.ettus.com/USRP_Host_Performance_Tuning_Tips_and_Tricks#Adjust_Network_Buffers>link?
> Are you getting any error messages or warnings (other than those
> shown) when you run benchmark_rate? How about rx_streaming - are you
> able to consume Rx streaming at high rates without errors?
>
Or there's something wrong with the ethernet interface hardware, or the
drivers for it.
Other folks have (quite recently) done 2-channel transmit experiments
with N320/N321 without any issue. I don't have one of these myself, so
I cannot do
experiments.
But if "benchmark_rate" fails at low sample rates (and 5Msps X 2 is
pretty modest), then something is woefully wrong with your system.
MD
Marcus D. Leech
Wed, Feb 2, 2022 3:14 PM
Hello Marcus,
I attached the related pictures of error counters. There is no
problem caused by NIC.
There is also problem with 5 Msps sample rate but it took longer
time (almost 4 minutes) to give error. When I increase the sample
rate it took shorter time to give error.
o
TX at 5 Msps gives error after 4 minutes
o
TX at 10 Msps gives error after 2.5 minutes
o
TX at 20 Msps gives error after 1.5 minutes
o
TX at 50 Msps gives error after 50 seconds
I am running on native hardware. Ubuntu 20.04 and UHD 4.1.
USRP-users mailing list --usrp-users@lists.ettus.com
To unsubscribe send an email tousrp-users-leave@lists.ettus.com
Can you use "htop" to look at the CPU consumption when you're using
benchmark_rate at 5Msps? Just working on a hunch.
On 2022-02-02 08:11, seckinoncu8070@gmail.com wrote:
>
> Hello Marcus,
>
> *
>
> I attached the related pictures of error counters. There is no
> problem caused by NIC.
>
> *
>
> There is also problem with 5 Msps sample rate but it took longer
> time (almost 4 minutes) to give error. When I increase the sample
> rate it took shorter time to give error.
>
> o
>
> TX at 5 Msps gives error after 4 minutes
>
> o
>
> TX at 10 Msps gives error after 2.5 minutes
>
> o
>
> TX at 20 Msps gives error after 1.5 minutes
>
> o
>
> TX at 50 Msps gives error after 50 seconds
>
> *
>
> I am running on native hardware. Ubuntu 20.04 and UHD 4.1.
>
>
> _______________________________________________
> USRP-users mailing list --usrp-users@lists.ettus.com
> To unsubscribe send an email tousrp-users-leave@lists.ettus.com
Can you use "htop" to look at the CPU consumption when you're using
benchmark_rate at 5Msps? Just working on a hunch.
