error running osld_rx_file.app

[ra]

Hi all,

I researched this forum but couldn't find an answer to a problem I'm having. When I run 

sudo runcf osld_rx_file.app config

and load, init, run, I get the following error:

>> r

Running...OK!

>> [rtdal]: Caught thread-specific signal

[rtdal]: Got signal SIGSEGV from pipeline thread 0. Current running process idx is 7

The changes I made to the default config file are the following:

cores = "4"

thread_sync_on_finish=true;

and the reason I changed it is because I choose cores="1", I get the following error upon load

>> l

Computing mapping for 66 modules...

Platform model. 1 processsors: C[0]=870, 

Error loading waveform. Not enough resources

[_run_main():122]: loading waveform

Could you please help me run the receiver? I get the same error even when I run osld_rx_usrp.app when using the radio.

Ra.

[Kevin McDermott]

Hello!

I have been working with OSLD and have run into similar issues as well.

As far as the resources issue to run the waveform goes: having "cores = "4" "  only dedicates core #4 (a single core).

To use multiple cores you must have the config file be as follow: "cores = "1,2,3,4" "

If you have a quad-core processor you have 8 logical cores. the Decimal point ( "1.") that follows the number also allocates a core.

so "1." dedicates two logical cores.

Try that out and report back if you have another error.

Hope this helps,
Kev

--
You received this message because you are subscribed to the Google Groups "ALOE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to flexnets+u...@googlegroups.com.
To post to this group, send email to flex...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/flexnets/51025ad5-c447-45d1-9b8f-1f012b23416d%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[ra]

Hi Kevin,

I did try "1,2,3,4" and it exits with an error "Error: Invalid argument" and "1,2" gives me "Not enough resources" error. However "4" works fine :O. But consider the following:

I have a i7-M620 processor which has 2 cores and 4 threads. In the email thread here, Ismail says "4," uses only core #4 but "4" uses 4 cores. But that story is about 6 months old and things might have changed. Also, in the comments section of the config file I see, 

/* Options:

     - single value n uses 0..n cores

     - comma-separated values, e.g. 0,3,5 uses cores 0, 3 and 5.

      - to use a single core, use 1, for instance, to use core 1

*/

So my guess is that I'm using all 4 cores since I see the utilization go up in the system monitor. 

Ra.

[Ismael Gomez]

Hi,

If you have 2 cores 4 threads, you should use 2 cores. Threads share a
single physical processing unit which is very bad for real-time. Then, you
have to set cores=0,2

Try always first running aloe with a long time slot and a single core. Then
try with 2 cores and reduce the timeslot length.

Regarding the "not enough resources" message, remember that the platform
model is defined manually. increasing the capacity should allow you
allocate the waveform.

cheers,
Ismael

To view this discussion on the web visit https://groups.google.com/d/msgid/flexnets/770882ce-96fb-470c-9bb1-1f9b307ca1f4%40googlegroups.com.

[ra]

Hi Ismael,

Thanks for the info. With a longer time slot, I see no errors. However when I run the osld_rx_file.app, I get BLER ~= 0.6. I saw that pss is being detected every 5 time slots in the log file, so at least they are synced. 

Any idea what could be happening?

Ra.

[ra]

I noticed one more thing that doesn't make sense to me yet. I hope you'll be able to shed some light on the internals. 

When I run osld_tx_file.app for sufficiently long time, I get BLER = 0. But when I run osld_tx_file.app for only a little while, I get BLER = 0.6 and increases from there. I am wondering if some internal buffers need valid future data to decode properly. Any ideas?

Ra.

[Vuk Marojevic]

That's because you try to decode data at the receiver that you have not generated by the transmitter. RX  should not run longer than the TX.
Regards,
Vuk

To view this discussion on the web visit https://groups.google.com/d/msgid/flexnets/350fc11c-3b97-4fba-b246-d20427c697e5%40googlegroups.com.

[Ravi Teja C]

But it appears like RX depends on future samples. No? 

--
You received this message because you are subscribed to a topic in the Google Groups "ALOE" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/flexnets/im6dOfdtyGE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to flexnets+u...@googlegroups.com.

To post to this group, send email to flex...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/flexnets/CANjVH%3DgU0DSQtqG1Ycw7V87DSekGAO0n4jep23W0JJhajFu-Fg%40mail.gmail.com.

[Vuk Marojevic]

RX reading from file runs for as long as you let it run. You can modify it to stop when all data has been processed. Would be useful.
We target the real-time over-the-air communication mode.
Regards,
Vuk

To view this discussion on the web visit https://groups.google.com/d/msgid/flexnets/CAAeyTY0WS6H3LneHXjJmHBzZZwF9dRiAPbyhgOCUTZYYpze4sA%40mail.gmail.com.

[ra]

Hi Ismael and Vuk,

Thanks for the help. 

1. Is time slot in your code equivalent to subframe since PSS arrives once every 5 time slots? 

2. If each time slot is 1920 samples at 1.92MHz, by altering the time_slot_ns, are we effectively changing the sampling rate? So, if we are using time_slot_ns = 10,000us, sampling rate is 0.192MHz? 

3. I did go through the slides in documents but I couldn't find answer to the above question. 

Ra.

On Saturday, March 1, 2014 5:59:20 AM UTC-5, Ismael Gomez wrote: