Assert at nr-mac-scheduler-ns3

[Daniel Rodrigues de Luna]

Hi all,

I am facing a problem regarding the scheduler in some simulations, and for now I have not understood the problem yet. Besides, it is been very hard to reproduce the error because it takes too long to reach the point of it.

Well, I am running the code (attached) with the default values, but using the RngRun = 227285083604620. After some time I got the error (also attached slurm-1773870.out), reproduced below:

"assert failed. cond="(dataSymPerSlot + m_ulCtrlSymbols) - ulSymAvail == totUlSym", msg="UL symbols available: 14 UL symbols available at end of sched: 0 total of symbols registered in the allocation: 10 slot type UL", +15.376000000s 10 file=../src/nr/model/nr-mac-scheduler-ns3.cc, line=1986
terminate called without an active exception
Command ['/home/drdluna/ns3-dev-nr/build/scratch/NrMmtc-M0_v5', '--RngRun=227285083604620', '--simTime=60', '--window=0.1', '--centralFrequencyBandn28=753000000', '--bandwidthBand=100000000', '--BwpConfiguration=1', '--BwpBand=1250000', '--Shadowing=0', '--UlMode=1', '--TrafficModel=TriggeredReport', '--numerology=0', '--PacketSize=125', '--interval=60', '--Lambda=0.1', '--BwpBand=1250000', '--Lambda=2', '--ueNumScenario=84'] terminated with signal SIGIOT. Run it under a debugger to get more information (./waf --run <program> --gdb")."

I have made some investigations and according to the assert msg, the expression "(dataSymPerSlot + m_ulCtrlSymbols) - ulSymAvail == totUlSym" should be (13 + 1) - 4 = 10. However, it is (13+1) - 0 = 10 and that is why the error is happening. For now, I have not figured out why is that, however I noted that happens when the number of UEs and the amount of traffic increases. That is, when I am using small number of UEs and low load of traffic, the error does not happen.

I would very much appreciate any help on this.

Regards, 

---------------------------------------------
Me. Daniel Rodrigues de Luna
Engenheiro de Telecomunicações
Doutorando em Telecomunicações
Pós-Graduação em Engenharia Elétrica e de Computação (PPGEEC)
Natal, Rio Grande do Norte
Universidade Federal do Rio Grande do Norte (UFRN)
daniel...@gmail.com   dani...@ufrn.edu.br | celular: +55 84 9 9809-4000 |

---------------------------------------------
MSc. Daniel Rodrigues de Luna
Telecommunication Engineering
PhD Student in Telecommunication
Post-Graduate Department of Electrical Engineering and Computation
Natal, Rio Grande do Norte, Brasil
Federal University of Rio Grande do Norte
daniel...@gmail.com   dani...@ufrn.edu.br | celular: +55 84 9 9809-4000 |

[David Segura]

Hi Daniel,

I'm getting a similar error. In my case, this error happens at 655s time simulation, but this is strange because the traffic is always constant, so I don't understand why it fails at 655s and not before. The scenario consists in two gNBs, the first one with 15 UEs attached and the second one with 20 UEs. Traffic model for the UEs is 500 kb/s. The UEs are static and only one is moving.

I attach the output obtained.

Kind regards,

David

[Katerina Koutlia]

Hi,

Could you please re-run your simulations having SRS disabled? You can do that by adding the following code in your example:

   nrHelper->SetSchedulerAttribute("EnableSrsInUlSlots", BooleanValue (false));

  nrHelper->SetSchedulerAttribute("EnableSrsInFSlots", BooleanValue (false));

And please let us know if with this addition you don't hit the assert.

BR,

Kat

--
You received this message because you are subscribed to the Google Groups "5G-LENA-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 5g-lena-user...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/5g-lena-users/59aa5b49-31ca-413d-9ac8-3ba89f3a5d01n%40googlegroups.com.

[David Segura]

Hi,

Do we also have to change RealisticBeamforming to IdealBeamforming if we disable SRS?

BR,

David

[Katerina Koutlia]

Well, if you use realistic BF, then yes, change to ideal and test with SRS disabled. Then if it doesn't crash, enable SRS, and see if it crashes. In all the cases print the logs of nr-mac-scheduler-ns3 and try to compare and locate the problem.

BR,

Kat

To view this discussion on the web, visit https://groups.google.com/d/msgid/5g-lena-users/353c3a92-aff8-43d3-b47c-1d1135ca794cn%40googlegroups.com.

[Biljana B.]

I have opened an issue: https://gitlab.com/cttc-lena/nr/-/issues/88.  

I will take a look as soon as possible.

Thanks for reporting!

[Daniel Rodrigues de Luna]

Hi Kat,

I have implemented your suggestion and, so far, the assert msg has not triggered. Thanks.

Regards,

[Biljana Bojovic]

Hi,

I tried to execute the example from Daniel but I could not because there are some custom EPS bearer types that do not exist in NR/LTE module.

David, if possible, could you please share your example so we can reproduce the error? You can upload it directly here: https://gitlab.com/cttc-lena/nr/-/issues/88. 

Thanks a lot,

Biljana

You received this message because you are subscribed to a topic in the Google Groups "5G-LENA-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/5g-lena-users/lMG1km6E8I8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to 5g-lena-user...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/5g-lena-users/70759e8e-958b-44cd-8021-bb6eee930cf5n%40googlegroups.com.

[David Segura]

Hi Biljana,

I have uploaded my code in gitlab: https://gitlab.com/cttc-lena/nr/-/issues/88.

For simulation, I do not change the parameters on the example. Real beamforming is used with SRS Count by default. With this configuration, at 655s simulation time the assert triggers. I recommend you to execute the code in a server (it takes a few days simulating to reach the error).

I 'm also executing the code without SRS and using Ideal beamforming instead of real to check if the error persists, as Kat suggested.

Kind regards,

David

[Katerina Koutlia]

HI David,

What you can try now to do (if ideal BF without SRS is not hitting the ASSERT) is to enable SRS (with ideal BF) and check the following steps:

  1. If it hits the assert with the SRS and ideal BF, try to reduce the number of symbols reserved for SRS transmissions to 1 and lets see if the assert msg changes

  2. If it doesn't hit the assert go back to real BF and again reduce the SRS symbols to 1 and lets see if the assert msg changes

BR,

Kat

To view this discussion on the web, visit https://groups.google.com/d/msgid/5g-lena-users/5b0866c5-53fe-4d70-b747-5d4ec7335c34n%40googlegroups.com.

[Biljana Bojovic]

Thanks a lot David. We will let you know as soon we have some news.

To view this discussion on the web, visit https://groups.google.com/d/msgid/5g-lena-users/CALMpQ68XHCSUeGbym6sCrnd09wKVv5QSiMJPtaRStnSM5yxgag%40mail.gmail.com.

[David Segura]

Thank you Kat, I will try that.

If I found any problem, I will let you know.

BR,

David

[David Segura]

Hi Biljana and Kat,

I did a simulation with real beamforming and 1 SRS symbol instead of 4 and now the assert does not trigger. It seems to work fine. 

So the problem occurs when SRS is set to 4, which is the default value in NR code.

BR,

David