Clarification on nprb

[Guy Gabay]

Hi all,
I’m looking for clarification regarding NrAmc::CalculateTbSize(uint8_t mcs, uint8_t rank, uint32_t nprb) when using the NR TDMA scheduler in ns-3 (5G-LENA).

Conceptually, nprb should represent the number of PRBs allocated to a UE (i.e., dci.m_rbLen).
However, in the TDMA scheduler I’m observing that nprb behaves like a cumulative counter (number of RBs allocated/scanned so far while building the schedule), and this same variable is then passed directly to CalculateTbSize().

This makes the TB size depend on:

• UE ordering in the scheduler loop

• internal traversal state rather than the actual allocation

My understanding is that the correct behavior should be:

uint32_t rbAllocatedToUe = dci.m_rbLen;

uint32_t tbSize = m_amc->CalculateTbSize(mcs, rank, rbAllocatedToUe);

i.e., pass a fixed number of PRBs actually allocated to the UE, and keep any cumulative counters separate for internal bookkeeping only.

This ensures that:

• TB size is stable and independent of UE order

• AMC and DCI allocations are consistent

• BLER and throughput results are realistic

Is the current behavior:

1. Intended by design?

2. A known simplification of the TDMA scheduler?

3. Or an implementation issue where nprb is reused for two different purposes?

Any guidance or clarification would be greatly appreciated. Thanks!

[Tom Henderson]

Hello Guy,

I don't know much about the TDMA scheduler (I work mainly on sidelink code) but I looked at the calls to CalculateTbSize() in the TDMA scheduler code (also in the nr-mac-scheduler-ue-info.cc) and it doesn't seem to me that cumulative counts are passed; each call is using either m_ulRBG or m_dlRBG, which are per-UE variables.  If you still think there is a problem, can you please provide more details?

In working on sidelink code, I recently looked at this method carefully and noticed that it does not calculate the TB size strictly (the name of the method is not fully accurate), and I wrote up some proposed documentation changes in a merge request that is waiting for review/comment.  If you or others want to review my proposal, or have other suggested improvements, I'd appreciate the review.

https://gitlab.com/cttc-lena/nr/-/merge_requests/341

- Tom

--
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, visit https://groups.google.com/d/msgid/5g-lena-users/f10240ab-1aae-4938-9634-d68e4e6fc245n%40googlegroups.com.