Add physim tests into ctest framework
Convert physim testcases listed in
cmake_targets/autotests/test_case_list.xml into ctest. Note: Tests for
dlsim_tm4 are excluded, due to compilation issues.
CTest evaluates success based on return code: '0' = pass, non-zero =
fail. For some simulators, the return code is non-zero even when the
test is successful (as indicated by success messages in the test logs).
For some simulators, the return code is always '0'. This inconsistency
has been address in the first two commits of this MR.
To build and execute these tests:
- using cmake
cd openairinterface5g
mkdir build && cd build # you can also do that in cmake_targets/ran_build/build
cmake .. -GNinja -DENABLE_PHYSIM_TESTS=ON && ninja tests
ctest
- using build_oai: physim-related ctests are activated when PHY
simulators are built (use build_oai with '-P' or '--phy_simulators'
option)
cd openairinterface5g/cmake_targets
./build_oai --ninja --phy_simulators
cd ran_build/build
ctest
Link to newly created 4g and 5g RAN-Physim-Cluster pipelines:
- RAN-Physim-Cluster-4G: https://jenkins-oai.eurecom.fr/job/RAN-Physim-Cluster-4G/
- RAN-Physim-Cluster-5G: https://jenkins-oai.eurecom.fr/job/RAN-Physim-Cluster-5G/
Setting the DEBUG_UE_TIMING flag caused a flood of UE timing logs, resulting
in dlsim output logs as large as 1.5MB per test in CI. This commit comments
out the line where we set the flag to reduce log verbosity.
- Change return value of nr_dlsim, nr_pbchsim and nr_pucchsim to:
- Return '0' if test succeeds (indicated also by "PDSCH/PBCH/PUCCH test
OK" in the test log).
- Return '1' if test fails.
Improvements for LDPC encoding
AVX512 modifications for LDPC encoding: interleaving, ZC384 BG1, output
formatting for encoder, segmentation (memcpy instead of loop). some
improvement in TX for aarch64 in same places where AVX512 support was
added. Also, rate matching and interleaving are done on bytes containing
8 segments and reformatting of the output is done at the end of segment
processing instead of after ldpc encoding.
This improves the overall performance of the NR DL transmitter in gNB
See the merge request description for numbers.
1. nrLDPC_coding_segment_encoder: pointer not aligned on 64 bits
An undefined behavior was happening because of a shift on a __m64 which
was actually a uint32_t which was not aligned on 64 bits.
This commit suggests a fix by properly loading into a __m64 with _mm_set_pi32
before performing the shift.
2. nr_rate_matching: Filler bits detection
Previously, if the bit selection was reaching Ncb and restarting selection from the beginning of encoded data, the filler bits were detected by testing the bytes again NR_NULL which was equal to 2.
But on this branch the bits of up to 8 segments are stored in one byte while doing rate matching so that one byte could be tested to be a filler while it is not.
So this commit changes the way to test filler bits.
Instead of relying on NR_NULL, bit selection now relies on the offset and size of filler bits.
3. nr_ulsch_coding: missing filling f with zeros
The array `f` of the UE UL HARQ process was not filled with zeros between transmissions.
This is a problem if the encoder does not properly overwrite its output.
In order to avoid such situation, this commit adds a memset to fill `f` with zeros before passing it to the encoder.
4. nrLDPC_coding_segment_encoder: missing init of f to zeros
In the encoder, `f` is not filled with zeros before being used.
`f` is instead somehow filled with arbitrary values.
When writing to `output` by 64 bit vectors, some arbitrary bits may be added after the segment.
Then the arbitrary added ones are not overwritten when writing the next segment because of writing with a logical or.
This commit properly fills `f` and `f2` with zeros.
5. nr_ulschsim & nr_dlschsim: Adapt modulation to new encoder output
The QPSK symbols that are feed to the channel model of nr_dlschsim were determined based on the bytes of the encoder output.
With the new output of the encoder, it had to be reworked to determine the QPSK symbols based on the bits of the encoder output.
1. 3GPP Rate-matching/interleaving on 8-segment blocks, deinterleaving at output of multi-segment processing.
2. Bit-packing done in nrLDPC_coding_segment_encoder.c now instead of nr_scrambling.
Interestingly, it improves performance significantly too.
Tested for AVX512 only for now.
- rewrite add_UE_to_list() and rewrite remove_UE_from_list() to simplify
list handling
- return pointer from get_new_nr_ue_inst(); it's completely freed by
delete_nr_ue_data()
- have add_new_UE_RA() add UE context to access list; remove with
nr_release_ra_UE()
- have add_connected_nr_ue() only add UE to connected list; to move from
access to connected use transition_ra_connected_nr_ue(); remove with
mac_remove_nr_ue() as previously
- free memory on problems during RA
- add some documentation
- in NSA/HO: check for pre-configured RA process
- there are asserts in all functions to clarify when a UE has RA
NR UE: add support for continuous frequency offset compensation (in particular NTN LEO Doppler)
- move function add_noise() to file
openair1/SIMULATION/TOOLS/phase_noise.c
- refactor (co)sinus calculation out of phase_noise() into get_sin_cos()
in nr_common.c
- NR UE: continuously estimate the frequency offset using PBCH DMRS
- NR UE: add command line parameter --initial-fo to specify known
initial frequency offset
- NR UE: add support for continuous DL frequency offset compensation
--cont-fo-comp
- NR UE: add support for continuous UL frequency offset
pre-compensation. Depending on the main source of frequency offset,
pre-compensate the UL frequency offset (FO) accordingly. Therefore
modify the flag --cont-fo-comp to accept an integer value, where
* 1 specifies, that the main FO contribution comes from the accuracy of the local oscillator (LO)
* 2 specifies, that the main FO contribution comes from Doppler shift
- NR UE: add PI controller to continuous frequency offset compensation
- rfsimulator: enable and correct NTN LEO doppler simulation
- use SIMD to speed up nr_fo_compensation()
- Extend NTN LEO example in RUNMODEM.md to include continuous Doppler
compensation
- NR UE: in nr_adjust_synch_ue() extend the peak search to +/- cyclic
prefix size, esp. for LEO NTN scenarios the timing drifts during
initial sync, which is mostly compensated with
--ntn-initial-time-drift. But there might be a residual time offset,
and by extending the peak search window we make this more tolerant.
- NR UE: remove unused nr_slot_tx_offset, it is unused since commit
f3e60a14
- NR UE: remove now unnecessary local variable
update_ntn_system_information, directly use
UE->ntn_config_message->update instead
nr_dlsim allocates memory on every SNR iteration without freeing it.
This can lead to memory exhaustion. Fix this by allocating only once
and reuse memory. Also, free it before exit.
Miscelaneous improvements in PHY simulators
1: Add a SIGINT handler to NR PHY simulators
When using T2 virtual functions, it is important to properly stop DPDK
and free the device. Otherwise the virtual functions may be blocked and
a restart of the admin application is necessary.
If not carefully done, such kind of operation can lead to losing cores
to DPDK processes that cannot be stopped. This can lead to situations
where the machine is locked and can only be unlocked by a power cycle.
Always properly stopping DPDK and freeing the device reduces the risk of
such situation to happen.
Up to now, SIGINT was shutting down the PHY simulators without freeing
the device. This commit adds a signal handler to handle SIGINT in a way
that allow to properly free the device.
This feature is added to all the NR PHY simulators whether they use T2
or not in case it is now or later of any use.
2: Make nr_ulschsim functional
There were two issues that were making nr_ulschsim non functional:
1. The channel output was not copied to decoder input (llr array)
2. The test on decoding successful outcome was wrong
The result was that nr_ulschsim was succesfull whatever were its
arguments. This changeset fixes the two issues so that nr_ulschsim is
now functional.
Harmonize and update Frequency Range computation to the current values
specified by the standard
3GPP TS 38.101-1 Version 19.0.0 Table 5.1-1: Definition of frequency
ranges
- FR1 from 410 MHz to 7125 MHz
- FR2 from 24.25 GHz to 71 GHz
1. Change the name of `openair1/SIMULATION/NR_PHY/nr_dummy_functions.c` into `openair1/SIMULATION/NR_PHY/nr_unitary_common.c` for naming coherence after adding the SIGINT handler which is not a dummy function.
2. Instead of being built once for every NR PHY simulators, the common functions source file is built once as an object and linked to each simulator.
Even though the remaining NR PHY simulators do not use the T2, we may still feature them with a SIGINT handler for proper termination on SIGINT and for a matter of completeness.
The remaining NR PHY simulators are:
* nr_pbchsim
* nr_prachsim
* nr_psbchsim
* nr_pucchsim
There was two issues that were making nr_ulschsim non functional:
1. The channel output was not copied to decoder input (llr array)
2. The test on decoding successful outcome was wrong
The result was that nr_ulschsim was succesfull whatever were its arguments.
This commit fixes the two issues so that nr_ulschsim is now functional.
When using T2 virtual functions, it is important to properly stop DPDK and free the device.
Otherwise the virtual functions may be blocked and a restart of the admin application is necessary.
Up to now, SIGINT was shutting down the PHY simulators without freeing the device.
This commit adds a signal handler to handle SIGINT in a way that allow to properly free the device.
This feature is first added to the PHY simulators that use the T2 which are `nr_ulsim`, `nr_dlsim`, `nr_ulschsim` and `nr_dlschsim`.
fix(nrLDPC_decoder): fix numIter increment
The logs of ./nr_ulsim -n100 -m19 -s12 -S12 -z2 -L4 were showing that sometimes
the decoding of one segment was failing while the CRC check succeeded. This was
happening because the iteration counter numIter was not properly managed so that
when the last iteration was successful to decode the segment still numIter at
the exit was equal to numMaxIter plus 1 so that decoding was considered to have
failed. In other words the last iteration if it occured was always wasted.
This commit offers to solve this problem by simply moving the iteration counter
increment at the end of the iteration loop while remaining careful at fixing the
tests relying on this counter.
Post-Scriptum:
After a first review, we decided to make the counter more standard from a C code
perspective. The counter now counts from 0 to numMaxIter minus 1. If it reaches
numMaxIter, it means that decoding was unsuccessful.
Robert Schmidt