How to fix high latency and low throughput? #300

ozankaraali · 2024-04-03T10:39:16Z

ozankaraali
Apr 3, 2024

Hello everyone,

I have successfully attached my iPhone 14 Pro Max using USRP N310 via 5G-SA-VONR setting with 999 code (I did not need to change it to 001 since private network was enough). And I managed to connect to the internet. I am getting 805/525 with 25 ping (20/19 load pings) on the host machine in the Speedtest. But I am having 44/4 with 39 ping (with 637/1419 loaded ping time). So I am assuming there are some processing issues, since ping gets worse on load.

USRP is connected to the machine with index 1 SFP port, 10GB/s. External Time and Clock source chosen for USRP with Octoclock with GPS lock. Also I did srsran_performance.sh to change governors and stuff. I am using 20 MHz bandwidth with srate of 30.72. I have Xeon Silver 4208 @ 2.10 GHz to 3.20 GHz with 96 GB of memory.

What else should I check to make my connection faster? Any ideas, or tips?

Thank you and best regards,
Ozan

Answered by herlesupreeth

Apr 4, 2024

Ah I see what you mean. Thats because you are using 20Mhz BW in 5G network (if not please post the gNB configuration file you are using). If you use 100Mhz then you get even greater throughput. Since I dont have USRP N310 I am not aware of the sampling rate to set for 100Mhz BW.

Also, if you use MIMO and 256QAM MCS table you can push throughput even higher.

View full answer

herlesupreeth · 2024-04-04T12:27:20Z

herlesupreeth
Apr 4, 2024
Maintainer

And I managed to connect to the internet. I am getting 805/525 with 25 ping (20/19 load pings) on the host machine in the Speedtest. But I am having 44/4 with 39 ping (with 637/1419 loaded ping time). So I am assuming there are some processing issues, since ping gets worse on load.

Didnt understand what you mean here. Can you please elaborate what are those numbers and what you mean by loaded ping time?

3 replies

ozankaraali Apr 4, 2024
Author

Hi Supreeth,
Sorry for my confusing message
Basically when I connect to Speedtest.net, it gives speed test results in Mbps, and it gives 3 different ping latencies, yellow one is regular latency, blue one is download latency when on load, purple one is upload latency when on load.

and in my computer which is connected to the internet has 805 Mbps download with 205 Mbps upload of consistent ~25 ms ping. But when I switch to the 5G network I created, I get:

So as you can see, my upload is way worse than my upload and heavy-load latencies are a bit too high.
Hope I could explain my issue this time better, thank you.

herlesupreeth Apr 4, 2024
Maintainer

Ah I see what you mean. Thats because you are using 20Mhz BW in 5G network (if not please post the gNB configuration file you are using). If you use 100Mhz then you get even greater throughput. Since I dont have USRP N310 I am not aware of the sampling rate to set for 100Mhz BW.

Also, if you use MIMO and 256QAM MCS table you can push throughput even higher.

Answer selected by ozankaraali

ozankaraali Apr 5, 2024
Author

Hi Supreeth,
Thank you for the quick response. You are correct, I was using it with 20 MHz.
I will attach gnb config on the most bottom of this reply. So I was successful with 30.72 srate with 20 MHz. If I make 10 MHz, iPhone does not connect but gnb is stable. If I make 30 MHz, I got 67,6 download and 4,22 upload, which is better than the initial value, thanks.
if I add

  pdsch:
    mcs_table: qam256

my upload goes worse to 1 mpbs.

If I increase srate to 61.44 or 122.88
I get

srsgnb  | The PRACH detector will not meet the performance requirements with the configuration {Format B4, ZCZ 0, SCS 30kHz, Rx ports 1}.
srsgnb  | Lower PHY in quad executor mode.
srsgnb  | --== srsRAN gNB (commit 342df75a9) ==--
srsgnb  | 
srsgnb  | Connecting to AMF on 172.22.0.10:38412
srsgnb  | Available radio types: uhd and zmq.
srsgnb  | [INFO] [UHD] linux; GNU C++ version 11.4.0; Boost_107400; UHD_4.6.0.0-0ubuntu1~jammy1
srsgnb  | [INFO] [LOGGING] Fastpath logging disabled at runtime.
srsgnb  | No device type given, found device with address 'mgmt_addr=192.168.20.2,type=n3xx,product=n310,serial=3198220,name=ni-n3xx-3198220,fpga=HG,claimed=False'
srsgnb  | Making USRP object with args 'addr=192.168.20.2,num_recv_frames=128,num_send_frames=128,clock_source=external,time_source=external,type=n3xx,master_clock_rate=122.88e6,send_frame_size=8000,recv_frame_size=8000'
srsgnb  | [INFO] [MPMD] Initializing 1 device(s) in parallel with args: mgmt_addr=192.168.20.2,type=n3xx,product=n310,serial=3198220,name=ni-n3xx-3198220,fpga=HG,claimed=False,addr=192.168.20.2,num_recv_frames=128,num_send_frames=128,clock_source=external,time_source=external,master_clock_rate=122.88e6,send_frame_size=8000,recv_frame_size=8000
srsgnb  | [WARNING] [MPM.RPCServer] A timeout event occured!
srsgnb  | [INFO] [MPM.PeriphManager] init() called with device args `clock_source=external,fpga=HG,master_clock_rate=122.88e6,mgmt_addr=192.168.20.2,name=ni-n3xx-3198220,num_recv_frames=128,num_send_frames=128,product=n310,recv_frame_size=8000,send_frame_size=8000,time_source=external'.
srsgnb  | [WARNING] [0/Radio#0] Attempting to set tick rate to 0. Skipping.
srsgnb  | Cell pci=1, bw=100 MHz, dl_arfcn=676666 (n77), dl_freq=4149.99 MHz, dl_ssb_arfcn=673920, ul_freq=4149.99 MHz
srsgnb  | 
srsgnb  | ==== gNodeB started ===
srsgnb  | Type <t> to view trace
srsgnb  | Late: 0; Underflow: 7; Overflow: 538;
srsgnb  | Error: exceeded maximum number of timed out transmissions.
srsgnb  | Error: exceeded maximum number of timed out transmissions.
srsgnb  | Error: exceeded maximum number of timed out transmissions.

I remember that if I don't use SFP, I was getting this message even in 30.72 srate / 20 MHz but I don't think 10 Gbits of connection is choking (unless I should run dual 10 Gbits).
Also as I said that, I am running with DRM KMS disabled, buffers maxed and cpu is on performance governor.

amf:
  addr: AMF_IP                                                    # The address or hostname of the AMF.
  bind_addr: SRS_GNB_IP                                           # A local IP that the gNB binds to for traffic from the AMF.

cu_cp:
  security:
    nea_pref_list: nea2,nea1,nea3,nea0     # Optional TEXT (nea0,nea2,nea1). Ordered preference list for the selection of encryption algorithm (NEA). Supported: [nea0,nea2,nea1,nea3].

ru_sdr:
  device_driver: uhd                                            # The RF driver name.
  device_args: addr=192.168.20.2,num_recv_frames=128,num_send_frames=128,clock_source=external,time_source=external
  # clock_source=external,time_source=external  # Optionally pass arguments to the selected RF driver.
  clock: external                                                       # Specify the clock source used by the RF.
  sync:  external
  # srate: 122.88                                                        # Specify the sync source used by the RF.
  # srate: 61.44 #46.08 122.88 184.32                             # RF sample rate might need to be adjusted according to selected bandwidth.
  srate: 30.72
  # otw_format: sc16                                                  # RF sample rate might need to be adjusted according to selected bandwidth.
  tx_gain: 50                                                   # Transmit gain of the RF might need to adjusted to the given situation.
  rx_gain: 50
  # time_alignment_calibration: -200
  # expert_cfg:
    # low_phy_thread_profile: quad                                                   # Receive gain of the RF might need to adjusted to the given situation.

cell_cfg:
  dl_arfcn: 676666                                                # ARFCN of the downlink carrier (center frequency).
  band: 77                                                        # The NR band.
  channel_bandwidth_MHz: 30                                       # Bandwith in MHz. Number of PRBs will be automatically derived.
  common_scs: 30                                                  # Subcarrier spacing in kHz used for data.
  plmn: "99970"                                                    # PLMN broadcasted by the gNB.
  tac: 7                                                          # Tracking area code (needs to match the core configuration).
  pci: 1  
  pusch:
    max_ue_mcs: 2         
  # pdsch:
  #   mcs_table: qam256
  # pusch:
  #   mcs_table: qam256
  # pucch:
  #   sr_period_ms: 4 # this crashes on my device

# expert_phy:
#   max_proc_delay: 5

# expert_execution:
#   # affinities:
#   #   isolated_cpus: 4
#   # cell_affinities:
#   #   l1_dl_cpus: [7,8]
#   #   l1_ul_cpus: [9,10]
#   #   l2_cell_cpus: [11,12]
#   threads:
#     upper_phy:
#       # nof_pdsch_threads: 4
#       # nof_pusch_threads: 2
#       pdsch_processor_type: auto              # Optional TEXT (auto). Sets the PDSCH processor type. Supported: [auto, generic, concurrent, lite].
#       # nof_pdsch_decoder_threads: 3
#       # nof_pusch_decoder_threads: 5            # Optional UINT (1). Sets the number of threads used to encode PUSCH.
#       nof_dl_threads: 3
#       nof_ul_threads: 3
#     lower_phy:
#       execution_profile: quad
#     ofh:
#       enable_dl_parallelization: 1

log:
  filename: /mnt/srsran/gnb.log                                   # Path of the log file.
  all_level: warning                                              # Logging level applied to all layers.

pcap:
  mac_enable: false                                               # Set to true to enable MAC-layer PCAPs.
  mac_filename: /mnt/srsran/gnb_mac.pcap                          # Path where the MAC PCAP is stored.
  ngap_enable: false                                              # Set to true to enable NGAP PCAPs.
  ngap_filename: /mnt/srsran/gnb_ngap.pcap                        # Path where the NGAP PCAP is stored.
  f1ap_enable: false                                              # Set to true to enable F1AP PCAPs.
  f1ap_filename: /mnt/srsran/gnb_f1ap.pcap                        # Path where the F1AP PCAP is stored.
  e1ap_enable: false                                              # Set to true to enable E1AP PCAPs.
  e1ap_filename: /mnt/srsran/gnb_e1ap.pcap                        # Path where the E1AP PCAP is stored.
  e2ap_enable: false                                              # Set to true to enable E2AP PCAPs.
  e2ap_filename: /mnt/srsran/gnb_e2ap.pcap                        # Path where the E2AP PCAP is stored.
  gtpu_enable: false                                              # Set to true to enable GTP-U PCAPs.
  gtpu_filename: /mnt/srsran/gnb_gtpu.pcap                        # Path where the GTP-U PCAP is stored.

Do you have more tips to make it better? Or do you have any ideas/tips for equipments that I can run a low latency 5G environment?

Thank you and best regards,
Ozan

herlesupreeth · 2024-04-08T10:01:42Z

herlesupreeth
Apr 8, 2024
Maintainer

If I make 10 MHz, iPhone does not connect but gnb is stable.

Almost all phones dont support 10Mhz on TDD band. In order know which BWs the phone support you need to check the UE Capabilites information packet in the MAC pcap

pusch:
max_ue_mcs: 2

The above restricts your UL throughput.

As per my tests with B210, on a 20Mhz BW I get around 50Mbps DL and around 14Mbps UL.

Can you send me the trace output on the gNB console? (i.e. what you get when you press t on gNB console). I suspect your gain settings maybe causing lower throughput.

2 replies

ozankaraali Apr 10, 2024
Author

Sorry for my late response, I was not in the lab until today, and thanks for your message:
Here's my latest gnb config:

amf:
  addr: AMF_IP                                                    # The address or hostname of the AMF.
  bind_addr: SRS_GNB_IP                                           # A local IP that the gNB binds to for traffic from the AMF.

cu_cp:
  security:
    nea_pref_list: nea2,nea1,nea3,nea0     # Optional TEXT (nea0,nea2,nea1). Ordered preference list for the selection of encryption algorithm (NEA). Supported: [nea0,nea2,nea1,nea3].

ru_sdr:
  device_driver: uhd                                            # The RF driver name.
  device_args: mgmt_addr=192.168.20.2,type=n3xx,product=n310,serial=3198220,name=ni-n3xx-3198220,fpga=HG,claimed=False,addr=192.168.20.2,num_recv_frames=64,num_send_frames=64,clock_source=external,time_source=external,master_clock_rate=122.88e6,send_frame_size=8000,recv_frame_size=8000
  # clock_source=external,time_source=external  # Optionally pass arguments to the selected RF driver.
  # clock: external                                                         # Specify the clock source used by the RF.
  # sync:
  # srate: 122.88                                                        # Specify the sync source used by the RF.
  # srate: 61.44 #46.08 122.88 184.32                             # RF sample rate might need to be adjusted according to selected bandwidth.
  srate: 30.72
  # otw_format: sc16                                                  # RF sample rate might need to be adjusted according to selected bandwidth.
  tx_gain: 60                                                   # Transmit gain of the RF might need to adjusted to the given situation.
  rx_gain: 60
  # time_alignment_calibration: -500
  # expert_cfg:
    # low_phy_thread_profile: quad                                                   # Receive gain of the RF might need to adjusted to the given situation.

cell_cfg:
  dl_arfcn: 676666                                                # ARFCN of the downlink carrier (center frequency).
  band: 77                                                        # The NR band.
  channel_bandwidth_MHz: 30                                       # Bandwith in MHz. Number of PRBs will be automatically derived.
  common_scs: 30                                                  # Subcarrier spacing in kHz used for data.
  plmn: "99970"                                                    # PLMN broadcasted by the gNB.
  tac: 7                                                          # Tracking area code (needs to match the core configuration).
  pci: 1  
  # pusch:
  #   max_ue_mcs: 2        
  nof_antennas_ul: 1
  nof_antennas_dl: 1 #2 
  pdsch:
    mcs_table: qam256
  pusch:
    mcs_table: qam256
  # pucch:
  #   sr_period_ms: 4
  # tdd_ul_dl_cfg:
  #   dl_ul_tx_period: 10 # default 10
  #   nof_dl_slots: 3 # default 6
  #   nof_dl_symbols: 0 # default 0
  #   nof_ul_slots: 6 # default 3
  #   nof_ul_symbols: 0 # default 0
  #   pattern2:
  #     dl_ul_tx_period: 10 # default 10
  #     nof_dl_slots: 3 # default 6
  #     nof_dl_symbols: 0 # default 0
  #     nof_ul_slots: 6 # default 3
  #     nof_ul_symbols: 0 # default 0

  # tdd_ul_dl_cfg:
  #   dl_ul_tx_period: 10
  #   nof_dl_slots: 3
  #   nof_dl_symbols: 9
  #   nof_ul_slots: 5
  #   nof_ul_symbols: 1

# expert_phy:
#   max_proc_delay: 5

expert_execution:
#   # affinities:
#   #   isolated_cpus: 4
#   # cell_affinities:
#   #   l1_dl_cpus: [7,8]
#   #   l1_ul_cpus: [9,10]
#   #   l2_cell_cpus: [11,12]
  threads:
    upper_phy:
      # nof_pdsch_threads: 4
#       # nof_pusch_threads: 2
      pdsch_processor_type: auto              # Optional TEXT (auto). Sets the PDSCH processor type. Supported: [auto, generic, concurrent, lite].
#       # nof_pdsch_decoder_threads: 3
#       # nof_pusch_decoder_threads: 5            # Optional UINT (1). Sets the number of threads used to encode PUSCH.
      nof_dl_threads: 2
      nof_ul_threads: 2
    lower_phy:
      execution_profile: quad
#     ofh:
#       enable_dl_parallelization: 1

log:
  filename: /mnt/srsran/gnb.log                                   # Path of the log file.
  all_level: warning                                              # Logging level applied to all layers.

pcap:
  mac_enable: false                                               # Set to true to enable MAC-layer PCAPs.
  mac_filename: /mnt/srsran/gnb_mac.pcap                          # Path where the MAC PCAP is stored.
  ngap_enable: false                                              # Set to true to enable NGAP PCAPs.
  ngap_filename: /mnt/srsran/gnb_ngap.pcap                        # Path where the NGAP PCAP is stored.
  f1ap_enable: false                                              # Set to true to enable F1AP PCAPs.
  f1ap_filename: /mnt/srsran/gnb_f1ap.pcap                        # Path where the F1AP PCAP is stored.
  e1ap_enable: false                                              # Set to true to enable E1AP PCAPs.
  e1ap_filename: /mnt/srsran/gnb_e1ap.pcap                        # Path where the E1AP PCAP is stored.
  e2ap_enable: false                                              # Set to true to enable E2AP PCAPs.
  e2ap_filename: /mnt/srsran/gnb_e2ap.pcap                        # Path where the E2AP PCAP is stored.
  gtpu_enable: false                                              # Set to true to enable GTP-U PCAPs.
  gtpu_filename: /mnt/srsran/gnb_gtpu.pcap                        # Path where the GTP-U PCAP is stored.

And this is the trace when I do the speed test:
I get around 91.3 Mbps download, 6.30Mbps Upload with 39 latency, with 311 download / 826 upload latency:
I will be trying to change some more configs and experiment a bit more but still things look kinda promising.

[WARNING] [0/Radio#0] Attempting to set tick rate to 0. Skipping.
Cell pci=1, bw=30 MHz, dl_arfcn=676666 (n77), dl_freq=4149.99 MHz, dl_ssb_arfcn=676032, ul_freq=4149.99 MHz

==== gNodeB started ===
Type <t> to view trace
t
Late: 0; Underflow: 37; Overflow: 0;

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4601   15   1   25    13k   11    0   0%      0 |   4.4    2    26k    7    0   0%      0
   1 4601   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4602   15   1   25    15k   13    1   7%      0 |   8.5   12    28k   10    5  33%      0
   1 4602   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4602   15   1   26   2.9k    3    0   0%      0 |   1.3   18   2.3k    1    5  83%      0
   1 4602   15   1   26   420k  113    0   0%      0 |   1.3    0   154k   83    0   0%      0
   1 4602   15   1   26   265k   72    0   0%      0 |   3.2    0   210k   99    0   0%      0
   1 4602   15   1   27   132k   43    0   0%      0 |  13.6   11    81k   16    0   0%      0
   1 4602   15   1   27   115k    9    0   0%      0 |   9.0    6    21k    5    0   0%      0
   1 4602   15   1   27   143k   38    0   0%      0 |   8.9    6    64k   21    1   4%      0
   1 4602   15   1   27   2.5M  338    0   0%      0 |   3.2    0   599k  292    0   0%  1.45k

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4602   15   1   27   1.8M  234    0   0%      0 |   3.8    1   320k  130    0   0%      0
   1 4602   15   1   27   259k   60    0   0%      0 |   5.0    1   164k   52    0   0%  5.45k
   1 4602   15   1   27   262k  104    0   0%      0 |   4.6    1   219k   65    0   0%      0
   1 4602   15   1   27   1.5M  230    0   0%      0 |   4.8    2   238k   82    0   0%      0
   1 4602   15   1   27   878k  140    0   0%      0 |   3.9    1   304k  115    0   0%      0
   1 4602   15   1   27    21k   19    0   0%      0 |   3.4    0    28k   11    0   0%      0
   1 4602   15   1   27    45M  712    0   0%   2.1M |   3.6    0   373k  162    0   0%    384
   1 4602   15   1   27    94M 1400    0   0%  1.51M |   4.8    2   475k  172    0   0%    745
   1 4602   15   1   27    94M 1400    0   0%  1.93M |   4.5    2   475k  153    1   0%    745
   1 4602   15   1   27    94M 1400    0   0%  2.14M |   4.7    2   465k  162    0   0%    384
   1 4602   15   1   27    94M 1400    0   0%  2.25M |   3.8    1   442k  171    0   0%    384

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4602   15   1   27    94M 1400    0   0%   2.1M |   4.7    2   481k  141    1   0%    745
   1 4602   15   1   27    94M 1400    0   0%  1.62M |   4.9    2   535k  140    0   0%  1.04k
   1 4602   15   1   27    94M 1400    0   0%  1.87M |   4.8    2   532k  135    3   2%  1.45k
   1 4602   15   1   27    94M 1400    0   0%  1.77M |   5.2    2   692k  157    2   1%  2.01k
   1 4602   15   1   27    94M 1400    0   0%  1.85M |   5.0    2   690k  147    0   0%      0
   1 4602   15   1   27    83M 1229    0   0%      0 |   5.2    3   465k  105    1   0%      0
   1 4602   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4602   15   1   27   5.9k    3    0   0%      0 |   5.1    3   9.2k    3    0   0%  1.04k
   1 4602   15   1   27   389k  344    0   0%      0 |   5.0    3   5.1M  560    1   0%   300k
   1 4602   15   1   27   466k  418    0   0%      0 |   5.0    4   6.3M  599    1   0%   300k
   1 4602   15   1   27   466k  411    0   0%      0 |   5.0    4   6.4M  599    1   0%   300k

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4602   15   1   27   426k  380    0   0%      0 |   5.1    4   6.5M  595    5   0%   300k
   1 4602   15   1   27   443k  394    1   0%      0 |   5.1    4   6.5M  592    8   1%   300k
   1 4602   15   1   27   418k  375    0   0%      0 |   5.1    4   6.5M  594    6   1%   300k
   1 4602   15   1   27   427k  383    0   0%      0 |   5.2    4   6.5M  591    9   1%   300k
   1 4602   15   1   27   413k  369    0   0%      0 |   5.2    4   6.5M  595    5   0%   300k
   1 4602   15   1   27   524k  412    0   0%      0 |   5.9    5   7.4M  594    6   1%   300k
   1 4602   15   1   27   527k  414    0   0%      0 |   5.9    5   7.2M  594    6   1%   300k
   1 4602   15   1   27   463k  395    0   0%      0 |   5.3    4   6.4M  593    7   1%   300k
   1 4602   15   1   27   455k  392    0   0%      0 |   5.5    4   6.9M  594    6   1%   300k
   1 4602   15   1   27   245k  222    0   0%      0 |   6.5    5   4.5M  326    6   1%      0
   1 4602   15   1   27   131k   34    0   0%      0 |   5.4    4   335k   38    0   0%      0

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4602   15   1   26   1.5M  193    0   0%      6 |   6.4    5   226k   50    4   7%      0
   1 4602   15   1   27   1.1M  111    2   1%      0 |  15.3   16   108k   19   14  42%      0
   1 4602   15   1   27   111k   49    1   2%      0 |  15.0   16   127k   25   11  30%      0
   1 4602   15   1   27   122k   47    0   0%    686 |  12.4    9    87k   24    0   0%      0
   1 4602   15   1   25    20k   11    0   0%      0 |   6.4    3    17k    4    0   0%      0
   1 4602   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4602   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4602   15   1   26    17k   11    1   8%      0 |  11.9    6    19k    5    0   0%      0
   1 4602   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4602   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4602   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0

But still any tips will be appreciated since I need to optimize latency for upload.
Thank you for your help, you really helped me achieve a lot! :)

ozankaraali Apr 10, 2024
Author

In addition:


--== srsRAN gNB (commit 342df75a9) ==--

Connecting to AMF on 172.22.0.10:38412
Available radio types: uhd and zmq.
[INFO] [UHD] linux; GNU C++ version 11.4.0; Boost_107400; UHD_4.6.0.0-0ubuntu1~jammy1
[INFO] [LOGGING] Fastpath logging disabled at runtime.
Making USRP object with args 'mgmt_addr=192.168.20.2,type=n3xx,product=n310,serial=3198220,name=ni-n3xx-3198220,fpga=HG,claimed=False,addr=192.168.20.2,num_recv_frames=64,num_send_frames=64,clock_source=external,time_source=external,master_clock_rate=122.88e6,send_frame_size=8000,recv_frame_size=8000'
[INFO] [MPMD] Initializing 1 device(s) in parallel with args: mgmt_addr=192.168.20.2,type=n3xx,product=n310,serial=3198220,name=ni-n3xx-3198220,fpga=HG,claimed=False,addr=192.168.20.2,num_recv_frames=64,num_send_frames=64,clock_source=external,time_source=external,master_clock_rate=122.88e6,send_frame_size=8000,recv_frame_size=8000
[INFO] [MPM.Magnesium-0] Re-initializing daughter board. This may take some time.
[INFO] [MPM.Magnesium-1] Re-initializing daughter board. This may take some time.
[INFO] [MPM.PeriphManager] init() called with device args `clock_source=external,fpga=HG,master_clock_rate=122.88e6,mgmt_addr=192.168.20.2,name=ni-n3xx-3198220,num_recv_frames=64,num_send_frames=64,product=n310,recv_frame_size=8000,send_frame_size=8000,time_source=external'.
[INFO] [MPM.Magnesium-0] Re-initializing daughter board. This may take some time.
[INFO] [MPM.Magnesium-1] Re-initializing daughter board. This may take some time.
[INFO] [MPM.Magnesium-0] Re-initializing daughter board. This may take some time.
[INFO] [MPM.Magnesium-1] Re-initializing daughter board. This may take some time.
WARNING] [0/Radio#0] Attempting to set tick rate to 0. Skipping.
Cell pci=1, bw=50 MHz, dl_arfcn=676666 (n77), dl_freq=4149.99 MHz, dl_ssb_arfcn=675360, ul_freq=4149.99 MHz

==== gNodeB started ===
Type <t> to view trace
Late: 8; Underflow: 3734; Overflow: 0;
tLate: 19; Underflow: 3689; Overflow: 0;

Late: 0; Underflow: 3725; Overflow: 0;
Late: 0; Underflow: 3738; Overflow: 0;
Late: 0; Underflow: 3732; Overflow: 0;
Late: 0; Underflow: 3687; Overflow: 0;
Late: 0; Underflow: 3725; Overflow: 0;
tLate: 33; Underflow: 3693; Overflow: 0;
Late: 0; Underflow: 3745; Overflow: 0;
Late: 73; Underflow: 3669; Overflow: 0;
t
Available commands:
        t: start/stop console trace
        q: quit application

Late: 0; Underflow: 3746; Overflow: 0;
Late: 0; Underflow: 3783; Overflow: 0;
t
t
Late: 0; Underflow: 3746; Overflow: 0;
tLate: 0; Underflow: 3736; Overflow: 0;
tLate: 0; Underflow: 3714; Overflow: 0;
ttLate: 0; Underflow: 3719; Overflow: 0;
tLate: 0; Underflow: 3766; Overflow: 0;

when I try to make srate 61.44 and BW of 50, I get a lot of underflow and I cannot even open trace, any ideas?

herlesupreeth · 2024-04-11T07:35:15Z

herlesupreeth
Apr 11, 2024
Maintainer

If you are after DL speeds then I would suggest to have the following TDD pattern

tdd_ul_dl_cfg:
    nof_dl_symbols: 7
    nof_dl_slots: 7
    nof_ul_slots: 2

I dont know much about expert_execution configuration. In order to get more throughout you could increase the following (you get increase only in DL since MIMO in UL is not yet implemented)

  nof_antennas_ul: 2
  nof_antennas_dl: 2

Regarding BW 50Mhz, srate looks to be good (atleast from perspective of B210). I would suggest removing expert_execution configuration and giving it a try.

1 reply

ozankaraali Apr 11, 2024
Author

No, upload low latency is the first goal for me but thanks for this tip.
Also I don’t know why but if I chose nof_antennas_ul: 2 but dl as 1, it works and it boosts the upload a bit, but if I make 2/2, it crashes. I assume I need to try that with a better CPU. Hopefully around next week I can try with it.
Regarding 50 MHz, srate should be >50, which is 61.44 for my device but it also gives underflow, so with 30.72 I can only go until 30 MHz, but thanks for the tips, that helped a lot. :)

Edit: I chose the most important tip as an answer but still I would not close the discussion, so if I have any more updates and tips that I can discover, I will post a reply to this thread.

herlesupreeth · 2024-04-11T08:09:34Z

herlesupreeth
Apr 11, 2024
Maintainer

If low latency is important then you can try the following config

  pucch:
    sr_period_ms: 10
    min_k1: 2
  csi:
    csi_rs_period: 40
  pusch:
    mcs_table: qam256
    min_k2: 2

1 reply

ozankaraali Apr 19, 2024
Author

Hi @herlesupreeth ,
Sorry for my late response, I was trying to test on a different computer to see if I can run 61.44 srate, it did not work. Meanwhile, I have read somewhere that I need to reduce my rx_gain to not catch noise but only active elements, I thought it was caused by that but it was not the case also.
Since I had two antennas, I decided to give it a go with the following:

amf:
  addr: AMF_IP                                                    # The address or hostname of the AMF.
  bind_addr: SRS_GNB_IP                                           # A local IP that the gNB binds to for traffic from the AMF.

cu_cp:
  security:
    nea_pref_list: nea2,nea1,nea3,nea0

ru_sdr:
  device_driver: uhd                                            # The RF driver name.
  device_args: addr=192.168.20.2,num_recv_frames=64,num_send_frames=64,clock_source=external,time_source=external  # Optionally pass arguments to the selected RF driver.
  clock: external                                                       # Specify the clock source used by the RF.
  sync: external                                                        # Specify the sync source used by the RF.
  srate: 30.72                                                  # RF sample rate might need to be adjusted according to selected bandwidth.
  tx_gain: 60                                                   # Transmit gain of the RF might need to adjusted to the given situation.
  rx_gain: 50                                                   # Receive gain of the RF might need to adjusted to the given situation.

cell_cfg:
  dl_arfcn: 676666                                                # ARFCN of the downlink carrier (center frequency).
  band: 77                                                        # The NR band.
  channel_bandwidth_MHz: 20                                       # Bandwith in MHz. Number of PRBs will be automatically derived.
  common_scs: 30                                                  # Subcarrier spacing in kHz used for data.
  plmn: "99970"                                                    # PLMN broadcasted by the gNB.
  tac: 7                                                          # Tracking area code (needs to match the core configuration).
  pci: 1                                                          # Physical cell ID.
  nof_antennas_ul: 2
  nof_antennas_dl: 1

log:
  filename: /mnt/srsran/gnb.log                                   # Path of the log file.
  all_level: warning                                              # Logging level applied to all layers.

pcap:
  mac_enable: false                                               # Set to true to enable MAC-layer PCAPs.
  mac_filename: /mnt/srsran/gnb_mac.pcap                          # Path where the MAC PCAP is stored.
  ngap_enable: false                                              # Set to true to enable NGAP PCAPs.
  ngap_filename: /mnt/srsran/gnb_ngap.pcap                        # Path where the NGAP PCAP is stored.
  f1ap_enable: false                                              # Set to true to enable F1AP PCAPs.
  f1ap_filename: /mnt/srsran/gnb_f1ap.pcap                        # Path where the F1AP PCAP is stored.
  e1ap_enable: false                                              # Set to true to enable E1AP PCAPs.
  e1ap_filename: /mnt/srsran/gnb_e1ap.pcap                        # Path where the E1AP PCAP is stored.
  e2ap_enable: false                                              # Set to true to enable E2AP PCAPs.
  e2ap_filename: /mnt/srsran/gnb_e2ap.pcap                        # Path where the E2AP PCAP is stored.
  gtpu_enable: false                                              # Set to true to enable GTP-U PCAPs.
  gtpu_filename: /mnt/srsran/gnb_gtpu.pcap                        # Path where the GTP-U PCAP is stored.

As you can see from the trace, my pusch value went well in the beginning and I managed to record 43.7 Mbps download, and 10.5 Mbps upload, which was okay, but the thing I don't understand, each experiment, my upload can burst even to 70-80 Mbps for a second, then, it drops. So, when I retried once more, second time I got 44 Mbps download, 3.58 upload and as you can see my pusch values go to minus. Do you have any ideas about that?

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4604   15   1   26   200k   64    0   0%      0 |  15.3   23    98k   29    0   0%      0
   1 4604   15   1   26    26k   25    0   0%      0 |  14.4   22    47k   12    0   0%      0
   1 4604   15   1   26   363k   77    0   0%    771 |  14.0   21   134k   21    0   0%  1.04k
   1 4604   15   1   26   391k  138    0   0%      0 |  16.1   24   360k   52    2   3%      0
   1 4604   15   1   27   809k  132    0   0%      0 |  14.7   23   473k   56    0   0%      0
   1 4604   15   1   27   2.0k    3    0   0%      0 |  13.9   22    39k    5    1  16%      0
   1 4604   15   1   27    32k   41    0   0%      0 |  14.3   21    77k   18    0   0%      0
   1 4604   15   1   27    25k   12    0   0%  8.86k |  14.0   21    56k    9    0   0%      0
   1 4604   15   1   28    43M 1340    0   0%  1.68M |  11.4   16   635k   98    0   0%    745
   1 4604   15   1   28    46M 1400    0   0%   1.9M |  11.3   15   570k   99    0   0%    384
   1 4604   15   1   28    46M 1400    0   0%  2.13M |  13.3   19   485k  100    0   0%  1.04k

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4604   15   1   28    45M 1387   13   0%  2.37M |  12.2   18   606k   98    0   0%    745
   1 4604   15   1   27    45M 1400    0   0%  2.47M |  13.8   21   572k  100    0   0%    745
   1 4604   15   1   28    46M 1398    2   0%  2.78M |  14.1   22   445k   97    2   2%    276
   1 4604   15   1   28    46M 1400    0   0%  2.68M |  14.2   22   571k  100    0   0%    535
   1 4604   15   1   28    46M 1398    2   0%  2.54M |  13.9   22   711k   98    1   1%    535
   1 4604   15   1   28    46M 1400    0   0%  2.87M |  14.2   22   561k   99    0   0%    535
   1 4604   15   1   28    45M 1392    8   0%  2.52M |  13.6   22   646k   97    3   3%  1.04k
   1 4604   15   1   28    45M 1398    2   0%   2.4M |  14.1   22   650k  100    0   0%  1.04k
   1 4604   15   1   27    45M 1391    9   0%  2.86M |  14.2   22   596k   91    3   3%    745
   1 4604   15   1   28    46M 1399    1   0%   285k |  14.1   21   464k   79    2   2%      0
   1 4604   15   1   28   2.6M  124    0   0%      0 |  14.3   22   111k   25    0   0%      0

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4604   15   1   28    67k   54    0   0%      0 |  11.7   17   1.2M   90    0   0%   150k
   1 4604   15   1   28   489k  471    0   0%     56 |  11.6   16   8.4M  598    2   0%   300k
   1 4604   15   1   28   589k  512    0   0%    112 |  11.7   17   9.1M  594    6   1%   300k
   1 4604   15   1   28   561k  490    0   0%      0 |  11.6   17   9.1M  597    3   0%   300k
   1 4604   15   1   28   537k  489    0   0%    151 |  11.9   18  10.0M  598    2   0%   300k
   1 4604   15   1   28   523k  459    0   0%     56 |  12.0   19    11M  597    3   0%   300k
   1 4604   15   1   28   561k  458    0   0%      0 |  12.9   19    10M  588   12   2%   300k
   1 4604   15   1   28   581k  479    0   0%     56 |  12.9   20    11M  597    3   0%   300k
   1 4604   15   1   28   991k  493    0   0%    115 |  12.9   21    12M  595    5   0%   300k
   1 4604   15   1   28   511k  445    0   0%     56 |  12.9   21    12M  595    5   0%   300k
   1 4604   15   1   28   507k  451    0   0%    209 |  12.9   20    12M  594    6   1%   300k

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4604   15   1   28   510k  449    0   0%     56 |  12.9   20    11M  590   10   1%   300k
   1 4604   15   1   28   486k  467    0   0%     56 |  12.9   20    11M  596    4   0%   300k
   1 4604   15   1   28   512k  475    0   0%     59 |  12.9   20    11M  594    6   1%   300k
   1 4604   15   1   28   507k  455    0   0%     56 |  13.0   20    12M  596    4   0%   300k
   1 4604   15   1   28   500k  444    0   0%    112 |  12.9   21    12M  594    6   1%   300k
   1 4604   15   1   28   755k  172    0   0%      6 |  13.3   22   2.7M  151   14   8%      0
   1 4604   15   1   27   118k   54    0   0%      0 |  15.6   25   154k   24    1   4%      0
   1 4604   15   1   28    13k   16    0   0%      0 |  15.5   25    57k   13    0   0%      0
   1 4604   15   1   28   1.5k    2    0   0%      0 |  17.7   27   4.2k    1    0   0%      0
   1 4604   15   1   28   133k   29    0   0%      0 |  -2.9    4    51k   41    9  18%      0
   1 4604   15   1   28   1.5k    2    0   0%      0 |  -4.9    0   2.8k    2    0   0%      0

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4604   15   1   28   1.5k    2    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4604   15   1   28   1.5k    2    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4604   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4604   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4604   15   1   28   1.5k    2    0   0%      0 |  -0.8    0   1.4k    1    0   0%      0
   1 4604   15   1   28   1.5k    2    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4604   15   1   27   1.4k    2    0   0%      0 |   n/a    0      0    0    0   0%      0
Late: 0; Underflow: 2; Overflow: 0;
   1 4604   15   1   28   166k   37    0   0%      0 |  -1.3    3   160k   66    0   0%      0
   1 4604   15   1   28    736    1    1  50%      0 |   n/a    0      0    0    0   0%      0
   1 4604   15   1   28    58k   10    0   0%      6 |  -0.6    3    85k   38    0   0%  2.81k
   1 4604   15   1   28    94k   38    4   9%      0 |   0.3    4   172k   57    0   0%      0

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4604   15   1   28   571k   88    1   1%      0 |   4.1    8   154k   35    0   0%      0
   1 4604   15   1   28    30k   13    0   0%     60 |  -2.1    2    12k    7    1  12%  2.81k
   1 4604   15   1   28   180k   70    0   0%      0 |   2.7    6   259k   72    0   0%      0
   1 4604   15   1   28   4.4M  157    0   0%   306k |   9.5   14   184k   29    1   3%    102
Late: 0; Underflow: 2; Overflow: 0;
   1 4604   15   1   25    39M 1349   51   3%   2.6M |   3.8    8   389k  118    5   4%    198
   1 4613   15   1   28    17M  530    0   0%   288k |  -0.6    4   190k   71    2   2%    276
   1 4613   15   1   25    39M 1377   23   1%  1.18M |  -3.2    1   371k  275    7   2%    142
   1 4613   15   1   26    43M 1396    4   0%   1.7M |  -2.0    2   417k  238    2   0%    745
   1 4613   15   1   28    45M 1396    4   0%  1.98M |  -1.8    3   545k  260    1   0%    384
   1 4613   15   1   28    46M 1400    0   0%  2.15M |  -1.6    3   591k  278    5   1%      0
   1 4613   15   1   28    46M 1400    0   0%  2.46M |  -1.3    3   390k  189    2   1%    745

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4613   15   1   28    46M 1396    4   0%  2.57M |  -1.1    3   464k  206    1   0%  1.04k
   1 4613   15   1   28    46M 1398    2   0%  2.67M |  -0.4    4   502k  203    0   0%    384
   1 4613   15   1   28    46M 1400    0   0%  2.78M |   1.7    5   444k  136    0   0%    535
Late: 0; Underflow: 1; Overflow: 0;
   1 4613   15   1   28    46M 1396    4   0%  2.89M |   1.7    6   445k  127    0   0%      0
   1 4613   15   1   28    46M 1400    0   0%  2.49M |   1.8    6   641k  197    0   0%      0
   1 4613   15   1   28    46M 1398    2   0%  2.54M |   1.8    6   678k  196    0   0%    745
Late: 37; Underflow: 2; Overflow: 0;
   1 4613   15   1   24    37M 1370   29   2%  2.87M |   2.0    7   463k  113    1   0%    745
   1 4613   15   1   25    40M 1396    4   0%   2.4M |   1.9    6   461k  125    0   0%      0
   1 4613   15   1   27    20M  653    0   0%      0 |   1.7    6   149k   37    0   0%      0
   1 4613   15   1   28   245k  269    0   0%    148 |   1.6    6   2.9M  568    0   0%   300k
   1 4613   15   1   28   260k  275    0   0%     56 |   1.6    7   3.2M  595    5   0%   300k

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4613   15   1   28   300k  325    0   0%     56 |   1.6    7   3.4M  597    3   0%   300k
   1 4613   15   1   28   350k  378    0   0%    112 |   1.9    8   3.8M  599    1   0%   300k
   1 4613   15   1   28   313k  343    0   0%     56 |   1.4    8   3.8M  599    1   0%   300k
   1 4613   15   1   28   416k  368    0   0%      0 |   1.5    8   4.1M  599    1   0%   300k
   1 4613   15   1   28   339k  350    0   0%     56 |   1.5    9   4.2M  594    6   1%   300k
   1 4613   15   1   28   820k  355    0   0%      0 |   1.3    8   4.2M  595    5   0%   300k
   1 4613   15   1   28   268k  284    0   0%     56 |   1.1    8   3.9M  590   10   1%   468k
   1 4613   15   1   28   255k  267    0   0%    112 |   1.0    7   3.7M  593    7   1%   300k
   1 4613   15   1   28   267k  277    0   0%     56 |   1.1    8   3.8M  597    3   0%   300k
   1 4613   15   1   28   176k  207    0   0%      0 |   1.1    8   3.6M  557    5   0%      0
   1 4613   15   1   28   314k   75    0   0%      0 |   3.1   10   383k   69   21  23%      0

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4613   15   1   28   211k   29    1   3%      0 |   2.7    8    58k   14    4  22%      0
   1 4613   15   1   28   243k   64    1   1%      0 |   2.7    8   141k   36    1   2%      0
   1 4613   15   1   28   4.4k    5    0   0%      0 |  -0.0    4    19k    7    0   0%      0
   1 4613   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0
   1 4613   15   1    0      0    0    0   0%      0 |   n/a    0      0    0    0   0%      0

also for the reference, I added these lines you suggested to the upper cfg

  ...
  nof_antennas_ul: 2
  nof_antennas_dl: 1
  pucch:
    sr_period_ms: 10
    min_k1: 2
  csi:
    csi_rs_period: 40
  pusch:
    mcs_table: qam256
    min_k2: 2
  ...

and traced once more, it was way worse 20.8/0.94:

 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4601   15   1   24   450k  149   16   9%      0 |   4.0    2   327k  136    0   0%      0
   1 4601   15   1   22    48k   37    5  11%      0 |   2.1    1    49k   26    0   0%      0
   1 4601   15   1   20   119k   64    7   9%      0 |   1.4    1   173k  122    2   1%      0
   1 4601   15   1   20    68k   55    3   5%      0 |   0.2    0   126k   86    1   1%      0
   1 4601   15   1   20    64k   69    2   2%      0 |   2.6    0   104k   72    1   1%      0
Late: 0; Underflow: 2; Overflow: 0;
   1 4601   15   1   20   165k   62    2   3%      0 |   2.3    0   184k  131    1   0%      0
   1 4601   15   1   20   475k   79    0   0%      0 |   1.8    0    62k   53    0   0%    535
   1 4601   15   1   20   5.2M  358    5   1%  28.6k |   0.8    0   362k  274    7   2%      0
   1 4601   15   1   20    22M 1084   72   6%   425k |   1.5    0   427k  293    2   0%  10.6k
   1 4601   15   1   22    34M 1400    0   0%   645k |   1.0    0   566k  400    0   0%  14.7k
   1 4601   15   1   25    40M 1369   14   1%   465k |   0.5    0   559k  395    0   0%  14.7k
Late: 0; Underflow: 1; Overflow: 0;

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4601   15   1   19    22M 1180  103   8%  92.5k |   0.1    0   412k  269   12   4%  3.91k
Late: 0; Underflow: 9; Overflow: 0;
   1 4601   15   1   19    19M 1022  125  10%   103k |   1.8    2   344k  129    9   6%  2.01k
Late: 0; Underflow: 2; Overflow: 0;
   1 4601   15   1   17    13M  778   60   7%  6.41k |   1.4    1   237k  126    7   5%    745
Late: 0; Underflow: 3; Overflow: 0;
   1 4601   15   1   20    13M  732   62   7%  1.18k |  -1.3    0   219k  176   11   5%      0
   1 4601   15   1   20    15M  772   58   6%  75.6k |  -0.7    0   245k  187   11   5%      0
Late: 0; Underflow: 2; Overflow: 0;
   1 4601   15   1   20    14M  755   57   7%   134k |   0.1    0   233k  174    9   4%     38
   1 4601   15   1   19    13M  781  128  14%      0 |  -0.3    0   216k  153   12   7%  1.45k
Late: 36; Underflow: 2; Overflow: 0;
   1 4601   15   1   20    15M  833   54   6%      0 |  -0.7    0   279k  206    9   4%  2.01k
   1 4601   15   1   20   9.7M  585   59   9%  2.41k |  -1.3    0   218k  172   12   6%      0
   1 4601   15   1   20    20M 1018   20   1%      0 |  -0.6    0   325k  257    5   1%      0
   1 4601   15   1   21   2.4k    5    0   0%      0 |   n/a    0      0    0    0   0%      0

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4601   15   1   21   6.7k    4    0   0%      0 |  -0.2    0   1.4k    1    0   0%      0
   1 4601   15   1   21    38k   57    0   0%      0 |  -0.5    0   475k  339    0   0%   150k
   1 4601   15   1   21    34k   57    0   0%      0 |  -0.5    0   566k  400    0   0%   300k
   1 4601   15   1   22   148k   72    0   0%      0 |  -0.5    0   566k  400    0   0%   300k
Late: 0; Underflow: 25; Overflow: 0;
   1 4601   15   1   22   136k  123    4   3%      0 |  -0.5    0   561k  395    0   0%   300k
   1 4601   15   1   21   135k  153    0   0%      0 |   0.1    1   848k  400    0   0%   300k
Late: 0; Underflow: 1; Overflow: 0;
   1 4601   15   1   21   141k  152    1   0%      0 |  -0.3    1   871k  394    2   0%   300k
   1 4601   15   1   22   196k  166    8   4%      0 |  -0.4    1   899k  396    0   0%   300k
   1 4601   15   1   20   165k  148    0   0%      0 |  -0.6    1   909k  400    0   0%   300k
   1 4601   15   1   21   128k  144    0   0%    130 |  -0.6    1   911k  400    0   0%   300k
   1 4601   15   1   21   128k  144    0   0%      0 |  -0.5    1   919k  400    0   0%   300k

           -----------------DL-----------------------|------------------UL--------------------
 pci rnti  cqi  ri  mcs  brate   ok  nok  (%)  dl_bs | pusch  mcs  brate   ok  nok  (%)    bsr
   1 4601   15   1   21   153k  169    0   0%      0 |  -0.3    1   1.2M  399    1   0%   300k
   1 4601   15   1   22   181k  191    0   0%      0 |   0.1    2   1.4M  399    1   0%   300k
   1 4601   15   1   22   153k  148    0   0%      0 |  -0.5    1   957k  400    0   0%   300k
   1 4601   15   1   22   715k  291    0   0%      0 |   1.3    3   2.3M  400    0   0%   300k
   1 4601   15   1   23   249k  262    0   0%      0 |   1.3    4   2.7M  399    1   0%   300k
   1 4601   15   1   23   239k  217    0   0%      0 |   2.1    4   2.8M  400    0   0%   300k
   1 4601   15   1   24   207k  175    0   0%      0 |   2.5    4   2.1M  297   20   6%      0
   1 4601   15   1   24   554k  120    0   0%      0 |  -0.5    1   220k  154    1   0%      0
   1 4601   15   1   24    43k   23    3  11%      0 |   4.1    5    54k   12    4  25%      0
   1 4601   15   1   24   225k   70    0   0%      0 |  -0.2    0   238k  156    0   0%      0
   1 4601   15   1   24   185k   28    0   0%      0 |  -0.7    0    71k   47    0   0%      0

herlesupreeth · 2024-04-23T18:06:27Z

herlesupreeth
Apr 23, 2024
Maintainer

In order to get consistent results on Intel platforms I would suggest executing following steps - https://gitlab.eurecom.fr/oai/openairinterface5g/-/wikis/OpenAirKernelMainSetup#power-management and then give it a try once more

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

How to fix high latency and low throughput? #300

{{title}}

{{editor}}'s edit

{{editor}}'s edit

Replies: 5 comments 7 replies

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

{{title}}

Select a reply

How to fix high latency and low throughput? #300

ozankaraali Apr 3, 2024

Replies: 5 comments · 7 replies

herlesupreeth Apr 4, 2024 Maintainer

ozankaraali Apr 4, 2024 Author

herlesupreeth Apr 4, 2024 Maintainer

ozankaraali Apr 5, 2024 Author

herlesupreeth Apr 8, 2024 Maintainer

ozankaraali Apr 10, 2024 Author

ozankaraali Apr 10, 2024 Author

herlesupreeth Apr 11, 2024 Maintainer

ozankaraali Apr 11, 2024 Author

herlesupreeth Apr 11, 2024 Maintainer

ozankaraali Apr 19, 2024 Author

herlesupreeth Apr 23, 2024 Maintainer

ozankaraali
Apr 3, 2024

Replies: 5 comments 7 replies

herlesupreeth
Apr 4, 2024
Maintainer

ozankaraali Apr 4, 2024
Author

herlesupreeth Apr 4, 2024
Maintainer

ozankaraali Apr 5, 2024
Author

herlesupreeth
Apr 8, 2024
Maintainer

ozankaraali Apr 10, 2024
Author

ozankaraali Apr 10, 2024
Author

herlesupreeth
Apr 11, 2024
Maintainer

ozankaraali Apr 11, 2024
Author

herlesupreeth
Apr 11, 2024
Maintainer

ozankaraali Apr 19, 2024
Author

herlesupreeth
Apr 23, 2024
Maintainer