Why are results not consistant if 100 or 500 packets used in 802-11ax-d​ownlink-of​dma-and-mu​lti-user-m​imo-throug​hput-simul​ation

https://ch.mathworks.com/help/wlan/ug/802-11ax-downlink-ofdma-and-multi-user-mimo-throughput-simulation.html#

states

« The results show for this channel realization at high SNRs (low pathloss) the throughput provided by the MU-MIMO configuration exceeds OFDMA configuration. The packet duration of the MU-MIMO configuration is roughly half that of the OFDMA configuration which provides the throughput gain. »

Surprisingly (for me), this is valid only if 10 (original code) or 100 packets are sent. But as the number of packets is increased (from 100 to 700 as shown below), the throughput provided by the MU-MIMO configuration decreases even below OFDMA configuration… because in the MU-MIMO case, all packets after the first 500 packets have errors.

I cannot find an explanation yet. Is this a code bug ? an artifact from the MU-MIMO behavior as implemented in the code?

Fred

jean-frederic dot wagen at hefr dot ch

for 100 packets (tgaxBase.Seed = 5 or 3);

Simulating OFDMA...

Pathloss 96.0 dB, AP throughput 66.1 Mbps (no error) or 65.8 Mbps (user 2: 2 errors)

Simulating MU-MIMO...

Pathloss 96.0 dB, AP throughput 110.5 Mbps or 110.5 Mbps (no error)

for 200 packets (tgaxBase.Seed 3);

Simulating OFDMA...

Pathloss 96.0 dB, AP throughput 65.1 Mbps (user 2: 12 errors)

Simulating MU-MIMO...

Pathloss 96.0 dB, AP throughput 110.5 Mbps or 110.5 Mbps (no error)

for 500 packets (tgaxBase.Seed 3);

Simulating OFDMA...

Pathloss 96.0 dB, AP throughput 58.7 Mbps (user 2: 225 errors)

Simulating MU-MIMO...

Pathloss 96.0 dB, AP throughput 68.0 (user 1-2-3-4: 261-143-260-106 errors)

for 700 packets (tgaxBase.Seed 3);

Simulating OFDMA...

Pathloss 96.0 dB, AP throughput 56.1 Mbps (user 2: 425 errors)

Simulating MU-MIMO...

Pathloss 96.0 dB, AP throughput 48.5 (user 1-2-3-4: 461-143-460-306 errors = exactly +200 errors from the 500 packets case)