What are things to improve NSA throughput?
Like in LTE we have 256qam, rank, mimo, cqi and bler adjustment.
This too would be in 5G?
MCS settings and so on.
Apart from that how does beamforming helps.
Without beamforming you have single SSB i.e. 2 layers so max rank 2.
With beamforming > 1 SSB i.e. you can use 2 SSB i.e. 4 layers so max rank 4. So your codeword split on 4 layers one of benefit.
Beamforming helps increasing the SINR.
In books only improved.
In field degraded 
No, beamforming works nice on AAU, otherwise you would not have a high SINR.
Also MU_MIMO would be impossible without beamforming.
But how many SSB possible for one UE?
During DL or UL.
Max.
Don’t confuse beamforming just with SSBs.
Beamforming can point the Beam in many other directions than just SSB.
Yes, main gain claimed with beamforming is improved SINR but in field reverse seen.
MU-MIMO is not able to achieve so much success due to less pairing ratio for UE in MU-MIMO.
Very difficult to maintain that orthogonal separation between UE.
Contradictory it’s seen with MU-MIMO user throughput tends to degrade.
Although it’s never claimed MU-MIMO for throughput improvement mainly it’s for spectral efficiency improvement.
But as per theory if beamforming improved SINR then why degradation in user throughput.
If you have different view please put.
Yes, pairing UEs is a challenge but Huawei has parameters for this.
Up to operator to decide how to pair UEs.
If difficult pairing no gain due to lack of MU-MIMO.
If easy pairing than layers will interfere each other and MU-MIMO gain will be low.
So it all depends on number of active users per cell.
The more users there are, spread out, the easier it will be pairing.
And cell throughput may increase 2x 3x due to MU-MIMO, even more.
Yes that possible in trial only.
Practically, in live network, very tough.
No such accurate tool available to give exact location where we need to plug massive mimo cell.
Huawei has ideally 16 PDSCH layers for MU-MIMO for DL and 8 for UL.
That is for ideal pairing of UEs.
Look at MU-MIMO counters and you will see that there is MU-MIMO activity on heavy traffic cells.
Yes we need to filter out cells having high traffic for MU-MIMO.
Huawei algorithm is tricky to understand. They are using time base tilt change for cells as per loading on cells.
There are lots of parameters for MU-MIMO, some of them not easy to understand, I agree.
Coverage scenario is an important parameter for footprint of the cell.
It can have 16 different values.
Depending if you want to cover in horizontal more, on vertical more, on specific target area with high SINR etc
Yes agree 
Now time for 5G MU-MIMO, but I don’t think any network deploy it.
I understand 6G will be cell free massive MIMO, that means cell concept will fade away.
Just thousands of TRPs.
Let’s see what will come out when standards for 6G will be closed.
It’s never ending race then 7Gg will come with Artificial Human World