Performance of WCN7850

[Jun Nie]

Jun Nie <jun.nie at linaro.org> 于2024年3月1日周五 16:15写道：
>
> Jeff Johnson <quic_jjohnson at quicinc.com> 于2024年2月29日周四 23:53写道：
> >
> > On 2/27/2024 5:54 AM, Jun Nie wrote:
> > > Hi,
> > >
> > > I am trying to maximize the bandwidth of WCN7850. Does anyone know the
> > > performance status of WCN7850 in mainline? It's about 160Mbps in 6GHz
> > > band and ~600Mbps in 5GHz band on sm8650 platform in my test. The
> > > firmware is from below links:
> > > https://github.com/kvalo/ath11k-firmware/raw/master/ath12k-testing/WCN7850/hw2.0/board.bin
> > > https://github.com/kvalo/ath11k-firmware/raw/master/ath12k-testing/WCN7850/hw2.0/regdb.bin
> > >
> > > While it is ~940Mbps in both 5GHz and 6GHz band with
> > > Android/linux6.1.25, . But the firmware names are different per log,
> > > amss20.bin on Android. Not sure whether the driver is totally
> > > different for Android/linux6.1.25 and mainline.
> > >
> > > Do you have any idea that the low performance on mainline is due to
> > > driver difference or TX/RX power configuration, or firmware version?
> >
> > To debug this issue the development team has asked:
> > #1) Please change ath12k debug_mask to 0xffffffff
> > #2) trigger association to 5G AP
> > #3) reproduce throughput low issue with 5G AP
> > #4) trigger association to 6G AP
> > #5) reproduce throughput low issue with 6G AP
> >
> > Then share the logs with me (redacting anything personal) and I'll pass
> > them on to the developers.
> >
> > Thanks
> > /jeff
> >
Hi Jeff,
The latest finding is that 6GHz TX performance of wcn7850 is as we talked about.
But the RX performance wcn7850 is better than 5GHz case. So the direction
impacts a lot. Not sure whether the bottoleneck is in scheduling side, or any
configuration on TX power.

> Jun